Download the AI & Machine Learning For Dummies PRO App: iOS - Android Our AI and Machine Learning For Dummies PRO App can help you Ace the following AI and Machine Learning certifications:
Unraveling August 2023: Spotlight on Generative AI, Tech, Sports and the Month’s Hottest Trends.
Welcome to the hub of the most intriguing and newsworthy trends of August 2023! In this era of rapid development, we know it’s hard to keep up with the ever-changing world of ai, technology, sports, entertainment, and global events. That’s why we’ve curated this one-stop blog post to provide a comprehensive overview of what’s making headlines and shaping conversations. From the mind-bending advancements in artificial intelligence to captivating news from the world of sports and entertainment, we’ll guide you through the highlights of the month. So sit back, get comfortable, and join us as we dive into the core of August 2023!
OpenAI has released a guide for teachers using ChatGPT in their classroom. This guide includes suggested prompts, explanations about ChatGPT’s functionality and limitations, as well as insights into AI detectors and bias.
The company also highlights stories of educators successfully using ChatGPT to enhance student learning and provides prompts to help teachers get started. Additionally, their FAQ section offers further resources and answers to common questions about teaching with and about AI.
OpenAI’s teaching with AI empowers teachers with resources and insights to effectively use ChatGPT in classrooms, benefiting students’ learning experiences. While Competitors like Bard, Bing, and Claude may face pressure to offer similar comprehensive guidance to educators. Failing to do so could put them at a disadvantage in the increasingly competitive AI education market.
Meta announced 2 new AI updates: DINOv2, FACET (FAirness in Computer Vision Evaluation)
Meta has announced the commercial relicensing and expansion of DINOv2, a computer vision model, under the Apache 2.0 license to give developers and researchers more flexibility for downstream tasks.
Meta also introduces FACET (FAirness in Computer Vision Evaluation), a benchmark for evaluating the fairness of computer vision models in tasks such as classification and segmentation. The dataset includes 32,000 images of 50,000 people, with demographic attributes such as perceived gender age group, and physical features.
Why does this matter?
FACET ensures more equitable experiences when interacting with computer vision technology, reducing the risk of bias based on demographics. On the other hand, DINOv2’s availability under the Apache 2.0 license as it empowers developers and researchers to create more versatile computer vision applications.
The Graph of Thoughts (GoT) framework improves the capabilities of LLMs by modeling information as a graph. LLM thoughts are represented as vertices, and edges represent dependencies between these thoughts. GoT allows for combining thoughts, distilling networks of thoughts, and enhancing thoughts using feedback loops.
It outperforms other paradigms like Chain-of-Thought or Tree of Thoughts (ToT) in various tasks, increasing sorting quality by 62% and reducing costs by over 31%. It is also extensible, allowing for new thought transformations and advancing prompting schemes.
This advancement brings LLM reasoning closer to human thinking and brain mechanisms such as recurrence, both of which form complex networks. It makes AI models more versatile and adaptable, with implications on various domains.
Google announced a slew of massive AI updates at the Google Cloud Next 2023 event. Here are some key announcements:
Vertex AI extends enterprise-ready generative AI development with new models and tooling. Google Cloud gets a curated collection of models across first-party, open-source, and third-party models, including Meta’s Llama 2 and Code Llama, Falcon, Anthropic’s Claude 2, and more. Google’s foundation models– PaLM, Codey, and Imagen– also get several updates.
Powered by DeepMind, a new tool called SynthID helps watermark and identify synthetic images created by Imagen.
Google is expanding its AI-optimized infrastructure with the general availability of Cloud TPU v5e and Nvidia-powered A3 VMs.
Duet AI in Workspace (aiding tasks across meetings, documents, Google Chat, Gmail, and more) is now generally available, and Duet AI in Google Cloud (to assist in code refactoring, improving, etc.) is expanding its preview and will be generally available later this year
.
Duet AI in Google Cloud also includes advancements for software development, application infrastructure and operations, data analytics, accelerating and modernizing databases, and security operations.
Search Generative Experience (SGE) launches in the first countries outside the U.S. — India and Japan (with multilingual and local language support).
Why does it matter?
The advancements seem to offer a complete solution for AI, from computing infrastructure to end-to-end software and services that support the full lifecycle of model training, tuning, and serving at global scale. It will help organizations harness the full potential of AI with data and cloud through a unified foundation.
Introducing Zapier AI Chatbot: Create custom AI chatbots with no code
Now you can build your own AI-powered chatbot through Zapier Interfaces, its no-code, automation-powered app builder currently in beta. You also have a variety of sharing options, so you can embed chatbots on your website or limit access to your team or external stakeholders.
The base AI Chatbot model is GPT-3.5. With Interfaces Premium, you can connect to other models (like GPT-4) using an API key from your personal OpenAI account.
Why does this matter?
This makes it easier for businesses and individuals to create custom AI chatbots, no coding required. It democratizes AI chatbot development, potentially increasing their accessibility across various industries/applications and fostering innovation in AI.
Meta researchers find AI “Déjà Vu”ing: Suggested ways to address the privacy risks; Meta’s ImageBind: The ultimate fusion of 6 data types in 1 AI model; Meta’s Sandbox: Where AI meets advertising; Meta bets big on AI with custom chips & a supercomputer; Meta scaling Speech Technology to 1,100+ languages; Meta’s MusicGen: The LLaMA moment for music AI; Meta disclosed AI behind Facebook and Instagram recommendations; Meta merges ChatGPT & Midjourney into one; Meta unveils Llama 2, a worthy rival to ChatGPT; Meta-Transformer lets AI models process 12 modalities; Meta collabs with Qualcomm to enable on-device AI apps using Llama 2; Meta’s AudioCraft is AudioGen + MusicGen + EnCodec; Meta challenges OpenAI with code-gen free software; Meta’s SeamlessM4T: The first all-in-one, multilingual multimodal AI; Meta to rival GPT-4 with a free Llama 3?
Meta researchers find AI “Déjà Vu”ing: Suggested ways to address the privacy risks
Researchers at Meta recently discovered an anomaly common across most Self Supervised Learning (SSL) algorithms and call it Déjà Vu. They said SSL models can unintendedly memorize specific parts in individual training samples rather than learning semantically meaningful associations.
The report shares the details of studies around this unintended memorization and also explores ways of avoiding it.
Meta’s ImageBind: The ultimate fusion of 6 data types in 1 AI model
Meta has announced the new open-source AI model called ‘ImageBind’ that links together multiple data streams- text, audio, visual data, temperature, and movement readings. ImageBind is the first to combine 6 data types into a single embedding space.
The company also notes that other streams of sensory input could be added to future models, including touch, speech, smell, and brain fMRI signals.
Meta’s Sandbox: Where AI meets advertising
Meta has introduced an AI Sandbox for advertisers, which includes features such as alternative copy generation, background creation through text prompts, and image cropping for Facebook or Instagram ads. This new tool aims to assist advertisers in creating more diverse and engaging content using AI.
The tools are still in beta, but they have the potential to revolutionize how ads are created and delivered.
Meta bets big on AI with custom chips & a supercomputer
Meta is making a big bet on AI by developing custom chips and a supercomputer. The company is developing its own chips called the Meta Training and Inference Accelerator (MTIA), which will be optimized for AI workloads and allow for more efficient training and running of complex models.
In addition, Meta is building a supercomputer, which will be used to train large-scale AI models for natural language processing and computer vision. These investments aim to enable the development of more advanced products and services, such as virtual assistants and augmented reality applications.
Meta scaling Speech Technology to 1,100+ languages
Meta’s Massively Multilingual Speech (MMS) project aims to address the lack of speech recognition models for most of the world’s languages, introduced Introducing speech-to-text, text-to-speech. Combining self-supervised learning techniques with a new dataset containing labeled data for over 1,100 languages and unlabeled data for nearly 4,000 languages.
The MMS models outperform existing ones and cover 10 times as many languages. The project’s goal is to increase accessibility to information for people who rely on voice as their primary means of accessing information. The models and code are publicly available for further research and development. The project aims to contribute to the preservation of the world’s diverse languages.
Meta’s AI Segmentation Game Changer
Meta’s researchers have developed HQ-SAM (High-Quality Segment Anything Model), a new model that improves the segmentation capabilities of the existing SAM. SAM struggles to segment complex objects accurately, despite being trained with 1.1 billion masks. HQ-SAM is trained on a dataset of 44,000 fine-grained masks from various sources, achieving impressive results on nine segmentation datasets across different tasks.
HQ-SAM retains SAM’s prompt design, efficiency, and zero-shot generalizability while requiring minimal additional parameters and computation. Training HQ-SAM on the provided dataset takes only 4 hours on 8 GPUs.
Meta plans to put AI everywhere on its platforms
Meta has announced plans to integrate generative AI into its platforms, including Facebook, Instagram, WhatsApp, and Messenger. The company shared a sneak peek of AI tools it was building, including ChatGPT-like chatbots planned for Messenger and WhatsApp that could converse using different personas. It will also leverage its image generation model to let users modify images and create stickers via text prompts.
META released MusicGen, a controllable music generation model for producing high-quality music. MusicGen can be prompted by both text and melody.
The best thing is anyone can try it for free now. It uses a single-stage transformer language model with efficient token interleaving patterns, eliminating the need for multiple models.
MusicGen will generate 12 seconds of audio based on the description provided. You can optionally provide a reference audio from which a broad melody will be extracted. Then the model will try to follow both the description and melody provided. You can also use your own GPU or a Google Colab by following the instructions on their repo.
Meta’s new human-like AI model for image creation
Meta has introduced a new model, Image Joint Embedding Predictive Architecture (I-JEPA), based on Meta’s Chief AI Scientist Yann LeCun’s vision to make AI systems learn and reason like animals and humans. It is a self-supervised computer vision model that learns to understand the world by predicting it.
The core idea: It learns by creating an internal model of the outside world and comparing abstract representations of images. It uses background knowledge about the world to fill in missing pieces of images, rather than looking only at nearby pixels like other generative AI models.
Captures patterns and structures through self-supervised learning from unlabeled data.
Predicts missing information at a high level of abstraction, avoiding generative model limitations
Delivers strong performance on multiple computer vision tasks while also being computationally efficient. Less data, less time, and less compute.
Can be used for many different applications without needing extensive fine-tuning and is highly scalable.
Meta’s all-in-one generative speech AI model
Meta introduces Voicebox, the first generative AI model that can perform various speech-generation tasks it was not specifically trained to accomplish with SoTA performance. It can perform:
Text-to-speech synthesis in 6 languages
Noise removal
Content editing
Cross-lingual style transfer
Diverse sample generation
One of the main limitations of existing speech synthesizers is that they can only be trained on data that has been prepared expressly for that task. Voicebox is built upon the Flow Matching model, which is Meta’s latest advancement on non-autoregressive generative models that can learn highly non-deterministic mapping between text and speech.
Meta disclosed AI behind Facebook and Instagram recommendations
Meta is sharing 22 system cards that explain how AI-powered recommender systems work across Facebook and Instagram. These cards contain information and actionable insights everyone can use to understand and customize their specific AI-powered experiences in Meta’s products.
Moreover, Meta also shared its top ten most important prediction models rather than everything in the system to not dive into much technical detail can sometimes obfuscate transparency.
Using an input audio sample of just two seconds in length, Voicebox can match the sample’s audio style and use it for text-to-speech generation.
Meta plans to dethrone OpenAI and Google
Meta plans to release a commercial AI model to compete with OpenAI, Microsoft, and Google. The model will generate language, code, and images. It might be an updated version of Meta’s LLaMA, which is currently only available under a research license.
Meta’s CEO, Mark Zuckerberg, has expressed the company’s intention to use the model for its own services and make it available to external parties. Safety is a significant focus. The new model will be open source, but Meta may reserve the right to license it commercially and provide additional services for fine-tuning with proprietary data.
Tesla is launching its highly-anticipated supercomputer today. The machine, employing 10,000 Nvidia H100 compute GPUs, will be used for various AI applications. It is said to be one of the most powerful machines in the world.
But NVIDIA is struggling to keep up with the GPU demand. Thus, Tesla is investing over $1B to develop its own supercomputer, Dojo, built on the company’s hyper-optimized custom-designed chip. Tesla is also activating Dojo simultaneously. Take a look at Tesla’s internal forecast for the compute power of Dojo.
Why does it matter?
Elon Musk recently revealed that Tesla plans to spend over $2B on AI training in 2023 and is hiring reputed AI engineers. But this move gives Tesla unparalleled compute power. It also underscores Tesla’s commitment to overcoming computational bottlenecks in AI and should provide substantial advantages over its rivals. Elon might be the next big thing in AI. What do you think?
OpenAI launches ChatGPT Enterprise, the most powerful ChatGPT version yet
Open has launched ChatGPT Enterprise, the most powerful version of ChatGPT yet. It offers enterprise-grade security and privacy, features for large-scale deployments, unlimited higher-speed GPT-4 access, 32K context for faster processing of longer inputs, advanced data analysis capabilities, customization options, and much more. OpenAI is also working on more features and will launch them soon.
Why does it matter?
This is a simple and safe way of deploying ChatGPT into core operations at organizations. It could be a solution for big companies that have banned ChatGPT at work over privacy concerns, like Apple, Amazon, Citigroup, and more. Maybe, this can pave the way for truly widespread adoption of AI in the business world.
Usage of ChatGPT among Americans rises, but only slightly
A recent survey conducted in July by Pew Research Center reveals 18% of U.S. adults have ever used ChatGPT. While 16% of those who have heard of the tool and are employed say they have used it for tasks at work.
The statistic is consistent with a similar survey conducted in March by the Pew Research Center that showed 14% of U.S. adults had tried ChatGPT. And about one in ten working adults who had heard of ChatGPT used it at work.
While this shows increased adoption of ChatGPT among Americans, it is not a significant one in the grand scheme of AI adoption today. In fact, only a few think it will have a major impact on their job.
Why does this matter?
These findings suggest AI’s penetration remains gradual. It is also clear that there is still work to be done in educating and acclimating the workforce to the benefits and implications of generative AI. Plus, given the lingering concerns and uncertainties about ChatGPT’s prowess, maybe it is too early to start worrying about AI replacing jobs.
What Else Is Happening in AI
Microsoft infuses AI with human-like reasoning via an “Algorithm of Thoughts”.
DoorDash launches AI-powered voice ordering to answer calls and curate recommendations.
Uber is working on an AI chatbot for its food delivery app.
Yahoo Mail introduces new AI-powered capabilities, including a ‘Shopping Saver’ tool.
Generative inbreeding, akin to inbreeding in genetics, is a concern as AI systems training on AI-generated content can degrade their performance and distort human culture.
Tesla’s $300M AI cluster is going live today – Tesla is launching its highly-anticipated supercomputer today. The machine, employing 10,000 Nvidia H100 compute GPUs, will be used for various AI applications. – But NVIDIA is struggling to keep up with the GPU demand. Thus, Tesla is investing over $1B to develop its own supercomputer, Dojo, built on the company’s hyper-optimized custom-designed chip. Tesla is also activating Dojo simultaneously.
OpenAI launches ChatGPT Enterprise, the most powerful ChatGPT version yet – It offers enterprise-grade security and privacy, features for large-scale deployments, unlimited higher-speed GPT-4 access, 32K context for faster processing of longer inputs, advanced data analysis capabilities, customization options, and much more. OpenAI is also working on more features and will launch them soon.
Usage of ChatGPT among Americans rises, but only slightly – A recent survey conducted in July by Pew Research Center reveals 18% of U.S. adults have ever used ChatGPT. While 16% of those who have heard of the tool and are employed say they have used it for tasks at work. The statistic is consistent with a similar survey conducted in March by the center. – While it shows increased adoption of ChatGPT among Americans, it is not a significant one in the grand scheme of AI adoption today. In fact, only a few think it will have a major impact on their job.
Microsoft infuses AI with human-like reasoning via an “Algorithm of Thoughts” – The technique guides the language model through a more streamlined problem-solving path. It utilizes in-context learning, enabling the model to explore different solutions in an organized manner systematically. The result? Faster, less resource-intensive problem-solving.
DoorDash launches AI-powered voice ordering service – It will answer calls and provide customers with curated recommendations.
Uber is working on an AI chatbot for its food delivery app – It will offer recommendations to food-delivery customers and help them more quickly place orders.
Yahoo Mail introduces new AI-powered capabilities – The rollout includes upgrades to several of Yahoo Mail’s existing AI features and introduces a new Shopping Saver tool.
Poe by Quora lets you use all the AI chatbots in one place – Its goal is to be the web browser for accessing AI chatbots, and it just got a bunch of updates.
IBM’s new analog AI chip challenges Nvidia
IBM has developed an analog AI chip that’s up to 14 times more energy-efficient than current digital chips, addressing the power-hungry nature of generative AI.
The analog chip’s ability to manipulate analog signals and its human brain-like operation could potentially challenge Nvidia’s dominance in AI hardware.
IBM’s prototype chip demonstrated significant energy efficiency gains, encoding millions of memory devices and modeling parameters while performing computations directly within memory.
AI’s promise and peril in cancer research
UK-based biotech startup Etcembly used generative AI to develop a novel immunotherapy targeting hard-to-treat cancers, demonstrating AI’s potential for medical advancements.
However, risks of AI in healthcare are evident, as a study reveals that AI-generated cancer treatment plans, like those from ChatGPT, contained factual errors and contradictory information.
While AI-powered tools hold promise, their clinical deployment without rigorous validation could lead to dangerous missteps, highlighting the importance of skepticism and human consultation.
Linkedin: Building soft (human) skills remains key in the age of AI
Summary: A new LinkedIn report reveals that AI skills are spreading quickly globally, with major growth in AI job postings and professionals adding AI abilities.
Job postings mentioning AI skills like GPT and ChatGPT have risen dramatically, with a 21x increase since November 2022.
LinkedIn members adding AI skills to profiles is accelerating globally. The number of members with AI skills was 9x larger in June 2023 compared to January 2016.
Singapore, Finland, Ireland, India and Canada have the fastest AI skills adoption rates based on LinkedIn’s AI Skills Index.
47% of US executives believe using generative AI will boost productivity. 40% think it will help drive revenue growth.
84% of US members have jobs that could use AI to automate at least 25% of repetitive tasks. This will also increase demand for people skills.
In the US, the fastest-growing in-demand skills since November 2022 are: Flexibility +158%, Professional ethics +120%, Social perceptiveness +118%, Self-management +83%.
Communication remains the top skill in demand in US job postings, with people skills like flexibility growing the fastest since ChatGPT launched.
92% of executives agree people skills are more important than ever in an AI-driven world.
Why It Matters: AI is transforming and disrupting every industry for sure, but it will never disrupt humanity. Human skills (also called soft skills) like creativity and emotional intelligence will only become more important.
YouTube and Universal Music Partner to Launch ‘AI Incubator’
YouTube is partnering with Universal Music to launch an incubator focused on exploring the use of AI in music. The incubator will work with artists and musicians, including Anitta, ABBA’s Björn Ulvaeus, and Max Ricther, to gather insights on generative AI experiments and research. YouTube CEO Neal Mohan stated that the incubator will inform the company’s approach as it collaborates with innovative artists, songwriters, and producers.
YouTube also plans to invest in AI-powered technology, including enhancing its copyright management tool, Content ID, to protect viewers and creators.
Why does this matter?
By partnering with renowned artists, the AI incubator explores the potential of AI-generated music, spotlighting the intersection of technology and artistry. This collab not only underscores AI’s growing role in creative industries but also demonstrates how industry giants can collaborate to drive innovation and shape the future of music production.
In the ever-evolving landscape of artificial intelligence, Large Language models (LLMs) like GPT-3/GPT-4/Claude-2 and others have exhibited astonishing capabilities across various domains, from mathematical problem-solving to creative writing. However, there’s been a limitation in their approach – the left-to-right, token-by-token decision-making process, which doesn’t always align with complex problem-solving scenarios that demand strategic planning and exploration.
But what if we could enable these LLMs to think more strategically, explore multiple reasoning paths, and evaluate the quality of their thoughts in a deliberate manner? Some researchers have created a framework called “Tree of Thoughts” (ToT) which aims to fix this by enhancing the problem-solving prowess of large language models.
The Essence of ToT
At its core, ToT reimagines the reasoning process as an intricate tree structure. Each branch of this tree represents an intermediate “thought” or a coherent chunk of text that serves as a crucial step toward reaching a solution. Think of it as a roadmap where each stop is a meaningful milestone in the journey towards problem resolution. For instance, in mathematical problem-solving, these thoughts could correspond to equations or strategies.
But ToT doesn’t stop there. It actively encourages the LM to generate multiple possible thoughts at each juncture, rather than sticking to a single sequential thought generation process, as seen in traditional chain-of-thought prompting. This flexibility allows the model to explore diverse reasoning paths and consider various options simultaneously.
Source: Yao et el. (2023)
The Power of Self-Evaluation
One of ToT’s defining features is the model’s ability to evaluate its own thoughts. It’s like having an inbuilt compass to assess the validity or likelihood of success for each thought. This self-evaluation provides a heuristic, a kind of mental scorecard, to guide the LM through its decision-making process. It helps the model distinguish between promising paths and those that may lead to dead ends.
Systematic Exploration
ToT takes strategic thinking up a notch by employing classic search algorithms such as breadth-first search or depth-first search to systematically explore the tree of thoughts. These algorithms allow the model to look ahead, backtrack when necessary, and branch out to consider different possibilities. It’s akin to a chess player contemplating multiple moves ahead before making a move.
Customizable and Adaptable
One of ToT’s strengths is its modularity. Every component, from thought representation to generation, evaluation, and search algorithm, can be customized to fit the specific problem at hand. No additional model training is needed, making it highly adaptable to various tasks.
Real-World Applications
The true litmus test for any AI framework is its practical applications. ToT has been put to the test across different challenges, including the Game of 24, Creative Writing, and Mini Crosswords. In each case, ToT significantly boosted the problem-solving capabilities of LLMs over standard prompting methods. For instance, in the Game of 24, success rates soared from a mere 4% with chain-of-thought prompting to an impressive 74% with ToT.
Source: Yao et el. (2023)
The above image is a visual representation of the Game of 24 which is a mathematical reasoning challenge where the goal is to use 4 input numbers and arithmetic operations to reach the target number 24.
The tree of thought (ToT) approach represents this as a search over possible intermediate equation “thoughts” that progressively simplify towards the final solution.
First, the language model proposes candidate thoughts that manipulate the inputs (e.g. (10 – 4)).
Next, it evaluates the promise of reaching 24 from each partial equation by estimating how close the current result is. Thoughts evaluated as impossible are pruned.
The process repeats, generating new thoughts conditioned on the remaining options, evaluating them, and pruning. This iterative search through the space of possible equations allows systematic reasoning.
For example, the model might first try (10 – 4), then build on this by proposing (6 x 13 – 9) which gets closer to 24. After several rounds of generation and evaluation, it finally produces a complete solution path like: (10 – 4) x (13 – 9) = 24.
By deliberating over multiple possible chains of reasoning, ToT allows more structured problem solving compared to solely prompting for the end solution.
Trained AI algorithms work by taking the input and providing the output without explaining its inner workings. XAI aims at pointing out the rationale behind any decision by AI in such a way that humans can interpret it.
Deep learning works with neural networks just like the human brain works with neurons, where it uses a massive amount of training data to learn and identify patterns. It would be very difficult, or rather impossible, to dig into the rationale behind Deep Learning’s decision. Decisions like credit card eligibility or loan sanction are quite important to be explained by XAI. However, a few wrong decisions would not impact much. Whereas, in the case of healthcare, as discussed earlier, a doctor could not provide the appropriate treatment without knowing the rationale behind AI’s decision. Surgery on the wrong organ could be fatal.
4 Principles of Explainable AI
The US National Institute of Standards and Technology has developed four principles as guidelines to adopt fundamental properties of Explainable Artificial Intelligence (XAI) efficiently and effectively. These principles apply individually and independently from each other and guide us to better understand the working of the AI models.
1. Explanation:
This principle obligates the AI to generate a comprehensive explanation for humans to understand the process of generating the decisions with the required evidence and reasons. The standard for this evidence and reasons is governed by the next three principles.
2. Meaningful:
This principle is satisfied when a stakeholder understands the explanation provided in the first guiding principle. The explanation should not be complex and understood by the users on a group as well as individual level.
3. Explanation Accuracy:
The accuracy at which the AI explains the complicated process of generating the output is critical. Accuracy metrics may differ for individual stakeholders in terms of their explanation. The expected accuracy is 100% for all the stakeholders to understand the logic.
4. Knowledge Limits:
The last principle of XAI explains that the model can only be operated under the special conditions it has been modeled for. It is expected to operate under its limited knowledge to avoid any sort of discrepancy or unjustified business outcomes.
How does XAI work?
These principles help us define the expected output from the XAI model and how an ideal XAI model should be. However, it doesn’t indicate how the output has been achieved. Subdividing the XAI into three categories to better understand the rationale:
1. Explainable data: What data is used to train the model? Why the particular data is selected? How much biased is the data?
2. Explainable predictions: What features did the model use that lead to the particular output?
3. Explainable algorithms: How is the model layered? How do these layers lead to the prediction?
Based on individual instances, the explainability may change. For example, the neural network can only be explained using the Expainable Data category. Research is ongoing that is focused on finding ways to explain the predictions and algorithms. At present there are two approaches:
a. Proxy Modeling:
A different model from the original is used to approximate the actual model. This may result in different outcomes from the true model outcomes, as it is just an approximation.
b. Design for Interpretability:
The actual model is designed in such a way that it is easy to understand its working. However, this increases the risk of reduced predictive power and overall accuracy of the model.
The XAI is referred to as the White Box, as it explains the rationale behind its working. However, unlike the black box, its accuracy may decrease in order to provide an explainable reason for its outcome. Decision trees, Bayesian networks, sparse linear models, and many more are used as explainable techniques. Hopefully, with the advancements in the field, new studies will come up to increase the accuracy of the explanations.
Critical Industries for XAI
XAI would be helpful in those industries where machines play a key part in decision-making. These use cases might also be useful in your industry, as the details may vary, but the core principles remain the same.
1. Healthcare in XAI
As discussed earlier, the decisions made by AI in healthcare impact humans in a very critical way. A machine with XAI would help the healthcare staff save a lot of time, which they might use to focus on treating and attending to more patients. For example, diagnosing a cancerous area and explaining the reason in a matter of time helps the doctor to provide appropriate treatment.
2. Manufacturing in XAI
In the manufacturing industry, fixing or repairing equipment often depends on personnel expertise, which may vary. To ensure a consistent repair process, XAI can help provide ways to repair a machine type with an explanation, record the feedback from the worker, and continuously learn to find the best process to be followed. The workers need to trust the decision made by the machine in order to risk working on the equipment repair, which is the reason XAI becomes useful.
3. Autonomous vehicles in XAI
A self-driving car seems great until and unless it has made a bad decision, which can be deadly. If an autonomous car faces an inevitable accident scenario, the decision it makes impacts greatly on its future use, whether it saves the driver or the pedestrians. Providing the rationale for each decision an autonomous car takes, helps to improve people’s security on the road.
Strategize Your Social Media Campaigns with ChatGPT
Try the propmpt below:
You are a social media strategist. I am launching a crowdfunding campaign for an innovative portable solar charger and need to create a buzz on social media. I need a comprehensive social media strategy that covers platform selection, content ideas, posting frequency, engagement tactics, and analytics tracking. Please provide suggestions considering the latest trends in social media marketing and the behavior of tech-savvy, environmentally-conscious consumers.
YouTube will pay artists and rights holders for AI-generated music used on the platform. This aims to balance creative innovation and fair compensation.
Unraveling August 2023: Spotlight on Generative AI, Tech, Sports and the Month’s Hottest Trends.
Welcome to the hub of the most intriguing and newsworthy trends of August 2023! In this era of rapid development, we know it’s hard to keep up with the ever-changing world of ai, technology, sports, entertainment, and global events. That’s why we’ve curated this one-stop blog post to provide a comprehensive overview of what’s making headlines and shaping conversations. From the mind-bending advancements in artificial intelligence to captivating news from the world of sports and entertainment, we’ll guide you through the highlights of the month. So sit back, get comfortable, and join us as we dive into the core of August 2023!
OpenCopilot allows you to have your own product’s AI copilot. With a few simple steps, it takes less than 5 minutes to build.
It integrates with your underlying APIs and can execute API calls whenever needed. It uses LLMs to determine if the user’s request requires calling an API endpoint. Then, it decides which endpoint to call and passes the appropriate payload based on the given API definition.
Why does this matter?
Shopify has an AI-powered sidekick, while Microsoft (Windows Copilot, Bing Copilot) and GitHub (GitHub Copilot) have copilots. The above innovation empowers every SaaS product to have its own AI copilots tailored for its unique products.
Google teaches LLMs to personalize
LLMs are already good at synthesizing text, but personalized text generation can unlock even more. New Google research has proposed an approach inspired by the practice of writing education for personalized text generation using LLMs. It has a multistage and multitask framework consisting of multiple stages: retrieval, ranking, summarization, synthesis, and generation.
In addition, they introduce a multitask setting that further helps the model improve its generation ability, which is inspired by the observation that a student’s reading proficiency and writing ability are often correlated. When evaluated on three public datasets, each covering a different and representative domain, the results showed significant improvements over various baselines.
Why does this matter?
Customizing style is essential for many domains like personal communication, dialogue, marketing copies, stories, etc., which is hard to do via pure prompt engineering or custom instructions. The research attempts to address this and highlights how we can take inspiration from how humans achieve tasks to apply it to LLMs.
Local Llama
For businesses, local LLMs offer competitive performance, cost reduction, dependability, and flexibility. This article by ScaleDown provides practical guidance on setting up and running LLMs locally using a user-friendly project.
Moreover, Llama-2 and its variants are the go-to models, and the community continually refines them. The article highlights some things to note when running Llama models locally, including memory and model loader challenges.
Why does this matter?
This helps make AI accessible to individuals and businesses while avoiding limitations and high expenses associated with commercial APIs. Locally deploying LLM also helps businesses have more over the model, customize it, integrate with existing systems, and enable full utilization of its capabilities.
AI creates lifelike 3D experiences from your phone video
Luma AI has introduced Flythroughs, an app that allows one-touch generation of photorealistic, cinematic 3D videos that look like professional drone captures. Record like you’re showing the place to a friend, and hit Generate– all on your iPhone. No need for drones, lidar, expensive real estate cameras, and a crew.
Flythroughs is built on Luma’s breakthrough NeRF and 3D generative AI and a brand new path generation model that automatically creates smooth dramatic camera moves.
Why does this matter?
This marks a significant leap in democratizing 3D content creation with AI and making it cost-efficient. It opens up new possibilities for storytelling and crafting stunning digital experiences for users across various industries.
Genetic Algorithm Optimized Neural Network Model for Malicious URL Detection
URL Genie is a web application implementing a Multilayer Perceptron Neural Network optimized using genetic algorithms. Detect whether a domain name or URL is malicious by inputting a URL.
– Boosted.ai – AI stock screening, portfolio management, risk management
– JENOVA – AI stock valuation model that uses fundamental analysis to calculate intrinsic value
– Danielfin – Rates stocks and ETFs with an easy-to-understand global AI Score
– Comparables.ai – AI designed to find comparables for market analysis quickly and intelligently
Daily AI Update News from OpenCopilot, Google, Luma AI, AI2, and more
AI Copilot for your own SaaS product – OpenCopilot allows you to have your own product’s AI copilot. It integrates with your underlying APIs and can execute API calls whenever needed. It uses LLMs to determine if the user’s request requires calling an API endpoint. Then, it decides which endpoint to call and passes the appropriate payload based on the given API definition.
Teach LLMs to Personalize – New Google research has proposed a general approach for personalized text generation using LLMs, inspired by the practice of writing education. Upon evaluation, the results showed significant improvements over a variety of baselines.
Introducing Flythroughs, an app that creates lifelike 3D experiences from your phone video – It allows one-touch generation of photorealistic, cinematic videos that look like professional drone captures. No need for drones, lidar, expensive real estate cameras, and a crew. Record like you’re showing the place to a friend, and hit Generate; all on your iPhone.
Big brands are increasingly using AI-generated ads, including Nestlé and Mondelez – More brands see generative AI as a means to make creating an ad less painful and costly. However, there are concerns over whether to let people know it’s AI-generated, whether AI ads can receive copyright protections, and security risks when using AI.
AI2 drops the biggest open dataset yet for training language models – Language models like GPT-4 and Claude are powerful and useful. Still, the data on which they are trained is a closely guarded secret. The AI2’s (Allen Institute for AI) new, huge text dataset, Dolma, is free to use and open to inspection.
Ex-Machine Zone CEO launches BeFake, an AI-based social media app – Alias Technologies has introduced BeFake, a social media app for digital self-expression. Now available on both the App Store and Google Play, it aims to offer a refreshing alternative to the conventional reality portrayed on existing social media platforms.
Some of the world’s biggest advertisers, from food giant Nestle to consumer goods multinational Unilever, are experimenting with using generative AI software like ChatGPT and DALL-E to cut costs and increase productivity.
The New York Times may sue OpenAI over its AI chatbot ChatGPT, which uses the newspaper’s stories to generate text. The paper is unhappy that OpenAI is not paying for the use of its content and is also worried that ChatGPT could reduce its online traffic by providing answers based on its reporting.
Mantella allows you to have natural conversations with NPCs in Skyrim using your voice by leveraging Whisper for speech-to-text, ChatGPT for text generation, and xVASynth for text-to-speech. NPCs also have memories of your previous conversations and have awareness of in-game events.
British Prime Minister Rishi Sunak is set to spend 100 million pounds ($130 million) to buy thousands of computer chips to power artificial intelligence amid a global shortage and race for computing power.
Imagine a world where you reside in a luxurious home, an architectural marvel adorned with every comfort and amenity that one could possibly fathom. But it doesn’t stop there; your creativity knows no bounds, and you envision entire universes with their own laws of physics, teeming with diverse civilizations.
As you journey through life, your passions take an intriguing turn, guiding you towards the realm of digital marketing.
Yet, amid this shift in interests, a captivating question continues to linger in your mind like an enigmatic riddle: “If I possessed the power to design anything in the world, what wondrous creation would spring forth from my imagination?”
As your knowledge expands and your expertise in digital marketing deepens, you become acquainted with the remarkable world of graphic design software. Herein lies the key to unlock the gateway to your wildest ideas and aspirations.
With the vast array of possibilities that graphic design software offers, you come to realize that you can bring to life virtually anything your mind can conceive – and that realization holds true for anyone daring enough to venture into this realm.
While some graphic design software tools are tailored to cater to specialized fields, such as web design software that masters the dynamic nature of webpages or CAD software that focuses on technical drawings, at its core, graphic design software is an all-encompassing and versatile tool. It empowers individuals to transform their creative visions into tangible realities.
Within the confines of this article, we shall embark on a journey exploring the finest AI design software tools currently available. These cutting-edge tools are poised to revolutionize the design process and elevate your artistic capabilities to unprecedented heights.
By leveraging the power of artificial intelligence, these tools open up new horizons, enabling you to streamline and automate your design workflow like never before.
So, fasten your seatbelts and prepare to delve into the realm of limitless creativity. In the following sections, we shall uncover the potentials of AI-driven design software and how they stand as testaments to the boundless human imagination.
It’s time to manifest your artistic dreams into reality – let the voyage commence!
When it comes to harnessing the power of AI for creating mesmerizing visual graphics, few tools can rival the prowess of Adobe Photoshop CC. Renowned across the globe, this software stands as a beacon of creativity and innovation, empowering artists, designers, and digital enthusiasts to bring their imaginations to life in the most astonishing ways.
At the heart of Adobe Photoshop CC lies an impressive array of features that cater to every aspect of design. Whether you aim to craft captivating illustrations, design stunning artworks, or manipulate photographs with unprecedented precision, this software has got you covered.
With its user-friendly interface and intuitive controls, even those new to the world of digital design can quickly find themselves delving into the realm of endless possibilities.
One of the standout strengths of Photoshop lies in its ability to produce highly realistic and detailed images. From refining minute details in portraits to creating breathtaking landscapes, the software’s tools and filters enable artists to achieve a level of precision that defies belief.
The result is a visual masterpiece that captures the essence of the creator’s vision with unparalleled fidelity.
But Photoshop is not merely limited to polishing existing images; it opens the gates to boundless creativity by allowing users to remix and combine multiple images seamlessly. Whether it’s composing fantastical scenes or crafting surreal montages, the software’s blending capabilities grant designers the freedom to construct their own visual universes.
What truly sets Adobe Photoshop CC apart from the rest is its ingenious integration of artificial intelligence. The inclusion of AI-driven features elevates the design process to a whole new dimension.
Dull and lackluster photographs transform into jaw-dropping works of art with just a few clicks, as the software’s AI algorithms intelligently enhance colors, textures, and lighting, breathing life into every pixel.
Adobe’s suite of creative tools, including the likes of Adobe Illustrator and others, work in seamless harmony with Photoshop. This synergy empowers designers to amplify their creative potential even further.
Whether you’re crafting a logo, designing a website, or creating intricate vector graphics, the integration of these tools allows you to transcend the boundaries of imagination.
Planner 5D stands as an ingenious AI-powered solution, offering you the gateway to realize your long-cherished dream of a perfect home or office space. With its cutting-edge technology, this software empowers you to dive into the realm of architectural creativity and interior design like never before.
The first remarkable feature that sets Planner 5D apart is its AI-assisted design capabilities. Imagine describing your ideal home or office, and watch as the AI effortlessly translates your vision into a stunning 3D representation. From grand entrances to cozy corners, the AI understands your preferences, ensuring that every aspect of your dream space aligns with your desires.
Gone are the days of struggling with pen and paper to create floor plans. Planner 5D streamlines the process, enabling you to effortlessly design detailed and precise floor plans for your dream space.
Whether you seek an open-concept layout or a series of interconnected rooms, this software provides the tools to bring your architectural visions to life.
But that’s not all – Planner 5D goes above and beyond to cater to every facet of interior design. With an extensive library of furniture and home décor items at your disposal, you can furnish and decorate your space with ease.
From stylish sofas and elegant dining tables to enchanting wall art and lighting fixtures, the possibilities are limitless.
The user-friendly 2D/3D design tool within Planner 5D is a testament to the software’s commitment to simplicity and innovation. Whether you’re an aspiring designer or a seasoned professional, navigating through the interface is a breeze, allowing you to create the perfect space for yourself, your family, or your business with utmost ease and precision.
For those seeking a more hands-off approach, Planner 5D also offers the option to hire a professional designer through their platform. This feature is a boon for individuals who desire a polished and expertly curated space but prefer to leave the intricate details to the experts.
By collaborating with skilled designers, you can rest assured that your dream home or office will become a reality, tailored to your unique taste and requirements.
Uizard emerges as a game-changing tool that holds the power to transform the creative process for founders and designers alike. This innovative software enables you to breathe life into your ideas by swiftly converting your initial sketches into high-fidelity wireframes and stunning UI designs.
Gone are the days of spending endless hours painstakingly crafting wireframes and prototypes manually. With Uizard, the transformation from a low-fidelity sketch to a polished, high-fidelity wireframe or UI design can occur within mere minutes.
The speed and efficiency afforded by this cutting-edge technology empower you to focus on refining your concepts and iterating through ideas at an unprecedented pace.
Whether your vision encompasses web apps, websites, mobile apps, or any digital platform, Uizard stands as a reliable companion, streamlining the design process with its versatility. You no longer need to possess extensive design expertise, as the tool is intuitively designed to cater to users of all backgrounds and skill levels.
From tech-savvy founders to aspiring entrepreneurs, Uizard ensures that the creative journey remains accessible and enjoyable for everyone.
The user-friendly interface of Uizard opens up a realm of possibilities, allowing you to bring your vision to life with ease. Its intuitive controls and extensive feature set empower you to craft pixel-perfect designs that align with your unique style and brand identity.
Whether you’re a solo founder or part of a dynamic team, Uizard fosters seamless collaboration, enabling you to share and iterate on designs effortlessly.
One of the most significant advantages of Uizard lies in its ability to gather invaluable user feedback on your designs. By sharing your wireframes and UI designs with stakeholders, clients, or potential users, you can gain insights and refine your creations based on real-world perspectives.
This not only accelerates the decision-making process but also ensures that your final product resonates with your target audience.
Enter the extraordinary realm of 3D animation with Autodesk Maya, a software that transcends conventional boundaries to grant you the power to breathe life into expansive worlds and intricate characters. Whether you’re an aspiring animator, a seasoned professional, or a visionary storyteller, Maya provides the tools to transform your creative visions into stunning reality.
Imagination knows no bounds with Maya, as its powerful toolsets empower you to embark on a journey of endless possibilities. From the grandest of cinematic tales to the most whimsical of animated adventures, this software serves as your creative canvas, waiting for your artistic touch to shape it.
Complexity is no match for Maya’s prowess, as it deftly handles characters and environments of any intricacy. Whether you seek to create lifelike characters with nuanced emotions or craft breathtaking landscapes that transcend the boundaries of reality, Maya’s capabilities rise to the occasion, ensuring that your artistic endeavors know no limits.
Designed to cater to professionals across various industries, Maya stands as the perfect companion for crafting high-quality 3D animations for movies, games, and an array of other purposes. Its versatility makes it a go-to choice for animators, game developers, architects, and designers alike, unleashing the potential to tell stories and visualize concepts with stunning visual fidelity.
The heart of Maya lies in its engaging animation toolsets, each one carefully crafted to nurture the growth of your virtual world. From fluid character movements to dynamic environmental effects, Maya opens the doors to your creative sanctuary, enabling you to weave intricate tales that captivate audiences across the globe.
But the journey doesn’t end there – with Autodesk Maya, you are the architect of your digital destiny. As you explore the depths of this software, you discover its seamless integration with other creative tools, expanding your capabilities even further.
The synergy between Maya and its counterparts unlocks new avenues for innovation, granting you the freedom to experiment, iterate, and refine your creations with ease.
Aimed at architects, engineers, and a myriad of other professionals, this cutting-edge tool serves as a gateway to bring your imaginative designs to life with astonishing realism.
Architects find solace in Autodesk 3Ds Max as it empowers them to create stunningly realistic models of their architectural wonders. Engineers, too, discover the power of this software in crafting intricate and precise 3D models of their mechanical and industrial designs.
The software becomes a haven for creative professionals seeking to visualize and communicate their concepts with exceptional clarity and visual fidelity.
Beyond the realms of architecture and engineering, Autodesk 3Ds Max knows no bounds. Its versatility allows you to explore various dimensions of creativity, from crafting intricate product prototypes to fashioning enchanting animated characters.
Whatever your design aspirations may be, this software stands as a reliable companion, ensuring that your visions manifest into awe-inspiring digital realities.
In the fast-paced world of business and design, having a tool that caters to multiple purposes becomes invaluable. Autodesk 3Ds Max stands tall as a versatile and adaptable solution, making it a coveted asset for businesses and individuals alike.
Its potential to enhance the visual storytelling capabilities of any venture unlocks a new era of creativity and communication.
One of the most cherished qualities of Autodesk 3Ds Max lies in its user-friendly interface, fostering a seamless and intuitive design process. With this tool at your disposal, iteration becomes a breeze, allowing you to refine your designs effortlessly.
Each new iteration becomes a steppingstone towards perfection, ensuring that your final creation exudes excellence.
With Foyr Neo at your disposal, you can witness the transformation of your design ideas into reality in as little as a fifth of the time it takes with other software tools.
Gone are the days of grappling with complex design interfaces and spending endless hours on a single project. Foyr Neo streamlines the journey from a floor plan to a finished render, presenting you with a user-friendly interface that simplifies every step of the design process.
With its intuitive controls and seamless functionality, the software becomes an extension of your creative vision, ensuring that your ideas manifest into remarkable designs with utmost ease.
To further elevate your experience, Foyr Neo provides a thriving community and comprehensive training resources. This collaborative ecosystem allows you to connect with fellow designers, share insights, and gain inspiration from the collective creative pool.
Additionally, the abundance of training materials and support ensures that you can unlock the full potential of the software, mastering its capabilities and expanding your design horizons.
Bid farewell to the hassle of juggling multiple tools to complete a single project – Foyr Neo serves as the all-in-one solution to cater to your design needs. By integrating various design functionalities within a single platform, the software streamlines your workflow, saving you precious time and effort.
This seamless experience fosters uninterrupted creativity, enabling you to focus on the art of design without the burden of managing disparate software tools.
With this cutting-edge software, you can witness a remarkable increase in image resolution of up to 16 times, all without sacrificing an ounce of quality.
Gone are the days of tedious manual editing, spending hours painstakingly enhancing your images pixel by pixel. Let’s Enhance simplifies the process, offering a swift and efficient solution to elevate your photos’ quality with ease.
Whether you’re a professional photographer seeking crisper images for print or a social media enthusiast aiming to enlarge your visuals, this software promises to deliver the perfect shot every time.
Let’s Enhance’s proficiency in improving image resolution, colors, and lighting automatically alleviates the burden of post-processing. By entrusting this task to the intelligent algorithms of the software, you gain more time to focus on the core aspects of your business or creative endeavors.
Embrace the art of delegation and allow Let’s Enhance to handle the technicalities while you concentrate on perfecting your artistic vision.
The applications of Let’s Enhance are vast and diverse. Photographers, designers, artists, and marketers alike can benefit from this powerful tool. Imagine effortlessly preparing your images for print, knowing they’ll boast impeccable clarity and sharpness.
Envision your social media posts grabbing attention with larger-than-life visuals, thanks to Let’s Enhance’s seamless enlargement capabilities.
But Let’s Enhance doesn’t stop at resolution enhancement. It also becomes a reliable ally in refining color palettes, breathing new life into dull or faded images, and balancing lighting for picture-perfect results.
Whether it’s subtle adjustments or dramatic transformations, the software empowers you to create visuals that captivate audiences and leave a lasting impression.
6 AI Text to Video compared (updated August 2023 ) Link
Runway Features
– Text-to-video feature – Automatic prompt suggestions – The option to upload an image for reference – Different previews to choose from before generating a video – Free plan to test the tool out
Pros
– Best of AI text-to-video research – Comprehensive set of tools for video editing – Available as both a desktop and mobile app
Cons
– Gen-2 has limitations in generating intricate details, like fingers – Gen-2 video generation is limited to 4 seconds per video – The tool does not offer text-to-speech capabilities
Synthesia AI Features
– 120+ voices and accents – 140+ diverse AI avatars – 60+ video templates designed by professional designers – The option to have a custom avatar created
In today’s world, messaging apps are becoming increasingly popular, with WhatsApp being one of the most widely used. With the help of artificial intelligence, chatbots have become an essential tool for businesses to improve their customer service experience. Chatbot integration with WhatsApp has become a necessity for businesses that want to provide a seamless and efficient customer experience. ChatGPT is one of the popular chatbots that can be integrated with WhatsApp for this purpose. In this blog post, we will discuss how to integrate ChatGPT with WhatsApp and how this chatbot integration with WhatsApp can benefit your business.
The site uses openAI to generate trivia on anything and everything you want ! You can then revisit trivia you or others have made and replay them at anytime.
Solo & real time multiplayer, daily challenge, infinite playability and is getting updates daily !
Current feature roadmap :
jeopardy mode ( multiple topics and large question count )
email / sms notifications for new daily challenges etc.
public lobbies / multiplayer against random players
IBM’s study indicates that 40% of the global workforce, or 1.4 billion people, will need to reskill in the next three years due to AI’s rise.
While AI technologies, such as generative models, might shift job responsibilities, 87% of surveyed executives believe AI will augment jobs rather than replace them.
The focus in job skills has shifted from technical STEM skills (most important in 2016) to people skills like team management and adaptability (most important in 2023).
Meta did it first… Generative AI for producers
Generative AI is revolutionizing this decade’s technology, breaking into the realm of creativity once reserved for humans. Jobs are shifting, with some roles being replaced and others benefiting from AI assistance.
Content creators, take note! Meta just revealed that platforms like Facebook and Instagram will employ AI to produce music. This means no more copyright issues or losing business. Simply choose a genre, provide a sample, and the AI crafts tailor-made music for your videos.
Facebook’s music library becomes obsolete as Meta leads the way, while YouTube and TikTok will likely follow suit. As a content creator, AI eliminates rights concerns. However, creators of original music may face challenges.
AI’s impact extends to various fields, affecting writers, musicians, artists, and photographers. While some might feel the pinch, the creative economy as a whole benefits, making custom content creation easier.
Imagine conceiving, designing, and animating with AI—a reality that even big players like Disney face. This emerging world is thrilling and transformative.
To prepare, embrace AI. Integrate it into your work wherever possible. If you want to stay ahead and not fall behind to AI, leverage its capabilities.
Trustworthy LLMs: A survey and guideline for evaluating LLMs’ alignment
Ensuring alignment, which refers to making models behave in accordance with human intentions, has become a critical task before deploying LLMs in real-world applications. This new research has proposed a more fine-grained taxonomy of LLM alignment requirements. It not only helps practitioners unpack and understand the dimensions of alignments but also provides actionable guidelines for data collection efforts to develop desirable alignment processes.
It also thoroughly surveys the categories of LLMs that are likely to be crucial to improve their trustworthiness and shows how to build evaluation datasets for alignment accordingly.
The tool curates high-quality data that leads to improved LLM downstream performance and will significantly benefit LLM developers attempting to build pretraining datasets.
Microsoft-DataBricks collab may hurt OpenAI
Microsoft is reportedly planning to sell a new version of Databricks software, It helps customers create AI applications for their businesses. This move could potentially harm OpenAI, as Databricks allows companies to develop AI models from scratch or repurpose open-source models instead of licensing OpenAI’s proprietary ones.
Microsoft has been aggressively investing in AI services and integrating AI functionality across its products. Neither Microsoft nor Databricks have commented on the report.
What else happened in AI this week of August 12-20?
Google appears to be readying new AI-powered tools for ChromeOS
Zoom rewrites policies to make clear user videos aren’t used to train AI
Anthropic raises $100M in funding from Korean telco giant SK Telecom
Modular, AI startup challenging Nvidia, discusses funding at $600M valuation
California turns to AI to spot wildfires, feeding on video from 1,000+ cameras
FEC to regulate AI deepfakes in political ads ahead of 2024 election
Google’s AI search offers AI-generated summaries, definitions, and coding improvements.
Google Photos introduce a new AI feature, ‘Memories view’!
Amazon using AI to enhance product reviews.
WhatsApp test beta upgrade with new feature ‘custom AI-generated stickers’.
Google is testing an AI assistant that will give you Life Advice.
Robomart adopts “store-hailing” for self-driving stores delivered to customers.
OpenAI acquires AI design studio Global Illumination to work on core products, ChatGPT
The Associated Press releases guidelines for Generative AI to its journalists
Consulting giant McKinsey unveils its own generative AI tool for employees: Lilli
Opera for iOS will now include Aria, its browser AI built in collaboration with OpenAI
UK is using AI road safety cameras to detect potential driver offenses in passing vehicles
Adobe Express with AI Firefly app, now out of beta, is available worldwide
Ex-Google Brain researchers have started an AI research company called Sakana AI in Tokyo.
Runway, a gen AI video startup, has launched a new ‘Watch’ feature.
Research shows AI bots beat CAPTCHA and humans.
ML startup Arthur launched an open-source tool to help find the best LLM.
Buildbox has launched a new tool called StoryGames.AI!
Latest Tech News and Trends on August 20th, 2023
Major concerns after Cruise robotaxi incidents
Following a recent collision between a Cruise robotaxi and a fire truck in San Francisco, the California DMV requested Cruise to halve its robotaxi fleet in the city.
The state agency is investigating “recent concerning incidents” with Cruise vehicles, emphasizing the need to ensure the safety of the public sharing the road with these autonomous vehicles.
This specific accident saw a Cruise Chevy Bolt EV hit by an emergency vehicle at an intersection, resulting in passenger injuries; it adds to a series of issues potentially affecting Cruise’s future operations.
As wildfires spread, Canadian leaders ask Meta to reverse its news ban
The Canadian government demands that Meta lift its ban on domestic news sharing, citing its impact on sharing information about wildfires.
Meta blocked news on Facebook and Instagram due to a new law requiring payment for news articles, but this move hampers access to crucial information.
Officials and citizens express concerns, urging Meta to reinstate news sharing for safety and emergency information during the wildfire crisis.
X to remove ‘block’ feature
Elon Musk suggests that Twitter’s block feature, except for direct messages, may be removed, causing concern among users.
Blocking is currently used to restrict interactions and visibility of accounts, while mute only hides posts; users value blocking for spam control and harassment prevention.
Musk’s statement prompts backlash and uncertainty about whether the feature will actually be removed.
Discover the OpenAI code interpreter, an AI tool that translates human language into code. Learn about its functions, benefits and drawbacks in this guide.
The basics of OpenAI code interpreter
OpenAI, a leading entity in the field of artificial intelligence, has developed OpenAI code interpreter, a specialized model trained on extensive data sets to process and generate programming code.
OpenAi code interpreter is a tool that attempts to bridge the gap between human language and computer code, offering myriad applications and benefits. It represents a significant step forward in AI capabilities. It is grounded in advanced machine learning techniques, combining the strengths of both unsupervised and supervised learning. The result is a model that can understand complex programming concepts, interpret various coding languages, and generate human-like responses that align with coding practices.
New Generations of People Are Becoming More and More Indistinguishable from AI
One of the most concerning aspects of this trend is the way that new generations are rewriting previous information. In the past, people would typically come up with their ideas and opinions. However, today, it is much more common for people to simply rewrite information that they have found online. This is a trend that is being exacerbated by the rise of large language models (LLMs), which can generate text that is nearly indistinguishable from human-written text. Article: new-generations-of-people-are-becoming-more-and-more-indistinguishable-from-ai/
Neolithics, an agritech company based in Israel, is using artificial intelligence and machine learning to reduce food waste and ensure food safety and quality through its optical sensing AI-powered solution known as Crystal.eye™. This technology, which can be mounted and configured in various ways, automates and upgrades quality control for fresh produce, in order to maximize utilization and reduce waste.
While the normal spectrum of visible light has 3 colors – red, green, and blue, Crystal.eye™ uses hyperspectral imaging, with over 400 spectra of light. This light can penetrate deep into a fruit or vegetable and allows the device to scan even inside the sample, eliminating the need to cut it open or grind it.
The images produce a unique fingerprint, which is then analyzed by Neolithics’ food scientists to identify various characteristics, such as firmness, moisture content, sugar content, acidity, and many more. The data is then fed to an AI machine learning engine, allowing the system to scan and analyze a large batch of samples in a matter of seconds.
The outcomes of the inspections are then instantly displayed on a digital dashboard and can be delivered as reports, tailored to each customer’s unique requirements. For example, french fry makers need to know how much dry matter is contained in the potatoes they process, while winemakers take into account the grapes’ acidity and sweetness to obtain the flavor profile they desire.
Using Crystal.eye™ allows growers and distributors to greatly expand their sampling, from the usual 1% to around 30% to 40%. This ensures greater accuracy and significantly reduces the chance of produce being discarded due to not meeting the customers’ requirements.
According to Wayne Nathanson, the company’s VP for Global Development, knowledge in food science is Neolithics’ main differentiator. While there are other companies that make the hardware to move around and sort fruits and vegetables, he says that usually these technologies work on exterior qualities, and aren’t able to analyze the produce’s interior. Most companies do not have a team of expert food scientists to fully harness the information gathered from the produce like Neolithics, he adds.
Currently, Crystal.eye™ can check the content or defects of produce, providing customers with various external or internal attributes. This solution has been launched and is being used by an increasing number of growers, distributors, and food processing companies. At the end of this year, Neolithics expects to update the technology with the capability to assess the produce’s maturity cycle, allowing customers to identify how long it will take before it spoils. The company is also working on being able to identify traces of pesticides and other banned chemicals on the produce, with release estimated for next year.
“Sustainability is very important to Neolithics, and our mission is to reduce food waste and improve food safety. Knowing how much food is wasted daily is a major motivator for making a difference. We want to eliminate food wastage across the supply chain, including removing the need to destroy the produce when it’s being inspected. We also want to get more edible quality produce to the consumer, by helping the various links of the supply chain distribute it better. There are 1.3 billion tons of wasted food annually, and there are roughly a billion people in the world experiencing hunger. We believe there’s an opportunity to feed more people with the food that is thrown out. This becomes more and more critical, the closer the world population gets to the 10 billion mark,” Nathanson says.
The new AI programming jobs that require only very basic programming skills
There has never been a more exciting and promising time to get into AI development. Forbes reports that job listings for ChatGPT-related positions increased 21 times since last November:
They need both prompt engineers and programmers. But because of Copilot and other advances in AI programming they are looking for people with some basic programming skills but who mainly excel in advanced critical analysis and reasoning skills.
They basically need people who know how to think so for people with IQs above 130, (in the genius range) this could be a dream career. But really it’s not so much about IQ as it is about the ability to think rather than just mostly learn and remember. In fact programming courses must already be teaching this brand new kind of prompt engineering and programming.
I imagine that computer programming instruction is going through very rapid evolution right now as teaching fundamental programming skills more and more gives way to teaching how to most quickly and intelligently prompt AIs to do whatever programming is needed.
If incumbent programming schools are not changing fast enough they risk losing a substantial market share to startups that begin teaching much more marketable skills.
Many businesses today want to start using AIs but they don’t know how to go about it. Computer programmers and prompt engineers who can explain all of this to them have a ready and rapidly growing job market.
Yeah there could never be a better time to get into computer programming!
The importance of making superintelligent small LLMs
Google’s Gemini will set a new standard in AI largely because of the massive data set that it is trained on.
If you’re not familiar with Gemini yet, watch this amazingly intelligent 8-minute YouTube video:
The next step would be for Google to train that stronger intelligence to shift from relying on data to relying on principles for its logic and reasoning.
Once AI’s intelligence is based on principles, subsequent iterations will no longer require massive data for their training.
That achievement will level the playing field so that Gemini is much sooner joined by competitive or stronger models.
Once that happens, everything will get very intelligent.
As Hollywood strikes, 96% of entertainment companies are boosting generative AI spend
As the Hollywood strike continues, 96% of entertainment companies are ramping up their investments in generative AI, revealing a shift in the industry’s approach to content creation and potential concerns for its workforce.
If you want to stay ahead of the curve in AI and tech, look here first.
The rise in AI spending amidst the Hollywood strike
The Hollywood writer’s strike underscores a shift in the entertainment industry’s investment strategy.
Lucidworks’ research, one of the largest of its kind, shows 96% of executives prioritize generative AI investments.
Countries like China, the UK, France, India, and the U.S. have companies heavily investing in this technology.
AI’s potential impact on Hollywood content creation
Generative AI can produce content, virtual environments, and images, posing a potential disruption to traditional methods.
Predictions suggest that by 2025, up to 90% of Hollywood content could be influenced by AI.
There’s a growing concern among Hollywood writers about the rapid integration of AI and its effect on their careers.
The future of the entertainment industry with generative AI
The emergence of synthetic actors could revolutionize the way movies and shows are produced.
AI-driven actors don’t strike, age, or demand pay raises, presenting potential benefits for studios but challenges for human actors.
Microsoft is reportedly planning to sell a new version of Databricks software, It helps customers create AI applications for their businesses. This move could potentially harm OpenAI, as Databricks allows companies to develop AI models from scratch or repurpose open-source models instead of licensing OpenAI’s proprietary ones.
Microsoft has been aggressively investing in AI services and integrating AI functionality across its products. Neither Microsoft nor Databricks have commented on the report.
Why does this matter?
Microsoft’s reported intention to introduce an AI-focused Databricks software version carries implications for OpenAI. This software empowers businesses to craft AI solutions without relying on OpenAI’s proprietary models, potentially impacting OpenAI’s market.
Meta AI’s new RoboAgent with 12 skills
Meta and CMU Robotics Institute’s New Robotics research: RoboAgent. It is a universal robotic agent that can efficiently learn and generalize a wide range of non-trivial manipulation skills. It can perform 12 skills across 38 tasks, including object manipulation and re-orientation, and adapt to unseen scenarios involving different objects and environments.
The development of the RoboAgent was made possible through a distributed robotics infrastructure, a unified framework for robot learning, and a high-quality dataset. The agent also utilizes a language-conditioned multi-task imitation learning framework to enhance its capabilities. Meta is open-sourcing RoboSet, a large, high-quality robotics dataset collected with commodity hardware, to support and accelerate open-source research in robot learning.
Why does this matter?
RoboAgent has the potential to accelerate automation, manufacturing, and daily tasks as the end users can enjoy more capable and helpful robots at home. Industries can streamline operations with efficient automation, technology could push AI and robotics boundaries, and innovation might surge across sectors.
Meta challenges OpenAI with code-gen free software
Meta is set to release Code Llama, an open-source code-generating AI model that competes with OpenAI’s Codex. The software builds on Meta’s Llama 2 model and allows developers to automatically generate programming code and develop AI assistants that suggest code.
Llama 2 disrupted the AI industry by enabling companies to create AI apps without relying on proprietary software from major players like OpenAI, Google, or Microsoft. Code Llama is expected to launch next week, further challenging the dominance of existing code-generating AI models in the market.
Why does this matter?
Meta’s Code Llama is set to rival OpenAI’s Codex; this open-source AI model is an update of Meta’s Llama 2. This tool challenges giants like OpenAI, Google, and Microsoft, giving developers more control and reducing dependence on their proprietary tools.
AP sets new AI guidelines for newsrooms
The Associated Press has established standards for the use of generative AI in its newsroom, emphasizing that AI is not a replacement for human journalists and cautioning against creating publishable content with AI-generated text or images.
AP journalists are directed to treat AI-generated content as “unvetted source material” and apply editorial judgment and sourcing standards before considering it for publication.
The organization warns about the potential for AI to spread misinformation and advises its journalists to exercise caution, skepticism, and verify sources when dealing with AI-generated content.
Latest Tech News and Trends on August 18th, 2023
Scientists are leaving X
A significant portion of scientific researchers using X have reduced their usage or left the platform altogether, with over 47% decreasing usage and nearly 7% quitting, according to a survey by Nature.
About 47% of polled researchers have turned to alternative platforms, with Mastodon being the most popular, followed by LinkedIn and Instagram.
The change in researcher behavior on X is attributed to the platform’s evolving dynamics, increased content prioritization, and limited accessibility of its API for researchers.
Amazon imposes fees on self-shipping sellers
Starting from October 1st, third-party merchants on Amazon who ship their own packages will be required to pay a 2% fee per product sold.
This new fee is in addition to other charges Amazon already receives from merchants, including selling plan costs and referral fees based on product categories.
The fee comes as Amazon’s marketplace is under scrutiny, with the FTC planning to file an antitrust lawsuit over allegations that Amazon rewards third-party merchants using its logistics services while penalizing those fulfilling their own orders.
NYC bans TikTok from government devices
New York City is banning TikTok from government devices within 30 days, with immediate prohibition on downloading and usage by employees.
The NYC Cyber Command cited TikTok as a security threat to the city’s technical networks, prompting the decision.
While some states have broadly banned TikTok, most have restricted its use on government-owned tech, amid ongoing debates about the app’s security risks.
Unraveling August 2023: August 17th, 2023
Latest AI News and Trends on August 17th, 2023
You can now write one sentence to train an entire ML model.
How does it work?
You just describe the ML model you want…a chain of AI systems will take that sentence…it generates a dataset based on that sentence…and it trains a model for you…in ten minutes 😳
What does that mean?
Custom models in AI just got a whole lot easier. You can go from an idea (“a model that writes Python functions”) to a fully trained custom Llama-2 model in minutes 😮
Why should I care?
If you aren’t thinking about the impact of change in your industry, start now. It’s not linear and continuous, it’s exponential with step functions. 3 out of 4 C-suite executives believe that if they don’t scale artificial intelligence in the next five years, they risk going out of business entirely.
What should I do about it?
Further proof that AI is changing our work processes rapidly. You need to build a team and org that’s first and foremost, ready for change. And if you haven’t started pulling together an AI working group to get cracking on your AI usage principles and first AI use case, do it.
GPT-4 Code Interpreter masters math with self-verification
OpenAI’s GPT-4 Code Interpreter has shown remarkable performance on challenging math datasets. This is largely attributed to its step-by-step code generation and dynamic solution refinement based on code execution outcomes.
Expanding on this understanding, new research has introduced the innovative explicit code-based self-verification (CSV) prompt, which leverages GPT4-Code’s advanced code generation mechanism. This prompt guides the model to verify the answer and then reevaluate its solution with code.
The approach achieves an impressive accuracy of 84.32% on the MATH dataset, significantly outperforming the base GPT4-Code and previous state-of-the-art methods.
Why does this matter?
The study provides the first systematic analysis of the role of code generation, execution, and self-debugging in mathematical problem-solving. This highlights the importance of code understanding and generation capabilities in LLMs. Plus, the ideas presented can help build high-quality datasets that could potentially help improve the mathematical capabilities in open-source LLMs like Llama-2.
Multi-level machine learning models for estimating the risk of delay between cancer diagnosis and treatment initiation in a large group of cancer patients.
Study significance
Machine learning models that incorporate multi-level data sources can effectively identify cancer patients who are at a greater risk of experiencing treatment delays of more than 60 days after their initial cancer diagnosis.
Although neighborhood-level social determinants of health are incorporated in the study model as contributing variables, no significant impact of these factors was observed on the model performance. Furthermore, the model exhibits lower predictive effectiveness in vulnerable populations.
Future studies should include a higher proportion of vulnerable populations and more relevant social variables to improve the model performance.
Journal reference:
Frosch Z. A. K., Hasler, J., Handorf, E., et al. (2023). Development of a Multilevel Model to Identify Patients at Risk for Delay in Starting Cancer Treatment. JAMA Network Open. doi:10.1001/jamanetworkopen.2023.28712, https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2808249
Discover the top AI jobs shaping the future. Navigate the booming AI industry with insights on rewarding careers.
1. AI product manager
An AI product manager is similar to other program managers. Both jobs require a team leader to develop and launch a product. In this case, it is an AI product, but it’s not much different from any other product in terms of leading teams, scheduling and meeting milestones.
AI product managers need to know what goes into making an AI application, including the hardware, programming languages, data sets and algorithms, so that they can make it available to their team. Creating an AI app is not the same as creating a web app. There are differences in the structure of the app and the development process.
2. AI research scientist
An AI research scientist is a computer scientist who studies and develops new AI algorithms and techniques. They develop and test new AI models, collaborate with other researchers, publish research papers and speak at conferences. So, programming is only a small portion of what a research scientist does.
The tech industry is extremely open to self-taught and non-formally trained programmers, but it makes an exception for AI research scientists. They need to have a strong understanding of computer science, mathematics and statistics. Typically, they need graduate degrees.
3. Big data engineer
AI works with large data sets and so does its precursor, big data. A big data engineer is similar to an AI engineer because they are responsible for designing, building, testing and maintaining complex data processing systems that work with large data sets. But, instead of working with GPT or LaMDA, they work with big data tools, like Hadoop, Hive, Spark and Kafka.
Like AI researchers, big data engineers often have advanced degrees in mathematics and statistics. These degrees are necessary for designing, maintaining and building data pipelines based on massive data sets.
Business intelligence (BI) is also a data-driven discipline that predates the modern AI rush. Like big data and AI, BI also relies on large data sets. BI developers use data analytics platforms, reporting tools and visualization techniques to turn raw data into meaningful insights to help organizations make informed decisions.
BI developers work with a variety of coding languages and tools from major vendors, including SQL, Python, Tableau from Salesforce and Power BI from Microsoft. They also need to have a strong understanding of business processes to help improve them through data insight.
5. Computer vision engineer
A computer vision engineer is a developer who specializes in writing programs that utilize visual input sensors, algorithms and systems. These systems see the world around them and act accordingly, such as self-driving and self-parking cars and facial recognition.
They use languages like C++ and Python, along with visual sensors, such as Mobileye from Intel. Examples of use cases include object detection, image segmentation, facial recognition, gesture recognition and scenery understanding.
6. Data scientist
A data scientist is a technology professional who collects, analyzes and interprets data to solve problems and drive decision-making within the organization. They are not necessarily programmers, although many do write their own applications. Mostly, they use data mining, big data and analytical tools.
Their use of business insights derived from data enables businesses to improve sales and operations; make better decisions; and develop new products, services and policies. They use predictive modeling to forecast future events, such as customer churn, and data visualization to display research results visually. Some also use machine learning to build models to automate these tasks.
7. Machine learning engineer
A machine learning engineer is responsible for developing and implementing machine learning training algorithms and models. Training is the demanding side of machine learning and is the most processor- and computation-intensive aspect of machine learning. Therefore, it requires the highest level of skill and training.
Because of the need for advanced math and statistics skills, most machine learning engineers have advanced degrees in computer science, math or statistics. They often continue training through certification programs or a master’s degree in machine learning, deep learning or neural networks.
8. Natural language processing engineer
A natural language processing (NLP) engineer is a computer scientist who specializes in the development of algorithms and systems that understand and process natural human language input.
One of the big differentiators between traditional search engines and generative AI interfaces, such as ChatGPT, is that search engines use keywords and gather information from large amounts of existing online data. Generative AI creates new content based on other examples and patterns, and it answers queries in a chat-type format.
Like machine learning engineers, NLP engineers are not necessarily programmers first. They need to understand linguistics as much as they need to understand programming. NLP projects require machine translation, text summarization, answering questions and understanding context.
9. Robotics engineer
A robotics engineer is a developer who designs, develops and tests software for running and operating robots. Robotics has advanced significantly in recent years, such as automated home cleaners and precision cancer surgery equipment. Robotics engineers may also use AI and machine learning to boost a robotic system’s performance.
As a result, robotics engineers are typically designing software that receives little to no human input but instead relies on sensory input. Therefore, a robotics engineer needs to debug the software and the hardware to make sure everything is functioning as it should.
Robotics engineers typically have degrees in engineering, such as electrical, electronic or mechanical engineering.
10. Software engineer
A software engineer can cover various activities in the software development chain, including design, development, testing and deployment. Engineering professionals are needed at all points of software development. The demands are so high that it’s rare to find someone well versed in all of them. Most engineers tend to specialize in one discipline.
We spoke with MIT CSAIL head Daniela Rus about the emerging technology of liquid networks and implications for robotics.
The initial research papers date back to 2018, but for most, the notion of liquid networks (or liquid neural networks) is a new one. It was “Liquid Time-constant Networks,” published at the tail end of 2020, that put the work on other researchers’ radar. In the intervening time, the paper’s authors have presented the work to a wider audience through a series of lectures.
Ramin Hasani’s TEDx talk at MIT is one of the best examples. Hasani is the Principal AI and Machine Learning Scientist at the Vanguard Group and a Research Affiliate at CSAIL MIT, and served as the paper’s lead author.
“These are neural networks that can stay adaptable, even after training,” Hasani says in the video, which appeared online in January. When you train these neural networks, they can still adapt themselves based on the incoming inputs that they receive.”
The “liquid” bit is a reference to the flexibility/adaptability. That’s a big piece of this. Another big difference is size. “Everyone talks about scaling up their network,” Hasani notes. “We want to scale down, to have fewer but richer nodes.” MIT says, for example, that a team was able to drive a car through a combination of a perception module and liquid neural networks comprised of a mere 19 nodes, down from “noisier” networks that can, say, have 100,000.
“A differential equation describes each node of that system,” the school explained last year. “With the closed-form solution, if you replace it inside this network, it would give you the exact behavior, as it’s a good approximation of the actual dynamics of the system. They can thus solve the problem with an even lower number of neurons, which means it would be faster and less computationally expensive.”
The concept first crossed my radar by way of its potential applications in the robotics world. In fact, robotics make a small cameo in that paper when discussing potential real-world use. “Accordingly,” it notes, “a natural application domain would be the control of robots in continuous-time observation and action spaces where causal structures such as LTCs [Liquid Time-Constant Networks] can help improve reasoning.”
AI reconstructs song from brain activity
Neuroscientists recorded electrical activity from areas of the brain (yellow and red dots) as patients listened to the Pink Floyd song “Another Brick in the Wall, Part 1.” Using AI software, they were able to reconstruct the song from the brain recordings. This is the first time a song has been reconstructed from intracranial electroencephalography recordings.
Why does this matter?
By capturing the musicality of speech through neural signals, this research presents an innovative application of AI that could redefine how we interact and communicate, particularly for those who struggle with traditional modes of communication.
Saudi Arabia and UAE join the race for scarce Nvidia chips
Saudi Arabia has purchased at least 3,000 of Nvidia’s H100 chips at $40,000 apiece, while UAE has ordered a fresh batch of semiconductors to power its LLM. This signals the Gulf states’ intention to become major players in AI by buying up thousands of Nvidia’s GPUs which are vital in powering the boom in generative AI that has swept markets this year.
Why does this matter?
This makes them the latest to join the ever-growing queue of buyers for Nvidia chips to power AI ambitions. But will Nvidia be able to produce enough GPUs to meet the massive demand? It was reported in June that Nvidia GPUs are already in short supply (and very expensive).
Snapchat’s AI chatbot creates unexpected chaos
Snapchat users reported an unexpected video posted on the My AI chatbot’s Story, which some interpreted as showing a corner between a ceiling and a wall.
The unexpected post led to concerns and fears among users, with some believing the AI feature had become sentient or evolved, prompting some to delete the app.
Snapchat described the event as a “temporary outage”, which has since been resolved, and the AI chat feature temporarily stopped responding during this period.
Exploring the Power of Mojo Programming Language
Mojo is a new programming language that combines the usability of Python with the performance of C. It is designed to be the perfect language for developing AI models and applications. Mojo is fast, efficient, easy to use, and open source. Mojo is based on the LLVM (Low Level Virtual Machine) compiler infrastructure, which is one of the most advanced compiler frameworks in the world right now. Mojo uses a new type of system that allows for better performance and error checking. Mojo has a built-in autotuning system that can automatically optimize your code for the specific hardware that you are using.
Genmo is an artificial intelligence-driven video generator that takes text beyond the two dimensions of a page. Algorithms from natural language processing, picture recognition, and machine learning are used to adapt written information into visual form. It can turn text, pictures, symbols, and emoji into moving images. Background colors, characters, music, and other elements are just some of how the videos can be personalized. The movie will include the text and any accompanying images that you provide. The videos can be shared on many online channels like YouTube, Facebook, and Twitter. Videos made by Genmo’s AI can be used for advertising, instruction, explanation, and more. It’s a fantastic resource for companies, groups, and people who must rapidly and cheaply make interesting movies.
D-ID is a video-making platform powered by artificial intelligence that makes producing professional-quality videos from text simple and quick. Using Stable Diffusion and GPT-3, the company’s Creative RealityTM Studio can effortlessly create videos in over a hundred languages. D-ID’s Live Portrait function makes short films out of still images, and the Speaking Portrait function gives a speech to written or spoken text. Its API has been refined with the help of tens of thousands of videos, allowing it to generate high-quality visuals. Digiday, SXSW, and TechCrunch have all recognized D-ID for their ability to help users create high-quality videos at a fraction of the expense of traditional approaches.
The LeiaPix Converter is a web-based, no-cost service that changes regular photographs into 3D Lightfield photographs. It employs AI to turn your images into lifelike, immersive 3D environments. Select the desired output format and upload your picture to LeiaPix Converter. The converted file can be exported in several forms, including the Leia Image Format, Side-by-Side 3D, Depth Map, and Lightfield Animation. The LeiaPix Converter’s output is great quality and straightforward to use. It’s a fantastic way to give your pictures a new feel and make unique visual compositions. It does a 3D Lightfield conversion from a 2D image. Leia Image Format, Side-by-Side 3D, Depth Map, and Lightfield Animation are only a few of the supported export formats that bring about excellent outcomes. Depending on the size of the image, the conversion procedure could take a while. The quality of your original photograph will affect the final conversion outcomes. Because the LeiaPix Converter is currently in beta, it may include problems or have functionality restrictions.
A new open-source framework called instaVerse makes building your dynamic 3D environments easy. The background can be generated in response to AI cues, and players can then create their avatars to explore it. The first step in making a world in InstaVerse is picking a premade layout. Forests, cities, and even spaceships are just some of the many premade options available. After selecting a starter document, an AI assistant will guide you through the customization process. A forest with towering trees and a flowing river are just one of the many landscapes instaVerse may create at your command. Characters can also be generated in your universe. Humans, animals, and even robots are all included in the instaVerse cast of characters. Once a character has been created, you can use the keyboard or mouse to direct its actions. While InstaVerse is still in its early stages, it shows great promise as a robust platform for developing interactive 3D content. It’s simple to pick up and use and lets you make your special universes.
Sketch is a web app for turning sketches into GIF animations. It’s a fun and easy method to make unique stickers and illustrations to share on social media or use in other projects. Using Sketch is as easy as posting your drawing online. Then, you may utilize the drawing tools to give your work some life with some animation. Objects can be repositioned, recolored, and given custom sound effects. You can save your finished animation as a GIF after you’re satisfied. Sketch is a fantastic program for both young and old. It’s a terrific opportunity to show off your imagination and get a feel for the basics of animation simultaneously. In terms of ease of use, Sketch is excellent. Sketch makes it easy to create beautiful animations, even if you have no prior experience with the medium. With Sketch’s many tools, you can design elaborate and intricate animations. You can save your finished animation as a GIF after you’re satisfied. After that, your animation is ready for sharing or further use.
NeROIC can reconstruct 3D models from photographs as an element of AI technology. NeROIC, created by a reputable tech company, has the potential to transform our perceptions and interactions with three-dimensional objects radically. NeROIC can create a 3D model of the user’s intended message using an approved image. The video-to-3D capabilities of NeROIC are comparable to its image-to-3D capability. This means a user can create an interactive 3D setting from a single video. Because of this, creating 3D scenes is faster and easier than ever.
The discipline of computer science concerned with creating 3D models from 2D photographs is advancing quickly. Deep learning-based techniques may be used to train point clouds and 3D meshes to depict real-world scenes better. A potential method, DPT Depth Estimation, employs a deep convolutional network to read depth data from a picture and generate a point cloud model of the 3D object. DPT Depth Estimation uses monocular photos to input a deep convolutional network pre-trained on data from various scenes and objects. Following data collection, the web will use the information to create a point cloud from which 3D models can be made. When compared to conventional techniques like stereo-matching and photometric stereo, DPT’s performance can surpass a human’s. Because of its fast inference time, DPT is a promising candidate for real-time 3D scene reconstruction.
RODIN is quickly becoming the go-to 2D-to-3D generator in artificial intelligence. The creation of 3D digital avatars is now drastically easier and faster than ever before, thanks to this breakthrough. Creating a convincing 3D character based on a person’s likeness has always been more difficult. RODIN is an artificial intelligence-driven technology that can generate convincing 3D avatars using private data such as a client’s photograph. Customers are immersed in the action by seeing these fabricated avatars in 360-degree views.
That part at least seems pretty clear beyond a shadow of a doubt: Generative Enhanced Multimodal Intelligent Network Interface.
The word “Gemini” comes from Latin and means “twins” in German. Some possible meanings in the context of Google’s AI system:
Gemini combines two components: Text and image processing. It is, in a sense, a “twin system.”
Gemini could refer to the „twins“ Sergey Brin and Larry Page, the founders of Google.
Astrology assigns communication strength and flexibility to the zodiac sign Gemini. Gemini as an AI assistant aims to adapt linguistically and situationally.
The name suggests a dual strength or ability. Gemini aims to unite Google’s text and image AI to outperform the competition.
As a twin system, Gemini combines different perspectives and approaches, similar to different human characters. So the name is both an allusion to the system’s integrative capabilities and a promising indication of Google’s ambitions with this AI product.
Why is Google superior?
To do that, you have to understand WHAT treasure trove of data Google is actually sitting on. Here are a few facts:
Google, through its various services such as Google Search, YouTube and others, has an enormous amount of data that is very useful for developing AI systems.
On YouTube alone, over 500 hours of video material are uploaded every day, according to Statista. The total video database is over 30 million hours of video. The subtitles and transcripts of these videos give Google a gigantic text dataset for training language models.
According to a report by ARK Invest, Google owns over 130 exabytes of data. For comparison, 1 exabyte is equal to 1 billion gigabytes. This means that the entire data set comprises more than 130,000,000,000,000,000 bytes of information.
Google Search accounts for a large part of this data. Google says it processes over 40,000 search queries per second. That’s over 3.5 trillion search queries per year. From these queries and the clicked results, Google gains further insights.
Overall, it shows that Google has virtually inexhaustible data resources for AI research. Both the breadth of different types of data and the sheer volume should give Google a significant edge in the AI field.
Google – The Research Giant
In 2020, Google published over 1300 artificial intelligence research papers, according to the Papers with Code database. In 2021, Google increased the number of publications significantly again to over 2000 papers on AI and machine learning.
Topics included:
Computer Vision (image recognition)
Natural Language Processing (NLP)
Speech Recognition
reinforcement learning
Robotics
Multimodal AI
Recommender Systems
Applications in medicine
With over 3300 AI publications in 2020 and 2021, Google has greatly expanded its research output in artificial intelligence. The company is one of the most active players in this research field. This intensive work over the past few years is now being incorporated into the development of Gemini.
According to the AI publication database Papers with Code, Google published more than 1,500 artificial intelligence research papers in 2022 alone. That’s far more than other tech corporations like Meta or Microsoft.
This is a partial selection of Google’s most groundbreaking developments in AI in recent years. The list shows the enormous range of research from machine learning and computer vision to robotics and autonomous systems.
AlphaGo: Go game AI that defeated world champion Lee Sedol in 2016.
BERT (Bidirectional Encoder Representations from Transformers): breakthrough language model for NLP from 2018.
PaLM (Pathways Language Model): enormous language model with 540 billion parameters from 2022
PaLM-SayCan: variant of PaLM that can carry on human-like conversations
Imagen: image generation AI for realistic and creative images
MusicLM: AI for music composition and production
RLHF (Reinforcement Learning with Human Feedback): Reinforcement learning with human feedback
Model Based RL: reinforcement learning with explicit models of the environment
RobustFit: Robust neural network against data noise
T5: Text-to-text transfer transducer for various NLP tasks
ViT (Vision Transformer): Image recognition with Transformer architecture
WAYMO: Autonomous driving and robot cab service
ProteinFold: Protein structure prediction with Deep Learning
FLOOD: AI for flood prediction and prevention
SLIDE: pixel-level image segmentation
Switch Transformers: efficient architecture for very large transformers
MuZero: reinforcement learning without environmental model in games
Meena: conversational AI from 2020
DALL-E & DALL-E 2: text-to-image generation.
When you look at the sheer amount of data Google has collected over the years, it initially makes you dizzy. Over 500 hours of video footage are uploaded to YouTube every day. The total video database is over 30 million hours. Add to that countless search queries, texts, images and conversations. It’s an almost unimaginable amount of data.
Coupled with intensive research activity in the AI field, it adds up to enormous potential. In recent years, Google has produced groundbreaking innovations such as the BERT language model, the AlphaGo Go AI, and the DALL-E image generator. When you put all these puzzle pieces together, things take on almost frightening proportions.
Project: Google Gemini
With the new Gemini AI system, Google now seems to have bundled the essence of these years of data aggregation and research. If the company succeeds in combining all of its AI developments and treasure trove of data in this system, it would be a demonstration of the sheer power of innovation. It will be interesting to see whether Gemini can deliver on this promise. In any case, the expectations are huge – here what we know and what the rumors say:
Facts Google Gemini
There are already some facts from the Google Blog:
Gemini is supposed to be released this fall
Gemini combines text and image generation
Can create contextual images based on text generation
Has been trained with YouTube transcripts
Google lawyers are monitoring the training to avoid copyright issues
Gemini is said to have multiple modalities, e.g., text, image, audio, video
Sergey Brin is involved in development
Rumors
From Reddit and countless other sources on the web, there could be other features as well:
Gemini is said to be capable of AI image understanding and modification
Is said to combine text capabilities like GPT-4 with image generation
Has been developed from the ground up as a multimodal model
Could handle audio, video, 3D renderings, graphics, etc.
Shall learn with user interactions and thus become effective AGI
Architecture could enable lifelong learning
There are concerns about privacy and information leaks between users
Google Gemini and the (then new) AI market:
The AI market situation is likely to change significantly with the introduction of Google Gemini:
For OpenAI:
Strong new competitor for ChatGPT and DALL-E.
Google has significantly more resources and data
OpenAI could lose market share and come under pressure
For Anthropic:
Claude must stand up to Google Assistant with Gemini
Advantage due to focus on security and control
Risk of falling behind
For Microsoft:
Partnership with OpenAI important to compete with Google
Microsoft must further develop Azure AI services
Advantage due to strong cloud infrastructure
For others:
Startups could have a very hard time against Google
Consolidation in the market possible
Significantly higher innovation speed
Overall, competitive pressure in the AI market will increase sharply. With its resources, Google is in a very good starting position to take a leading role with Gemini. It will be more difficult for other providers to keep pace with Google. It remains to be seen whether the high expectations for Gemini are justified.
Google Gemini Conclusion
Google Gemini seems to be a very ambitious AI project that should give the company a competitive edge. The combination of different modalities in one model is new and could improve AI capabilities tremendously. However, there are still many unanswered questions regarding the specific capabilities and data security. The release this fall will show whether Google can deliver on its promise to outperform the competition. Much is still speculation, but expectations are high.
#ai #ki #google #gemini #text #image #multimodal
Artificial intelligence steps in to assist dementia patients with ‘AI Powered Smart Socks’
People suffering from dementia could live more independently thanks to a pair of AI-powered socks that can track everything from a patient’s heart rate to movement.
Called “SmartSocks,” the AI-powered apparel was created in partnership between the University of Exeter and researchers at the start-up company Milbotix, according to SWNS. The socks can monitor a patient’s heart rate, sweat levels and motion to prevent falls while also promoting independence for those with dementia.
“I came up with the idea for SmartSocks while volunteering in a dementia care home,” SmartSocks creator Zeke Steer, CEO of Milbotix, told SWNS. “The current product is the result of extensive research, consultation and development.”
Steer’s great-grandmother suffered from dementia, which also helped spark the creation of the socks.
“The foot is actually a great place to collect data about stress, and socks are a familiar piece of clothing that people wear every day; our research shows that socks can accurately recognize signs of stress, which could really help not just those with dementia but their caregivers, too,” Steer, who has a background in robotics and AI, told SWNS.
The socks send the data collected from the patient to an app, which flags caregivers when the patient appears to be in distress. The warning could prevent falls and even tragedies as caregivers can respond to a patient before their stress escalates.
“I think the idea of SmartSocks is an excellent way forward to help detect when a person is starting to feel anxious or fearful,” said Margot Whittaker, director of nursing and compliance at Southern Healthcare in the U.K.
A handful of care homes overseen by Southern Healthcare, including The Old Rectory in Exeter, are already testing the tech-powered socks on patients, who report they are happy with how easy the socks are to use.
“Anything that’s simple and easy to do, and is improving our look at life as a whole, I’m happy with,” dementia patient John Piper, 83, told the BBC.
The socks do not need to be recharged, according to Milbotix’s website, and can be machine washed.
There are other products on the market that can also track a dementia patient’s heart rate or sweat levels, but they often come in the form of wristbands and watches, which can pose issues to those with dementia.
“Wearable devices are fast becoming an important way of monitoring health and activity,” Imperial College London’s Health and Social Care Lead Sarah Daniels told SWNS. “At our center, we have been trialing a range of wristbands and watches. However, these devices present a number of challenges for older adults and people affected by dementia.”
Daniels said wristbands or watches often don’t hold long charges and are taken off by patients and then lost.
“SmartSocks offer a new and promising alternative, which could avoid many of these issues,” Daniels said.
The University of Exeter is investigating how beneficial the socks are for dementia patients.
Artificial intelligence platforms are revamping health care across many disciplines, including another U.K.-based system called CognoSpeak, which can monitor speech patterns in a bid to detect early signs of dementia or Alzheimer’s.
U.K.-based start-up SmartSocks has developed hosiery that can monitor a dementia patient’s heart rate, motion and sweat levels with AI and alert caregivers to potential problems.
GPT-4 Code Interpreter can enhance math skills with code-based self-verification – OpenAI’s GPT-4 Code Interpreter’s remarkable performance in math datasets is largely attributed to its step-by-step code generation and dynamic solution refinement based on code execution outcomes. Expanding on this understanding, new research has introduced the innovative explicit code-based self-verification (CSV) prompt, which leverages GPT4-Code’s advanced code generation mechanism. This prompt guides the model to verify the answer and then reevaluate its solution with code. – The approach achieves an impressive accuracy of 84.32% on the MATH dataset, significantly outperforming the base GPT4-Code and previous state-of-the-art methods.
AI just reconstructed a Pink Floyd song from brain activity, and it sounds shockingly clear – Neuroscientists recorded electrical activity from areas of the brain as patients listened to the Pink Floyd song “Another Brick in the Wall, Part 1.” Using AI software, they were able to reconstruct the song from the brain recordings. This is the first time a song has been reconstructed from intracranial electroencephalography recordings.
Saudi Arabia and UAE join the race for scarce Nvidia chips – Saudi Arabia has purchased at least 3,000 of Nvidia’s H100 chips at $40,000 apiece, while UAE has ordered a fresh batch of semiconductors to power its LLM. This signals their intention to become major players in AI.
OpenAI acquires Global Illumination to work on core products, including ChatGPT – Its team leverages AI to build creative tools, infrastructure, and digital experiences. It previously designed and built products early on at Instagram and Facebook and has made significant contributions at YouTube, Google, Pixar, Riot Games, and other notable companies.
McKinsey unveils its own generative AI tool for employees: Lilli – It is a chat application for employees designed that serves up information, insights, data, plans, and even recommends the most applicable internal experts for consulting projects, all based on 100K+ documents and interview transcripts.
Opera’s iOS web browser will now include Aria – The AI assistant, Aria, is Opera’s browser AI product built in collaboration with OpenAI, integrated directly into the web browser, and free for all users.
Adobe Express with AI Firefly app is available worldwide – The web app is now out of beta and can be used free of charge in web browsers.
The Associated Press releases guidelines for Generative AI to its journalists
UK is using AI road safety cameras to detect potential driver offenses in passing vehicles
The founder of Centricity, a data analytics firm using AI, is indicted for defrauding investors by manipulating financial data.
Leaders with a Montana digital academy say bringing artificial intelligence to high schools is an opportunity to embrace the future.
Google said to be testing new life coach AI for providing helpful advice to people.
Alibaba Cloud MagicBuild Community has launched the digital human video generation tool called LivePortrait. It can generate digital human videos from photos, text, or voice, which can be applied in scenarios such as live broadcasting and corporate marketing.
Are physical SIMs about to be a thing of the past? Jump into the latest, and discover eSIMs #sponsored
Latest Sport Football Soccer News and Trends on August 17th, 2023
Atletico Madrid takes the alleged ‘threat’ of João Félix terminating his contract as a joke. If he wants to do so, he will have to pay his €350m release clause. Link
BREAKING: Theo Walcott is set to retire. Walcott, 34, who left Southampton at the end of last season, scored more than 100 goals for Arsenal and won 47 England caps. Link
From epic open-world adventures to mind-bending puzzles, these are the best iOS games to play on the upcoming iPhone 15.
Unraveling August 2023: August 16th, 2023
Latest AI News and Trends on August 16th, 2023
GPT-4 to replace content moderators
OpenAI aims to use its GPT-4 to solve the challenge of content moderation at scale. Also, they already used GPT-4 to develop and refine their own content policies. It provides three major benefits: consistent judgments, faster policy development, and improved worker well-being. However, perfect content moderation remains elusive, as both humans and machines make mistakes, particularly in handling misleading or aggressive content.
GPT-4 can interpret complex policy documentation and adapt instantly to updates, reducing the cycle from months to hours. This AI-assisted approach offers a positive future for digital platforms, where AI can help moderate online traffic and relieve the burden on human moderators.
Why does this matter?
GPT-4 can alleviate content moderation challenges and improve the efficiency and effectiveness of content moderation. This could be a solution for platforms like Facebook and Twitter, who’ve been grappling with content moderation for ages. OpenAI’s this approach could also appeal to smaller companies lacking resources.
Shepherd is a language model designed to critique and improve the outputs of other language models. It uses a high-quality feedback dataset to identify errors and provide suggestions for refinement. Despite its smaller size, Shepherd’s critiques are either equivalent or preferred to those from larger models like ChatGPT. In evaluations against competitive alternatives, Shepherd achieves a win rate of 53-87% compared to GPT-4.
Shepherd outperforms other models in human evaluation and is on par with ChatGPT. Shepherd offers a practical and valuable tool for enhancing language model generation.
Why does this matter?
Despite Shepherd’s smaller size, its critiques match or surpass those of larger models like ChatGPT, with a win rate of 53-87% against GPT-4. It excels in human evaluations and offers practical value in improving language model generation.
Microsoft now offers OpenAI’s ChatGPT model in its Azure OpenAI service, allowing developers and businesses to integrate conversational AI into their applications. ChatGPT can be used to power custom chatbots, automate emails, and provide summaries of conversations.
Azure OpenAI users can access a preview of ChatGPT starting today, with pricing set at $0.002 for 1,000 tokens. ChatGPT on Azure solution accelerator is an enterprise option. This solution provides a similar user experience to ChatGPT but is offered as your private ChatGPT.
Microsoft Azure ChatGPT offers several benefits to organizations:
Ensures data privacy with built-in guarantees and isolation from OpenAI-operated systems.
Allows full network isolation and offers enterprise-grade security controls.
Enhances business value by integrating internal data sources and services like ServiceNow.
Why does this matter?
Amid the excitement around ChatGPT, Microsoft has cleverly introduced an enterprise version to meet strong market demand. By prioritizing security, Azure simplifies and enhances companies’ access to AI advantages. Also, Microsoft’s move aims to boost productivity through code editing, task automation, and more and offers enterprises a more secure way to share their data with AI.
Nvidia’s stock rises 7% as investors see its GPUs remaining dominant in powering large language models.
Morgan Stanley reiterates Nvidia as a “Top Pick” due to strong earnings, AI spending shift, and ongoing supply-demand imbalance.
Despite recent fluctuations, Nvidia’s stock has tripled in 2023, and analysts anticipate long-term benefits from AI and favorable market conditions.
The Strength and Realism of AI Models While artificial intelligence models demonstrate immense computational power, there’s a debate regarding their biological plausibility. How do these digital frameworks compare to the natural intelligence of living organisms? Are they accurate representations or mere simulations?
Transportation Systems: The Paradox of Choice More choices in transportation systems might seem beneficial, but there’s a hidden challenge. With increased variety comes complexity, leading to inefficiencies and potential gridlocks.
AI’s Role in Pinpointing Cancer Origins Recent advancements in AI have developed a model that can assist in determining the starting point of a patient’s cancer, a crucial step in identifying the most effective treatment method. [Read more at MedicalTechNews.com]
AI’s Defense Against Image Manipulation In the era of deepfakes and manipulated images, AI emerges as a protector. New algorithms are being developed to detect and counter AI-generated image alterations. [Read more at DigitalSafetyWatch.com]
Streamlining Robot Control Learning Researchers have uncovered a more straightforward approach to teach robots control mechanisms, making the integration of robotics into various industries more efficient.
Accelerated Robotics Training Techniques A revolutionary methodology promises to slash the time required to instruct robots, optimizing their utility and deployment speed in multiple applications.
Armando Solar-Lezama: The Beacon of Computing Armando Solar-Lezama has been honored as the inaugural Distinguished Professor of Computing, acknowledging his invaluable contributions to the world of computer science.
Efficient Planning for Household Robots with AI AI integration has enabled household robots to plan tasks more efficiently, cutting their preparation time by half and allowing for more seamless operations in domestic environments.
The ChatGPT Impact: Boosting Writing Productivity A recent study highlights how ChatGPT enhances workplace productivity, particularly in writing tasks. The AI-driven tool provides a significant advantage for professionals in diverse sectors.
Reimagining Data Privacy in the Modern Era Data privacy is evolving, and it’s time to approach it with a fresh perspective. As digital footprints expand, there’s an urgent need to revisit and redefine what personal data protection means.
Daily AI News on August 16th, 2023
OpenAI’s GPT-4 for more reliable and higher quality content moderation – OpenAI aims to use its GPT-4 to solve the challenge of content moderation at scale. GPT-4 could replace human moderators, offering similar accuracy and more consistency. OpenAI has already used GPT-4 to develop and refine its own content policies. – It provides three major benefits: consistent judgments, faster policy development, and improved worker well-being. While AI has been used for content moderation before, OpenAI’s approach could be appealing to smaller companies lacking resources.
Microsoft launches ChatGPT for enterprises with Azure – Microsoft is now offering OpenAI’s ChatGPT model in its Azure OpenAI service, allowing developers and businesses to integrate conversational AI into their applications. ChatGPT can be used to power custom chatbots, automate emails, and provide summaries of conversations. – Azure OpenAI users can access a preview of ChatGPT starting today, with pricing set at $0.002 for 1,000 tokens and it promises more control and privacy compared to the public model.
Google is progressing with new AI updates! – Search experience adds AI-powered summaries, definitions, and coding improvements. In addition it will include related diagrams or images for various topics, color-coded syntax highlighting for code snippets, making it easier for programmers to understand and debug generated code. – Google Photos adds a scrapbook-like Memories view feature aided by AI which allows users to relive and share their most memorable moments. The feature creates a scrapbook-like timeline that includes trips, celebrations, and daily moments with loved ones. The new Memories view is launching today for U.S. users and is similar to a combination of Stories and Facebook Memories.
Amazon using AI to enhance product reviews – Amazon is tapping into generative AI to create handy highlights that collects key points from customer reviews which will help shoppers quickly gauge product review. – The feature is part of ongoing efforts to improve utility of 125M+ reviews from shoppers. It uses only trusted reviews from verified purchases, and Amazon.
WhatsApp test beta upgrade with new feature ‘custom AI-generated stickers’ – The feature is currently available to a limited number of beta testers, includes a “Create” button under the stickers tab, which opens a keyboard for users to type prompts for the AI model to generate custom stickers. The feature is a server-side change and is currently only available in version 2.23.17.8 of the beta version.
Apple’s AI advancements in the last few months
Don’t sleep on Apple’s AI plans. Here’s how they’ve been slowly ramping up their AI efforts in the last few months.
Apple’s AI-powered health coach might soon be at your wrists Apple is reportedly developing an AI-powered health coaching service called Quartz, aimed at helping users improve their exercise, eating habits, and sleep quality. The service will use AI and data from the user’s Apple Watch to create personalized coaching programs, with plans to introduce a monthly fee. The company is also working on emotion-tracking tools and plans to launch an iPad version of the iPhone Health app this year.
Apple enters the AI race with new features Apple announced a host of updates at the WWDC 2023. Yet, the word “AI” was not used even once, despite today’s pervasive AI hype-filled atmosphere. The phrase “machine learning” was used a couple of times. (And AI is nothing but machine learning). However, here are a few announcements Apple made that use AI as the underlying technology.
Apple Vision Pro, a revolutionary spatial computer that seamlessly blends digital content with the physical world. It uses advanced ML techniques.
Upgraded Autocorrect in iOS 17 that is powered by a transformer language model for improved prediction capabilities.
Improved Dictation in iOS 17 that leverages a new speech recognition model to make it even more accurate.
Live Voicemail that turns voicemail audio into text on the fly, which is powered by a neural engine.
Personalized Volume, which uses ML to understand environmental conditions and listening preferences over time to automatically fine-tune the media experience.
Journal, a new app for users to reflect and practice gratitude, uses on-device ML for personalized suggestions to inspire entries.
Apple Trials a ChatGPT-like AI Chatbot Apple is developing AI tools, including its own large language model called “Ajax” and an AI chatbot named “Apple GPT.” They are gearing up for a major AI announcement next year as it tries to catch up with competitors like OpenAI and Google.
Apple bets big on AI Apple’s CEO Tim Cook has stated that AI and ML are embedded in every company product. This comes after concerns were raised about Apple’s lack of discussion on its AI plans while competitors have been actively incorporating the technology into their products. He also emphasized that AI is central to the design of Apple’s products, contradicting suggestions that the company has not yet integrated the technology.
Apple gearing up for an AI showdown Apple has reportedly ordered servers from Foxconn Industrial Internet, a division of its supplier Foxconn, for the testing and training of AI services. The servers are specifically for Apple’s AI work, which has been a focus for the company for years. While Apple does not currently have a ChatGPT-like app for external use, it is believed that this division of Foxconn already supplies servers to ChatGPT OpenAI, Nvidia, and Amazon Web Services. The news comes amidst reports about Apple’s plans to compete in the AI chatbot market.
The U.S. Consumer Financial Protection Bureau (CFPB) plans to regulate data brokers selling personal data due to concerns about their impact on privacy, including sensitive data from vulnerable groups.
CFPB aims to prevent illegal collection and sharing of personal data by data brokers in the surveillance industry.
The proposal expands coverage under the Fair Credit Reporting Act to include data derived from payment histories, personal income, and criminal records, addressing concerns such as credit header data disclosure.
Tesla unveils cheaper Model S and Model X variantsLINK
Tesla introduces lower-priced options for the Model S and Model X with reduced range.
The “standard range” trim brings the Model S starting price to $78,490 with 320 miles of range and the Model X starting price to $88,490 with 269 miles of range.
This trim, previously discontinued, likely uses the same battery pack with a software lock, and Tesla may offer range unlocking for an additional fee.
Singapore pioneers in stablecoin crypto regulationLINK
Singapore’s financial regulator has finalized rules for stablecoins, making it one of the first jurisdictions to do so globally.
Stablecoins are digital currencies designed to maintain a constant value against fiat currency, with a market value of around $125 billion and dominated by tokens like USDT and USDC.
The Monetary Authority of Singapore’s framework outlines requirements including holding reserves in low-risk assets, timely redemption, and proper user disclosures for stablecoins mimicking the Singapore dollar or other G10 currencies.
Amazon Pharmacy offers major savings on insulinLINK
Amazon Pharmacy offers automatic coupons to help uninsured insulin-requiring patients save on medication costs.
The digital pharmacy shows pricing with and without insurance, estimates savings with eligible coupons, and simplifies the process compared to existing coupon programs.
Amazon Pharmacy’s initiative has been praised by healthcare advocates and aims to improve access to affordable treatments, but manufacturer coupons are not available for patients benefiting from certain healthcare programs.
Other Tech news you might like
Apple will soon start making settlement payments to claimants of the “Batterygate” class-action lawsuit, with potential payments of up to $65 per person.LINK
YouTube is removing cancer treatment content that’s “harmful or ineffective,” implementing new guidelines to remove unproven treatments and harmful advice.LINK
California regulators approved round-the-clock robotaxi service in San Francisco for Waymo and Cruise, but a group of Cruise vehicles caused a traffic backup, highlighting challenges.LINK
Microsoft is implementing an eight-strike suspension policy for Xbox community standards violations, scaling suspensions based on the number of strikes, with players able to appeal and view enforcement history.LINK
Unraveling August 2023: August 15th, 2023
Latest AI News and Trends on August 15th, 2023
Do It Yourself Custom AI Chatbot for Business in 10 Minutes (Open Source)
If you’re looking to “train” a custom chatbot on your data (SOPs, legal docs, financial reports, etc), I’d strongly suggest checking out AnythingLLM. It’s the first chatbot with enterprise-grade privacy & security. When using ChatGPT, OpenAI collects your data including: – Prompts & Conversations – Geolocation data – Network activity information – Commercial information e.g. transaction history – Identifiers e.g. contact details – Device and browser cookies – Log data (IP address etc.) However, if you use their API to interact with their LLMs like gpt-3.5 or gpt-4, your data is NOT collected. This is exactly why you should **build your own private & secure chatbot**. That may sound difficult, but Mintplex Labs (backed by Y-Combinator) just released AnythingLLM, which gives you the ability to build a chatbot in 10 minutes without code. AnythingLLM provides you with the tools to easily build and manage your own private chatbot using API keys. Plus, you can expand your chatbot’s knowledge by importing data such as PDFs, emails, etc. This can be confidential data as only you have access to the database. ChatGPT currently allows you to upload PDFs, videos and other data to ChatGPT via vulnerable plug-ins, BUT there is no way to determine if that data is secure or even know where it’s stored. Easily build your own business-compliant and secure chatbot at useanything.com. All you need is an OpenAI or Azure OpenAI API key. Or, if you prefer using the open source code yourself, here’s the GitHub repo: https://github.com/Mintplex-Labs/anything-llm.
AI powered tools for the recruitment industry
AI-driven recruiting and retention strategies utilize data-driven strategies for better candidate experiences and better hiring decisions. Here’s a list of a few tools that are useful for this purpose : – Conversational AI To Recruit And Retain At Scale | Humanly.io : It is designed for high scale hiring in organizations. It enhances candidate engagement through automated chat interactions.
– MedhaHR : It’s an AI-driven healthcare talent sourcing platform that automates resume screening, provides personalized job recommendations, and offers cost-effective solutions.
– ZappyHire : It offers features such as candidate sourcing, resume screening, automated communication, and collaborative hiring.
– Sniper AI : It uses AI algorithms to source potential candidates, assess their suitability, and integrates with ATS for workflow optimization.
– PeopleGPT : PeopleGPT, developed by Juicebox (YC S22), is a tool that simplifies the process of searching for people data. Recruiters can input specific queries to find potential candidates. Which tools have you been using, and more importantly is AI really helping you with recruitment? More resources along with their pricing plans here
American companies are vigorously seeking AI specialists, leading to soaring salaries for high-demand roles. Amidst this recruitment frenzy, some organizations are offering nearly a million-dollar salary, especially to those experienced in AI.
Surge in AI Talent demand and salaries
American firms are hunting for AI experts, with some offering salaries nearing a million dollars.
Industries like entertainment and manufacturing want data scientists and machine-learning specialists.
Competition is fierce, with companies like Accenture investing in internal training and others considering acquisition of AI startups for talent.
The compensation landscape for AI roles
As AI expertise becomes more sought-after, compensation packages are rising.
Companies are offering mid-six-figure salaries, bonuses, and stock grants to lure experienced professionals.
While top positions like Netflix’s machine-learning platform product manager can reach up to $900,000 in total compensation, othersalike a prompt engineer might average $130,000 annually.
How to Manage Your Remote Team Effectively with ChatGPT?
Leading a remote team comes with unique challenges, from ensuring clear communication to fostering a sense of community. ChatGPT can be your expert consultant, offering suggestions based on best practices for remote team management.
You are a seasoned consultant in remote team management. I am the leader of a remote team working on a [define project]. I need advice on how to effectively manage my team, ensure clear communication, monitor progress, and maintain a positive team culture. Your suggestions should include strategies for scheduling and conducting virtual meetings, task assignment, progress tracking tools, and methods to promote team building in a virtual setting.
I asked ChatGPT to remove password protection from an Excel document, and it worked flawlessly
How are you uploading an excel document to chat gpt?
Using ChatGPT code interpreter: It’s a feature for GPT plus member as the old “bing search” which got disabled, You have code interpreter now where you can directly upload files.
Can it analyze conversations/texts? Yes it can analyse data and even give u back charts and feedback for gpt plus users.
Johns Hopkins Engineers and Cancer Researchers have collaboratively pioneered a breakthrough in personalized cancer therapy with their cutting-edge deep-learning technology.
Summary: Microsoft Azure allows organizations to run ChatGPT within their network for smoother work experiences. Think of it as your private, controlled, and extra valuable AI assistant. (source)
Key points:
Azure allows companies to run ChatGPT privately on their own networks, touting built-in data isolation from OpenAI.
The model connects to internal data services and sources, and is available on GitHub to install and deploy.
Benefits include privacy, control, and unique business value through internal data integration.
Why It Matters: For enterprises, this merger between ChatGPT and Azure opens a new realm of possibilities, with the cozy feeling of privacy and control. It’s more than a tech tool; it’s a tailored solution that could redefine how businesses work with AI.
Apple’s AI-powered health coach might soon be at your wrists
Apple is reportedly developing an AI-powered health coaching service called Quartz, aimed at helping users improve their exercise, eating habits, and sleep quality. The service will use AI and data from the user’s Apple Watch to create personalized coaching programs, with plans to introduce a monthly fee. The company is also working on emotion-tracking tools and plans to launch an iPad version of the iPhone Health app this year.
Why does this matter?
It’s only a matter of time before AI is deployed on IoT devices such as smartwatches. This confluence can definitely revolutionize our daily lives. AI can direct IoT devices to adapt and optimize settings based on external circumstances making them a lot more autonomous and helpful.
Apple announced a host of updates at the WWDC 2023. Yet, the word “AI” was not used even once, despite today’s pervasive AI hype-filled atmosphere. The phrase “machine learning” was used a couple of times. (And AI is nothing but machine learning). However, here are a few announcements Apple made that use AI as the underlying technology.
Apple Vision Pro, a revolutionary spatial computer that seamlessly blends digital content with the physical world. It uses advanced ML techniques.
Upgraded Autocorrect in iOS 17 that is powered by a transformer language model for improved prediction capabilities.
Improved Dictation in iOS 17 that leverages a new speech recognition model to make it even more accurate.
Live Voicemail that turns voicemail audio into text on the fly, which is powered by a neural engine.
Personalized Volume, which uses ML to understand environmental conditions and listening preferences over time to automatically fine-tune the media experience.
Journal, a new app for users to reflect and practice gratitude, uses on-device ML for personalized suggestions to inspire entries.
Why does this matter?
To the average user, AI can be scary. Perhaps it was Apple’s deliberate choice not to mention the word “AI”? Nevertheless, these updates and features demonstrate that Apple is indeed utilizing AI technologies in various aspects of its products and services, joining the likes of Google and Microsoft.
Apple is developing AI tools, including its own large language model called “Ajax” and an AI chatbot named “Apple GPT.” They are gearing up for a major AI announcement next year as it tries to catch up with competitors like OpenAI and Google.
The company has multiple teams developing AI technology and addressing privacy concerns. While Apple has been integrating AI into its products for years, there is currently no clear strategy for releasing AI technology directly to consumers. However, executives are considering integrating AI tools into Siri to improve its functionality and keep up with advancements in AI.
Why does this matter?
Apple’s development of AI tools, such as the language model “Ajax” and chatbot “Apple GPT,” signals the company’s efforts to catch up with competitors OpenAI and Google. The focus on addressing privacy concerns and the potential integration of AI into Siri shows Apple’s aim to enhance its product functionality and stay competitive.
Apple’s CEO Tim Cook has stated that AI and ML are embedded in every company product. This comes after concerns were raised about Apple’s lack of discussion on its AI plans while competitors have been actively incorporating the technology into their products. He also emphasized that AI is central to the design of Apple’s products, contradicting suggestions that the company has not yet integrated the technology.
Cook reassured that Apple has invested in AI for years, and this year’s Research & Development spending has hit $22.61 billion. They are also hiring dozens of AI jobs in the US, France, and China, looking to fill roles that could help build Gen AI tools.
Why does this matter?
This move signifies the potential for enhanced personalization and contextual relevance in user interactions, leading to a more intuitive and tailored experience within the Apple ecosystem. The seamless integration of AI may also pave the way for groundbreaking applications in health, home automation, and more. Ultimately redefining how users interact with and benefit from Apple’s ecosystem of products and services.
Apple has reportedly ordered servers from Foxconn Industrial Internet, a division of its supplier Foxconn, for the testing and training of AI services. The servers are specifically for Apple’s AI work, which has been a focus for the company for years. While Apple does not currently have a ChatGPT-like app for external use, it is believed that this division of Foxconn already supplies servers to ChatGPT OpenAI, Nvidia, and Amazon Web Services. The news comes amidst reports about Apple’s plans to compete in the AI chatbot market.
Why does this matter?
Apple’s this latest move to order servers from Foxconn’s division for AI testing and training has caught attention. While Apple hasn’t launched a ChatGPT-like app yet, the supplier’s involvement with ChatGPT OpenAI, Nvidia, and Amazon Web Services hints at potential AI ventures. Apple seems like a potentially new big player in the AI game.
The ‘SGE while browsing’ feature can break down a long web page into easy-to-read key points.
Daily AI News August 15th, 2023
Talon Aerolytics, a leading innovator in SaaS, Digital Twin capture services and AI technology, has announced ha its groundbreaking cutting-edge AI-powered computer vision platform enables wireless operators to visualise and analyse network assets using end-to-end AI and machine learning. Link
Beijing is poised to implement sweeping new regulations for artificial intelligence services this week, trying to balance state control of the technology with enough support that its companies can become viable global competitors. Link
Saudi Arabia and the United Arab Emirates are buying up thousands of the high-performance Nvidia chips crucial for building artificial intelligence software, joining a global AI arms race that is squeezing the supply of Silicon Valley’s hottest commodity. Link
OpenAI likely to go bankrupt by the end of 2024. Link
Latest Tech News on August 15th, 2023
Youtube algorithm flaws?
Personally I’ve always been a huge fan of youtube but I always thought that their algorithm have actually gotten worse since the 2010s.
Supposedly Google should have perfected the algorithm at making simple recommendations; they have teams working on it yet i could think of a few things that could improve it.
From my experience, youtube always recommends the same stuff. If you like one video or click on it, it would keep showing that channels videos until you’re bombarded by it. It rarely gives you anything new, or reminds you of old topics you enjoyed. Sometimes videos are just stuck there for weeks, when i’m clearly not watching it. Sometimes something i really want to watch disappears and never comes back again. Furthermore It’s other sections/buttons do not show the videos i mention above, but rather completely unrelated content.
Just off the top of my head, I can think of a few things – becoming more ambitious every refresh; recommend new topics more often; remind you of old topics you like more often.
YouTube announced a new policy for combatting medical misinformation on the video sharing platform
Latest World and Sport News on August 15th, 2023
Fulton County grand jury returns an indictment in 2020 election probe for Georgia. Link
The highest paid football players in the world in 2023 according to Le Parisien
After Al-Hilal move, Neymar leapfrogs Romelu Lukaku as the player with the highest combined transfer fee in football history.
Asked ChatGPT “Can you write an essay about web browsers but using a tone of a depressed teenager?” And this was the result
Ugh, like, seriously? Web browsers, man, they’re just like another thing we have to deal with in this messed up digital world. It’s like, we can’t even escape the internet, even if we wanted to. I mean, sure, web browsers are supposed to be all convenient and stuff, making it, like, easier for us to, you know, “surf” the internet. But honestly, it’s just one more thing on my list of things that stress me out.
First off, there’s like a gazillion web browsers out there. Chrome, Firefox, Safari, Edge – it’s like a popularity contest or something. Everyone’s like, “Use this one, it’s the best,” but then someone else is like, “Nah, that one sucks, use this one.” And, like, who even has the energy to care anymore? It’s like trying to pick the coolest outfit to wear, but for your computer. Can’t I just, like, not care and use whatever comes pre-installed?
And then there’s all these updates. Like, seriously, can’t they just leave me alone? Every time I want to, like, chill and watch some random videos, there’s this annoying pop-up telling me to update my browser. And I’m like, “Dude, I just updated you, like, last week. Can you give me a break?” It’s like they’re trying to control my life or something.
And don’t even get me started on cookies and privacy stuff. Like, I get it, they want to, like, track my every move and sell my info to, like, who knows who. But, hello, can a girl get some privacy? It’s like they’re stalking me, following me around, knowing all my secrets. It’s not like I asked for this, you know?
Oh, and the extensions. Like, yeah, I get it, they’re supposed to, like, make my life easier or whatever. But honestly, half the time I don’t even know what they do. It’s like going to a party and not knowing anyone – you’re just there, hoping for the best. And, of course, they slow down my browser, because why not add more misery to my life?
But you know what’s the worst? Those error messages. It’s like the browser is mocking me, laughing at my inability to even load a simple webpage. “Aw, you wanted to check your favorite blog? Too bad, here’s an error message just for you.” It’s like they’re out to get me, to ruin my day, to remind me that technology hates me.
So, yeah, web browsers are just another thing in this never-ending cycle of frustration and confusion. They’re like that friend who always says they’ll be there for you but ends up ghosting you when you really need them. Maybe one day they’ll figure themselves out and actually make my life better. But until then, I’ll just be here, rolling my eyes every time I open a new tab. Whatever, man.
People are having sex in self-driving cars, apparently. Link
Top physicist says chatbots are just ‘glorified tape recorders’ Link
Spain are just one step away from Women’s World Cup glory after their dramatic 2-1 semi-final win over Sweden on Tuesday and forward Salma Paralluelo said ‘La Roja’ had proved they have what it takes to be crowned champions.
LLM (Large Language Model) is a type of AI model designed to understand and generate human-like text. These models are trained on vast amounts of text data and use deep learning techniques, such as deep neural networks, to process and generate language.
LLMs are capable of performing various natural language processing (NLP) tasks, including
Language translation
Text summarization
Question-answering
Sentiment analysis
Generating coherent and contextually relevant responses to user inputs
They are trained on a wide range of textual data sources, such as books, articles, websites, and other written content, allowing them to learn grammar, vocabulary, and contextual relationships in language.
Examples of Large Language Models
Some of the most popular large language models are:
GPT-3 by OpenAI: GPT-3 is a large language model that was first released in 2020. It has been trained on a massive dataset of text and code, and it can generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way.
T5 by Google AI: T5 is a large language model that was first released in 2021. It is specifically designed for text generation tasks, and it can generate text that is more accurate, consistent, and creative than smaller language models.
LaMDA by Google AI: LaMDA is a large language model that was first released in 2022. It is specifically designed for dialogue applications, and it can hold natural-language conversations with users.
PaLM by Google AI: PaLM is a large language model that was first released in 2022. It is the largest and most powerful language model ever created, and it can perform a wide range of tasks, including text generation, translation, summarization, and question-answering.
FlaxGPT by DeepMind: FlaxGPT is a large language model that was first released in 2022. It is based on the Transformer architecture, and it can generate text that is more accurate and consistent than smaller language models.
Large language models (LLMs) have a number of advantages over traditional machine learning models. These advantages include:
Improved accuracy and performance: LLMs can be trained on massive datasets of text and code, which allows them to learn the nuances of human language and generate more accurate and consistent results than traditional machine-learning models.
Increased efficiency: LLMs can automate many tasks that were previously done manually, such as text classification, summarization, and translation. This can save businesses time and money, and free up human workers to focus on more creative and strategic tasks.
Expanded possibilities: LLMs can be used to create new and innovative products and services. For example, they can be used to develop chatbots that can hold natural-language conversations with customers or to create virtual assistants that can help users with tasks such as scheduling appointments or finding information.
Enhanced creativity: LLMs can be used to generate creative text formats, such as poems, code, scripts, musical pieces, emails, letters, and more with endless possibilities. This can be used to improve the quality of content or to create new and innovative forms of art and entertainment.
Reduced bias: LLMs can be trained on datasets that are more diverse than traditional datasets, which can help to reduce bias in their results. This is important for businesses and organizations that want to ensure that their products and services are fair and equitable for all users.
Challenges of LLM
Large language models (LLMs) are a powerful new technology, but they also come with several challenges. These challenges include:
Data requirements: LLMs require massive datasets of text and code to train. This can be a challenge for businesses and organizations that do not have access to large datasets.
Computational resources: LLMs require a lot of computational resources to train and run. This can be a challenge for businesses and organizations that lack the necessary resources.
Interpretability: LLMs are often difficult to interpret. This makes it difficult to understand how they work and to ensure that they are not generating harmful or biased results.
Bias: LLMs can be biased, depending on the data they are trained on. This can be a challenge for businesses and organizations that have ensured that their products and services are fair and equitable for all users.
Safety: LLMs can be used to generate harmful or misleading content. This can be challenging for businesses and organizations having a reputation for safe and secure services.
Use cases of LLM
The future of LLM models is bright. As this technology continues to develop, we can expect to see even more innovative and groundbreaking applications for LLMs in the future.
Some of the promising applications of LLMs include:
Virtual Assistants: LLMs could be used to power virtual assistants that are even more human-like and helpful than they are today. These virtual assistants could be used to provide a wide range of services, such as scheduling appointments, finding information, and controlling smart home devices.
Content Generation: LLMs could be used to generate more engaging and informative content. This content could be used to improve the customer experience, educate users, and entertain people.
Translation: LLMs could be used to translate text from one language to another more accurately and efficiently than ever before. This could help businesses to reach a wider audience and to provide better customer service.
Research: LLMs could be used to conduct research in a wider range of fields, such as natural language processing, machine translation, and artificial intelligence. This could help to advance our understanding of these fields and to develop new and innovative applications.
Education: LLMs could be used to create personalized learning experiences for students. These experiences could be tailored to each student’s individual needs and interests.
Healthcare: LLMs could be used to diagnose diseases, develop new treatments, and provide personalized care to patients.
Art and entertainment: LLMs could be used to create new forms of art and entertainment. This could include poems, code, scripts, musical pieces, emails, letters, etc.
Now that we have gone through the examples of Large Language Models, let us see how to utilize an LLM Library in different use cases along with code build. The LLM library used is provided by Hugging Face, called Transformer Library.
Introducing the Transformer Library
The transformer package, provided by huggingface.io, tries to solve the various challenges we face in the NLP field. It provides pre-trained models, tokenizers, configs, various APIs, ready-made pipelines for our inference, etc.
It is a large language model (LLM) developed by Hugging Face and a community of over 1000 researchers. It is trained on a massive dataset of text and code, and it can generate text, translate languages, and answer questions. Here we are going to see the following application of the Transformer Library:
Sentiment Analysis
Named Entity Recognition
Text Generation
Translate language
Question Answering Pipeline
Summarization
Before jumping to the examples of Transformer Library, we need to install the library to use it.
Install the Transformer Library
pip install transformers
By using the pipeline feature of the Transformers Library, you can easily apply LLMs for text generation, question answering, sentiment analysis, named entity recognition, translation, and more.
from transformers import pipeline
Example: Question Answering Pipeline
To perform question-answering using the Transformers library, you can utilize the pipeline feature with a pre-trained question-answering model. Here’s an example:
from transformers import pipeline
# Define the list of file paths
file_paths = ['document1.txt', 'document2.txt', 'document3.txt']
# Read the contents of each file and store them in a list
documents = []
for file_path in file_paths:
with open(file_path, 'r') as file:
document = file.read()
documents.append(document)
# Concatenate the documents using a newline character
context = "\n".join(documents)
# Use the pipeline with the updated context
nlp = pipeline("question-answering")
result = nlp(question="When did Mars Mission Launched?", context=context)
print(result['answer'])
The code prints the below output correctly to the question – When did Mars Mission Launch?
Output - 5 November 2013
IBM’s AI chip mimics the human brain
The human brain can achieve remarkable performance while consuming little power. IBM’s new prototype chip works similarly to connections in human brains. Thus, it could make AI more energy efficient and less battery draining for devices like smartphones.
The chip is primarily analogue but also has digital elements, which makes it easier to put into existing AI systems.
It addresses the concerns raised about emissions from warehouses full of computers powering AI systems. It could also cut the water needed to cool power-hungry data centers.
Why does this matter?
The advancements suggest the emergence of brain-like chips in the near future. It would mean large and more complex AI workloads could be executed in low-power or battery-constrained environments, for example, cars, mobile phones, and cameras. It promises new and better AI applications with reduced costs.
NVIDIA’s tool to curate trillion-token datasets for pretraining LLMs
Most software/tools made to create massive datasets for training LLMs are not publicly released or scalable. This requires LLM developers to build their own tools to curate large language datasets. To meet this growing need, Nvidia has developed and released the NeMo Data Curator– a scalable data-curation tool that enables you to curate trillion-token multilingual datasets for pretraining LLMs. It can scale the following tasks to thousands of compute cores.
The tool curates high-quality data that leads to improved LLM downstream performance and will significantly benefit LLM developers attempting to build pretraining datasets.
Why does this matter?
Apart from improving model downstream performance with high-quality data, applying the above modules to your datasets helps reduce the burden of combing through unstructured data sources. Plus, it can potentially lead to greatly reduced pretraining costs, meaning relatively faster and cheaper development of AI applications.
Trustworthy LLMs: A survey and guideline for evaluating LLMs’ alignment
Ensuring alignment, which refers to making models behave in accordance with human intentions, has become a critical task before deploying LLMs in real-world applications. This new research has proposed a more fine-grained taxonomy of LLM alignment requirements. It not only helps practitioners unpack and understand the dimensions of alignments but also provides actionable guidelines for data collection efforts to develop desirable alignment processes.
It also thoroughly surveys the categories of LLMs that are likely to be crucial to improve their trustworthiness and shows how to build evaluation datasets for alignment accordingly.
Why does this matter?
The proposed framework facilitates a transparent, multi-objective evaluation of LLM trustworthiness. And it enables systematic iteration and deployment of LLMs. For instance, OpenAI has to devote six months to iteratively align GPT-4 before release. Thus, with clear and comprehensive guidance, it can facilitate faster time to market for AI applications that are safe, reliable, and aligned with human values.
Amazon’s push to match Microsoft and Google in generative AI LINK
Amazon is developing proprietary chips, named “Inferentia” and “Trainium,” to rival Nvidia GPUs in terms of training and speeding up generative AI models.
The company’s late entry into the generative AI market has put it in a position of catch-up, with competitors like Microsoft and Google already investing heavily and integrating AI models into their products.
Despite Amazon’s cloud dominance, it aims to differentiate by leveraging its custom silicon capabilities, with Trainium offering significant price-performance improvements, although Nvidia remains dominant for training models.
World first’s mass-produced humanoid robots with AI brains LINK
Chinese start-up Fourier Intelligence showcased its humanoid robot GR-1, capable of walking on two legs at 5km/h carrying a 50kg load, highlighting the potential of bipedal robots.
Fourier originally focused on rehabilitation robotics, but in 2019, it embarked on creating humanoid robots, with GR-1 achieving success after three years of development.
While challenges remain in commercializing humanoid robots, Fourier aims to mass-produce GR-1 by year-end and sees potential applications in elderly care, education, and more.
Microsoft Designer: An AI-powered Canva: a super cool product that I just found!
I just found out about Microsoft Designer, which is an AI-powered tool for creating all types of graphics, from logos to invitations to social media posts. If you like Canva, you should check this out.
Some cool features:
Prompt-to-design: From just a short description, Designer uses DALLE-2 to generate original and editable designs.
Brand-kit: stay on-brand by instantly applying your fonts and color pallets to any design; it an even suggest color combinations.
Other AI tools: suggests hashtags and captions; replace background of an image with your imagination; erase items from an image; auto-fill a section of the image with generated image.
OpenAI is reportedly in “financial trouble” due to the astronomical costs of running ChatGPT, which is losing $700,000 daily. The article states OpenAI may go bankrupt in 2024 but I disagree because of their investment from Microsoft totaling $10B… there’s no way they can spend all of that right? let me know in the comments.
Top talent being poached by rivals like Google and Meta.
GPU shortages hindering ability to train better models.
Increasing Competition
Cheaper open-source models can replace OpenAI’s APIs.
Musk’s xAI working on more right wing biased model.
Chinese firms buying up GPU stockpiles.
With ChatGPT’s massive costs outpacing revenue and problems like declining users and talent loss mounting, OpenAI seems to be in a precarious financial position as competition heats up.
Google appears to be readying new AI-powered tools for ChromeOS (Link)
Zoom rewrites policies to make clear user videos aren’t used to train AI (Link)
Anthropic raises $100M in funding from Korean telco giant SK Telecom (Link)
Modular, AI startup challenging Nvidia, discusses funding at $600M valuation (Link)
California turns to AI to spot wildfires, feeding on video from 1,000+ cameras (Link)
FEC to regulate AI deepfakes in political ads ahead of 2024 election (Link)
AI in Scientific Papers on August 14th, 2023
This research paper has found that LLMs can naturally read docs to learn how to use tools without any training. Instead of showing demonstration, just provide tool documentation. LLMs figured out how to use programs like image generators and video tracking software, without any new training [Link]
This paper analyses and visualises the political bias of major AI language models. ChatGPT and GPT-4 were most left-wing while Meta’s Llama was right-wing [Link]. This type of research is very important and highlights the inherent bias in these models. It’s practically impossible to remove bias also, and we don’t even know what they’ve been trained on. People need to understand, you control the models, you control what people see, especially as AI models are used more frequently and become mainstream
Remember the Westworld style paper with the 25 AI agents living their lives? It’s now open-source. It’s implications in gaming cannot be overstated. Can’t wait to see what comes of this [Link]
MetaGPT is framework using multiple agents to behave as an entire company – engineer, pm, architect etc. It has over 18k stars on github. This specialised for industries and companies will be powerful [Link]
This paper discusses reconstructing images from signals in the brain. Soon we’ll have brain interfaces that could read these signals consistently, maybe map everything you see? Potential is limitless [Link]
Nvidia is partnering with HuggingFace with DGX Cloud platform allowing people to train and tune AI models. They’re offering a “Training Cluster as a Service” which will help companies and individuals build and train models faster than ever [Link]
Stability AI has released their new AI LLM called StableCode. 16k context length and 3b params with other version on the way [Link]
This paper discusses a framework for designing and implementing complex interactions between AI systems called Flows [Link] Will be very important when building complex AI software in industry. Github will be uploaded soon [Link]
Nvidia announced that Adobe Firefly models will be available as APIs in Omniverse [Link] This thread breaks down what the Omniverse will look like [Link]
Anthropic CEO Dario Amodei thinks AI will reach educated levels of humans in 2-3 years [Link] For reference, Claude 2 is probably the second most powerful model alongside GPT4
Layerbrain is building AI agents that can be used across Stripe, Hubspot and slack using plain english [Link] Looks very cool
LLMs picking random numbers almost always pick the numbers 6-8 [Link]
Inflection founder Mustafa Suleyman says we’ll probably rely on LLMs more than the best trained and most experienced humans within 5 years [Link]. For context, Mustafa is one of the co founders of Google DeepMind – this guys knows AI
Writer, a startup using Nvidia’s NeMo discuss how it helped them build and scale over 10 models. NeMo isn’t publicly available but seems like a massive advantage considering Writer’s cloud infra, which is managed by 2 people, hosts a trillion API calls a month [Link] Link to NeMo [Link] Link to NeMo guardrails blog [Link]
Someone open-sourced smol-podcaster – it transcribes and labels speakers, formats the transcription, creates chapters with timestamps [Link]
Ultra realistic AI generated videos are coming. It’s impossible to tell they’re fake now [Link] Signup for early access here [Link]
Anthropic released Claude Instant 1.2. Its very fast, better at math and coding and hallucinates less [Link]
This guy released the code for his modded Google Nest Mini using OpenAI’s function calling to take notes and control his lights. Once Amazon & Apple integrates better LLMs into their prods, AI will truly be everywhere [Link]
If you search “As an AI language model” in Google Scholar a lot of papers come up… [Link]
OpenAI released custom instructions for ChatGPT free users, except for people in the US or UK [Link]
OpenAI, Google, Microsoft and Anthropic partnered with Darpa for their AI cyber challenge [Link]
PlayHT released their new text-to-voice ai model and it looks crazy good. Change the way its delivered by describing an emotion and much more [Link] [Link]
A paper by Google showcasing that AI models tend to repeat a user’s opinion back to them, even if its wrong. Thread breaking it down [Link] Link to paper [Link]
Medisearch comes out of YC and claims to have the best model for medical questions [Link]
Someone made a way to one-click install AudioLDM with gradio web ui [Link]
WizardLM released a new math model that outperforms ChatGPT on math skills [Link]
A team of researchers trained an AI model to hear the sounds of keystrokes and steal data. Apparently it has a 95% success rate. Link to article [Link] Link to paper [Link]
Yann LeCun gave a talk at MIT about Objective-Driven AI [Link]
Implement text-to-SQL using langchain, a breakdown[Link]
SDXL implemented in 520 lines of code in a single file [Link]
OpenAI released a blog on Special Projects – one of them involved trying to find secret breakthroughs in the world [Link]
Google announced Project IDX, a browser-based code environment. Brings app dev to the cloud and has AI features like code gen, completion etc [Link] A shot at replit it seems
Meta open-sourced AudioCraft – musicgen, audiogen and encodec. Definitely worth checking out [Link]
If you’re interested in fine-tuning open-source models like Llama-2, definitely check out this blog [Link] In some cases, fine-tuned llama2 is better than gpt4 (for sql generation for example). Overall a great read if you’re interested in fine tuning
Nvidia released the code for Neuralangelo, an AI model that reconstructs 3d surfaces from 2d videos [Link]
Create digital environments in seconds with Blockade labs. Wild stuff [Link]
This paper compares the answers of ChatGPT and stackoverflow for software engineering questions [Link] “52% of chatgpt answers are incorrect and 77% are verbose but are still preferred 39% of the time due to their comprehensiveness and well-articulated language style”. Only issue is this uses 3.5. Need this test with gpt4
Latest Tech News and Trends on August 14th, 2023
Privacy win: Starting today Facebook must pay $100.000 to Norway each day for violating our right to privacy. Link
College professors are going back to paper exams and handwritten essays to fight students using ChatGPT. Link
New Footage Shows Tesla On Autopilot Crashing Into Police Car After Alerting Driver 150 Times. Link
IBM’s prototype brain-like chip promises efficient, greener AI – The human brain can achieve remarkable performance while consuming little power. IBM’s new prototype chip works similarly to connections in human brains. Thus, it could make AI more energy efficient and less battery draining for devices like smartphones. The chip is primarily analogue but also has digital elements, which makes it easier to put into existing AI systems.
NVIDIA’s tool to curate trillion-token datasets for pretraining LLMs – To meet the growing demands for curating pretraining datasets for LLMs, Nvidia has released Data Curator as part of the NeMo framework. It is a scalable data-curation tool that enables you to curate trillion-token multilingual datasets for pretraining LLMs. It also curates high-quality data that leads to improved LLM downstream performance and will significantly benefit LLM developers attempting to build pretraining datasets.
Trustworthy LLMs: A survey and guideline for evaluating LLMs’ alignment – New research has surveyed the categories of LLMs that are likely to be important for practitioners to focus on in order to improve LLMs’ trustworthiness. It explains in detail how to evaluate an LLM’s trustworthiness according to the above categories and build evaluation datasets for alignment accordingly in a more fine-grained manner.
ChromeOS might get some new AI-powered tools – Google appears to be readying an AI writing tool for ChromeOS. Its code has hints of some AI tools for suggestions and rewrites.
Zoom rewrites policies to make clear your videos aren’t used to train AI tools – Zoom has updated its terms of service and reworded a blog post explaining the recent changes. The company now explicitly states that “communications-like” customer data isn’t being used to train AI models for Zoom or third parties.
Anthropic raises $100M from Korean telco giant SK Telecom – They plan to co-develop a multilingual LLM customized for global telco firms.
Modular, AI startup challenging Nvidia, to be valued at $600M – It is to raise Series A funding that would value it at roughly $600 million. Nvidia makes Cuda, the dominant software for writing ML apps that works only with Nvidia chips. Modular’s software aims to make it easier for AI developers to train and run their ML models on chips designed by other companies, including AMD, Intel, and Google.
AI avatars are coming. In my mind the biggest market for this might be content creators. People who need to appear on video and are tired of ensuring pitch perfect recordings.
Amazon is introducing Amazon One, a biometric hand-scanning service that allows users to pay at Whole Foods, Amazon Fresh stores, Panera restaurants, airports, stadiums, and Starbucks locations using their palm.
This move is part of Amazon’s effort to compete with Google and Apple in the digital wallet space, aiming to create a universal identity provider that goes beyond payments, potentially connecting to various services, including health records.
Amazon One uses near-infrared light to capture palm vein patterns and surface features, with a focus on security through encrypted hand scan transmission, but it faces privacy concerns and the challenge of convincing merchants to adopt the technology.
The California Department of Forestry and Fire Protection (Cal Fire) has launched the Alert California AI program in collaboration with UCSD, using AI and 360-degree cameras to detect potential wildfires by identifying abnormalities in camera feeds.
The program successfully detected and prevented a fledgling fire in the Cleveland National Forest, alerting firefighters who extinguished the flames within 45 minutes.
Alert California utilizes LiDAR scans and machine learning to differentiate between smoke and other particles, aiming to combat wildfires in the face of extreme climate conditions.
White House’s $1.2B carbon capture initiative LINK
The Department of Energy is providing grants of up to $1.2 billion to two direct air capture (DAC) projects aiming to remove over 2 million metric tons of CO2 annually, equivalent to emissions from 445,000 gas-powered cars.
The DAC projects in Texas and Louisiana, supported by the Regional Direct Air Capture Hubs program, will create jobs and could potentially remove up to 30 million tons of CO2 per year, contributing to the US goal of emissions neutrality by 2050.
The DOE aims to lower DAC costs below $100 per metric ton of CO2-equivalent and is funding feasibility studies, engineering projects, and a carbon removal credits program to achieve global impact on carbon reduction.
Sam Bankman-Fried, former CEO of FTX, had his bail revoked ahead of his trial following allegations of leaking a diary to the New York Times.
Bankman-Fried faces charges including defrauding FTX investors and was initially under house arrest on a $250 million bond.
US District Court Judge revoked his bail due to alleged misconduct and possible witness intimidation, leading to potential detention at a detention center during trial.
AI can now outperform humans in Captcha tests LINK
A study reveals that humans are slower and less accurate than bots in solving Captcha tests, raising questions about their effectiveness.
Captchas are intended to deter bots from accessing services, preventing malicious activities like DDoS attacks and spam accounts.
Bots can outperform humans in solving certain types of Captchas, indicating an ongoing challenge in maintaining their efficacy.
Not only are bots faster at solving captchas than humans they’re also more accurate.
Unraveling August 2023: August 12th, 2023
Latest AI News and Trends on August 12th 2023: Week Recap
Apple bets big on AI
Apple’s CEO Tim Cook has stated that AI and ML are embedded in every company product. This comes after concerns were raised about Apple’s lack of discussion on its AI plans while competitors have been actively incorporating the technology into their products. He also emphasized that AI is central to the design of Apple’s products, contradicting suggestions that the company has not yet integrated the technology.
Cook reassured that Apple has invested in AI for years, and this year’s Research & Development spending has hit $22.61 billion. They are also hiring dozens of AI jobs in the US, France, and China, looking to fill roles that could help build Gen AI tools.
Jupyter brings AI to notebooks
Jupyter AI is a tool that brings generative AI to Jupyter notebooks, allowing users to explore and work with AI models. It offers an %%ai magic command that turns the notebook into a reproducible generative AI playground, a native chat UI for working with generative AI as a conversational assistant, and support for various generative model providers.
Jupyter AI is compatible with JupyterLab, with version 1.x supporting JupyterLab 3.x, and version 2.x supporting JupyterLab 4.x. The main branch of Jupyter AI focuses on the newest supported version of JupyterLab, with features and bug fixes backported to JupyterLab 3 if deemed valuable.
ChatGPT’s emotional awareness is more than humans’. What?
A study found that ChatGPT has higher emotional awareness than humans. The machine was subjected to a standardized test measuring human emotional awareness and scored significantly higher. The test required participants to show empathy in fictional scenarios.
ChatGPT outperformed humans in all categories, achieving an overall score of 85 compared to 56 for men and 59 for women. The researchers suggest that ChatGPT could be helpful in psychotherapy, cognitive training, and diagnosing mental illness. Previous studies have shown that people perceive ChatGPT’s responses as more empathetic than medical professionals.
Microsoft’s many AI monetization plans
Microsoft has announced new Azure AI infrastructure advancements and availability to bring its customer closer to the transformative power of generative AI.
Azure OpenAI Service goes global: OpenAI’s most advanced models, including GPT-4 and GPT-35-Turbo, will now be available in multiple new regions and locations.
General availability of ND H100 v5 VMs for unprecedented AI processing and scale: -It also announced general availability of the ND H100 v5 Virtual Machine series, featuring the latest NVIDIA H100 Tensor Core GPUs and low-latency networking, propelling businesses into a new era of AI applications.
OpenAI launches a web crawler to train ChatGPT
Called GPTBot, the crawler will comb through the internet to train and enhance AI’s capabilities. It can be identified by the following user agent and string.
Web pages crawled with the GPTBot user agent may potentially be used to improve future models and are filtered to remove sources that require paywall access, are known to gather personally identifiable information (PII), or have text that violates our policies.
Moreover, OpenAI also revealed how websites can prevent GPTBot from accessing their sites, either partially or by opting out entirely.
AI deep fake audios are getting scarily realistic
Speech deepfakes are artificial voices generated by AI models. While studies investigating human detection capabilities are limited, a new experiment presented genuine and deep fake audio to individuals and asked them to identify the deep fakes. Listeners could correctly spot the deep fakes only 73% of the time.
The experiment was done in English and Mandarin to understand if language affects detection performance and decision-making rationale. However, there was no difference in detectability between the two languages.
NVIDIA’s Biggest AI Breakthroughs
Reveals anew chip GH200
Nvidia announced a new chip GH200, designed to run AI models. It has the same GPU as the H100, Nvidia’s current highest-end AI chip, but pairs it with 141 gigabytes of cutting-edge memory and a 72-core ARM central processor. This processor is designed for the scale-out of the world’s data centers.
The adoption of Universal Scene Description (OpenUSD)
Announced new frameworks, resources, and services to accelerate the adoption of Universal Scene Description (USD), known as OpenUSD. Through its Omniverse platform and a range of technologies and APIs, including ChatUSD and RunUSD, NVIDIA aims to advance the development of OpenUSD, a 3D framework that enables interoperability between software tools and data types for creating virtual worlds.
An AI Workbench
Introduced AI Workbench, a developer toolkit that simplifies creating, testing, and customizing pre-trained generative AI models. The toolkit allows developers to scale these models to various platforms, including PCs, workstations, enterprise data centers, public clouds, and NVIDIA DGX Cloud. This will speed up the adoption of custom generative AI for enterprises worldwide.
The Partnership between NVIDIA and Hugging Face
NVIDIA and Hugging Face have partnered to bring generative AI supercomputing to developers. Integrating NVIDIA DGX Cloud into the Hugging Face platform will accelerate the training and tuning of large language models (LLMs) and make it easier to customize models for various industries. This partnership aims to connect millions of developers to powerful AI tools, enabling them to build advanced AI applications more efficiently.
Google’s AI Surprise for Developers
Project IDX is an experiment by Google to improve full-stack, multi-platform app development. It aims to simplify the complex app development process across mobile, web, and desktop platforms. It is a browser-based development experience built on Google Cloud and powered by Codey, Google’s PaLM 2-based foundation model for programming tasks.
It allows developers to work from anywhere, import existing projects, and preview apps across platforms. It supports frameworks like Angular, Flutter, Next.js, React, Svelte, Vue and languages like JavaScript and Dart. AI capabilities like smart code completion and contextual code actions are also included. Google plans to add support for more languages like Python and Go in the future. Additionally, Project IDX integrates with Firebase hosting for easy deployment of web apps.
Stability AI launches LLM code generator
Stability AI has released StableCode, an LLM generative AI product for coding. It aims to assist programmers in their daily work and provide a learning tool for new developers. StableCode uses three different models to enhance coding efficiency. The base model was trained in various programming languages, including Python, Go, Java, and more. It was then further trained on 560B tokens of code.
The instruction model was tuned for specific use cases by training it on 120,000 code instruction/response pairs. StableCode offers a unique solution for developers to improve their coding skills and productivity.
Anthropic’s Claude Instant 1.2- Faster and safer LLM
Anthropic has released an updated version of Claude Instant, its faster, lower-priced yet very capable model which can handle a range of tasks including casual dialogue, text analysis, summarization, and document comprehension.
Claude Instant 1.2 incorporates the strengths of Claude 2 in real-world use cases and shows significant gains in key areas like math, coding, and reasoning. It generates longer, more structured responses and follows formatting instructions better. It has also made improvements on safety. It hallucinates less and is more resistant to jailbreaks, as shown below.
Google attempts to answer if LLMs generalize or memorize
LLMs can certainly seem like they have a rich understanding of the world, but they might just be regurgitating memorized bits of the enormous amount of text they’ve been trained on. How can we tell if they’re generalizing or memorizing?
In this research, Google examines the training dynamics of a tiny model and reverse engineers the solution it finds – and in the process provides an illustration of the exciting emerging field of mechanistic interpretability. It seems that LLMs start by generalizing reasonably well but then change towards memorizing things.
IBM plans to make Meta’s Llama 2 available on watsonx.ai
IBM will host Llama 2-chat 70B model in the watsonx.ai studio, with early access available to select clients and partners. This will build on IBM’s collaboration with Meta on open innovation for AI, including work with open-source projects developed by Meta. It will also support IBM’s strategy of offering both third-party and its own AI models.
Apple has reportedly ordered servers from Foxconn Industrial Internet, a division of its supplier Foxconn, for the testing and training of AI services. The servers are specifically for Apple’s AI work, which has been a focus for the company for years.
While Apple does not currently have a ChatGPT-like app for external use, it is believed that this division of Foxconn already supplies servers to ChatGPT OpenAI, Nvidia, and Amazon Web Services. The news comes amidst reports about Apple’s plans to compete in the AI chatbot market.
Midjourney is rolling out a GPU cluster upgrade today. Pro and Mega users should see speedups of ~1.5x (/imagine from ~50 sec to ~30 sec). These renders should also be 1.5x cheaper.
They’re releasing V5.3, possibly next week, which will include features like inpainting and a new style (aesthetic) and may be only available on desktop. V6 is also in the works, aiming to enhance performance and language understanding. The website’s frontend is being worked on by a team, and it will be available for both desktop and mobile users. The launch date is approaching, but no specific date has been announced.
MetaGPT is a new framework that improves multi-agent collaboration by incorporating human workflows and domain expertise. It addresses the problem of hallucination in LLMs by encoding Standardized Operating Procedures (SOPs) into prompts, ensuring structured coordination.
The framework also mandates modular outputs, allowing agents to validate outputs and minimize errors. By assigning diverse roles to agents, MetaGPT effectively deconstructs complex problems.
Latest Tech News and Trends on August 12th 2023
Robotaxis greenlit for 24/7 operations in San Francisco LINK
California approved all-day paid robotaxi service in San Francisco, allowing unlimited self-driving car fleets.
The decision came amid objections from San Francisco officials, after a six-hour public comment session, and was a result of applications from Cruise (backed by GM) and Waymo (an Alphabet subsidiary).
Despite some challenges with driverless cars on the city’s streets, Cruise and Waymo see this approval as a pivotal step towards making their investments in self-driving technology profitable.
Russia launches its first lunar mission in 47 years LINK
Russia launches Luna-25, its first lunar mission since 1976, targeting the Moon’s south pole to potentially uncover water ice beneath its surface.
The mission is symbolic, referencing the Soviet Space Program era, and aims to project Russia as an influential world power amidst tensions following its 2022 Ukraine invasion.
Luna-25 is in competition with India’s Chandrayaan-3 mission, with both crafts expected to reach the Moon’s south pole around the same time.
Virgin Galactic debuts with its first civilian spaceflight LINK
Virgin Galactic’s second commercial flight, Galactic 02, took three private citizens to suborbital heights, including a historic mother-daughter duo.
The VSS Unity reached a peak altitude of 55 miles (88 kilometers) in an hour-long flight, with Kelly Latimer becoming the first woman pilot of a commercial spaceflight.
Following recent successes, Virgin Galactic aims for monthly commercial launches and is developing its Delta Class spacecraft for 2026, though substantial revenue from these flights is not anticipated.
Chinese internet giants, in response to US sanctions, are purchasing vast numbers of Nvidia GPUs to bolster their AI capabilities.
Companies like Alibaba, Baidu, ByteDance, and Tencent have reportedly spent around $1 billion on 100,000 Nvidia A800 GPUs, with further orders amounting to an additional $4 billion.
The GPUs are crucial for training large language models, and while the US seeks stricter export limitations on AI tech to China, US companies continue to design specific AI chips for the Chinese market.
Latest Football and Sport News on August 11th 2023
As Australia’s victorious players embarked on a lap of honor, home fans danced and cheered joyously inside a sold out Brisbane Stadium on Saturday. Neither players or supporters had experienced anything like it before.
Cristiano Ronaldo guided nine-man Al-Nassr to their first ever Arab Club Champions Cup title after scoring twice in a 2-1 extra-time win over fellow Saudi side Al-Hilal on Saturday at the King Fahd Stadium.
Alexander Isak’s clinical finishing helped Newcastle United to an emphatic victory against Aston Villa on the opening weekend of the new Premier League campaign.
Unraveling August 2023: August 11th 2023
Latest AI News and Trends on August 11th 2023
AI Tutorial: Applying the 80/20 Rule in Decision-Making with ChatGPT
The Pareto Principle, or the 80/20 rule, is the idea that 80% of results come from 20% of efforts. This concept is integral to many aspects of life, including productivity, business, and personal growth. By embracing this principle with tools like ChatGPT, you can make more efficient decisions and concentrate on what’s most important.
Try the prompt below:
Employing the 80/20 rule, please help me analyze my e-commerce business. I want to know which 20% of my products are generating 80% of my sales and which 20% of my marketing efforts are leading to 80% of my traffic. Additionally, provide insights on how I can optimize my operations based on this principle.
MetaGPT tackling LLM hallucination
MetaGPT is a new framework that improves multi-agent collaboration by incorporating human workflows and domain expertise. It addresses the problem of hallucination in LLMs by encoding Standardized Operating Procedures (SOPs) into prompts, ensuring structured coordination.
The framework also mandates modular outputs, allowing agents to validate outputs and minimize errors. By assigning diverse roles to agents, MetaGPT effectively deconstructs complex problems.
Why does this matter?
Experiments on collaborative software engineering benchmarks show that MetaGPT generates more coherent and correct solutions than chat-based multi-agent systems. And Integrating human knowledge into multi-agent systems opens up new possibilities for tackling real-world challenges.
Summary: The Federal Election Commission (FEC) has initiated a process that may lead to the regulation of AI-generated deepfakes in political ads before the 2024 election, aiming to protect voters against this form of election disinformation. (source)
Key Points:
A petition to regulate ads using AI for mis-representation.
Existing campaigns, including Florida GOP Gov. Ron DeSantis’s, are already using this technology.
The FEC will decide on rules only after a 60-day public comment window, likely starting next week.
Regulation could impose guidelines for disclaimers, but not cover all deepfake-related threats, such as those from individual social media users.
Why It Matters: With elections around the corner, the potential use of AI in misleading political ads is a hot topic. The decision to possibly regulate AI shows an understanding of its possible risks, but the real test will be in getting rules on the books. It’s not just about politics; it’s about truth in a world where seeing is no longer believing.
What Else Is Happening in AI on August 11th 2023
Microsoft introduced new tools for global frontline workers, enhancing their capabilities. (Link)
Google keyboard’s new update could include AI-powered proofreading, AI emojis & more. (Link)
Runway’s new update allows you to extend your Gen-2 videos up to 18 seconds! (Link)
China’s internet giants, including Baidu, TikTok-owner, Alibaba have reportedly ordered $5B worth of Nvidia chips! (Link)
PlayHT2.0 is a new AI model that can “talk”? (Link)
A new AI algorithm has detected a potentially hazardous asteroid that had gone unnoticed by human observers, slated to fly by Earth. The algorithm, HelioLinc3D, was explicitly designed for the Vera Rubin Observatory currently under construction in Northern Chile.[Link]
The U.S. Defense Department has created a task force to evaluate and guide the application of generative artificial intelligence for national security purposes, amid an explosion of public interest in the technology. [Link]
China’s largest web and cloud providers (Alibaba, Baidu, ByteDance, and Tencent)are lining up to buy as many Nvidia GPUs as they can while they still can get their hands on them. [Link]
At Black Hat USA 2023, DARPA issued a call to top computer scientists, AI experts, software developers, and beyond to participate in the AI Cyber Challenge (AIxCC) – a two-year competition aimed at driving innovation at the nexus of AI and cybersecurity to create a new generation of cybersecurity tools. [Link]
Apple is working aggressively on AI – Apple has reportedly ordered servers from Foxconn Industrial Internet, a division of its supplier Foxconn, for the testing and training of AI services. The servers are specifically for Apple’s AI work, which has been a focus for the company for years. While Apple does not currently have a ChatGPT-like app for external use, it is believed that this division of Foxconn already supplies servers to ChatGPT OpenAI, Nvidia, and Amazon Web Services. The news comes amidst reports about Apple’s plans to compete in the AI chatbot market.
Midjourney’s future plans revealed – They’re rolling out a GPU cluster upgrade today. Pro and Mega users should see speedups of ~1.5x (/imagine from ~50 sec to ~30 sec). These renders should also be 1.5x cheaper. – They’re releasing V5.3 possibly next week, will include features like inpainting and a new style (aesthetic) and may be only available on desktop.
Microsoft introduced new tools for global frontline workers, enhancing their capabilities – The company’s Copilot offering utilizes generative AI to enhance the efficiency of service professionals. Microsoft highlights the significant size of the frontline workforce, estimating it to be 2.7 billion globally. The new tools and integrations are designed to empower these workers and address labor challenges faced by businesses.
Google keyboard’s new update could include AI-powered proofreading, AI emojis & more – Google is enhancing its Gboard keyboard with new features powered by AI. These features include AI emojis, proofreading, and a drag mode that allows users to resize the keyboard to their liking. The updates have been discovered in the latest beta version of Gboard.
PlayHT2.0 is a new AI model that can “talk” – It has an Instant Voice Cloning capability that can capture any voice and accent from just 3s of a speaker’s voice and synthesize speech in a truly conversational tone. – Trained on over a million hours of speech across multiple languages, accents, and speaking styles.
Runway’s new update allows you to extend your Gen-2 videos up to 18 seconds. – Available now in the browser and coming soon to iOS.
China’s internet giants, including Baidu, TikTok-owner ByteDance, Tencent, and Alibaba, have reportedly ordered $5 billion worth of Nvidia chips to power their AI ambitions. The orders, totaling about 100,000 A800 processors, are crucial for building generative AI systems. The chips are expected to be delivered this year. This move highlights China’s growing focus on AI technology and its desire to become a global leader in the field.
TikTok is reportedly adding a toggle that enables creators to label AI-generated content, aiming to prevent content removal and enhance transparency.
Belva: Empower an AI agent to manage your phone calls effectively—an ideal solution for call management optimization.
Broadcast: Streamline the drafting and distribution of weekly updates using this AI-automated tool. It offers collaboration features, readership insights, and workflow optimization across platforms like Slack and Email.
Zefi: Enhance your product development process with this AI tool, integrating with development platforms to gather data, cluster feedback, assist in prioritization, and align stakeholders.
YT Transcripts by Editby: Download and edit YouTube videos easily with this tool, making it perfect for content creators seeking to repurpose their YouTube content.
AI Tools Database: Explore a comprehensive Notion database featuring 1350 useful AI tools curated by The Intelligo.
Sick of notifications lighting up your Android screen? You can prevent this in your app settings. This guide explains how.
Usually, a notification will buzz on your phone or beep at you while displaying on the screen to be noticed. However, this behavior can drain your battery faster and become annoying in general to deal with. You can turn off an app’s notification behavior in your device’s settings.
There isn’t a universal setting to prevent all apps from waking the lock screen, so you’ll need to manage them individually. Here’s how.
How to Disable App Wake Screen Settings on Android
Unless you enable Airplane Mode or turn on your device’s Do Not Disturb option, apps will continue to wake your screen by default. So, you need to manage each app you want to stop notifications from tediously.
To stop notifications from turning on the screen on Android:
Swipe down from the top of the screen and tap Settings (gear icon) in the top-right corner.
Select the Notifications option from the Settings menu.
Tap the App notifications option to view your complete list of installed apps.
Select the app that you don’t want to wake your screen.
Tap the Silent option under the Alerts section.
You can also disable all app notifications by toggling off the Allow notifications switch. You won’t have access to the notification settings for all apps when you turn this off, however.
It’s also important to note that some apps will allow you to manage specific notifications by selecting the Notification categories option and toggling individual notification types on or off.
How to Use In-App Settings to Stop Apps Waking Your Android Screen
Depending on the app, you may be able to stop app notifications from turning on the screen from within the app itself. For example, in the Snapchat app’s settings menu, you can turn off the Wake Screen option for notifications that’s enabled by default.
Following last month’s announcement, Google is rolling out the ‘Unknown Tracker Alerts’ feature to Android users. Thanks to this alert notification, Android users will know if someone has placed an unknown tracker on them. Also, the unknown tracker alert will be sent when someone else’s tracking device is separated from them and is found to be traveling with another person.
Contact Posters offer an engaging canvas to express creativity via otherwise mundane-looking contact cards. Here’s how to create them in iOS 17 on your iPhone.
he whole process of creating a Contact Poster is fairly easy. You can create a Contact Poster for your own number, or any other acquaintance in your contact list.
For this guide, we’ll go through the step-by-step process of creating a contact poster for fellow Digital Trends contributor Tushar Mehta. The process is identical if you are creating a contact poster for yourself. To do that, just tap on your name when it appears at the top of the contacts list in the Phone app.
Step 1: Open the Phone app on your iPhone and select the contact that needs a poster makeover. As you tap on a name, it will open the detailed contact page as shown in the image below.
Step 2: On the contact page, tap on the Edit button in the top-right corner of the screen. On the next page, either tap on the circle with the contact name initials, or the pill-shaped Add Photo button.
The US just invested more than $1 billion into carbon removal / The move represents a big step in the effort to suck CO2 out of the atmosphere—and slow down climate change. Link
Latest World USA Sport News on August 11th 2023
Orange juice prices to surge as US crops ravaged by disease and climate. Link
Teenage girl dies after being forced to stay in a ‘period hut’ in Nepal. Link
Nearly 50,000 Americans died by suicide in 2022, a record-high number: CDC. Link
Supreme Court blocks OxyContin maker’s bankruptcy deal that would shield Sackler family members. Link
New school bus routes a ‘disaster,’ Kentucky superintendent admits. Last kids got home at 10 pm. Link
2 minutes daily football news: Spain 2-1 Ned; Japan 1 – 2 Sweden; Harry Kane Caicedo; #soccer #footy
Liverpool have agreed a British record transfer fee of £111m with Brighton for midfielder Moises Caicedo.
England captain Harry Kane is set to have a medical at Bayern Munich after being given permission to travel to Germany by Tottenham.
Sweden produced a magnificent performance to book a semi-final date with Spain and leave Japan’s Women’s World Cup dreams in tatters
Teenage winger Salma Paralluelo came off the bench to score a 111th-minute winner as Spain beat the Netherlands to reach the Women’s World Cup semi-finals for the first time.
Off the pitch, few teams at this Women’s World Cup have been as dysfunctional and wracked by controversy as Spain.
Soccer Football Saudi Pro League kicks off after raiding Europe’s top football clubs.
Unraveling August 2023: August 10th 2023
Latest AI News and Trends on August 10th 2023
Advanced Library of 1000+ free GPT Workflows with HeroML – To Replace most “AI” Apps. Byu/papsamir
Disclaimer: all links below are free, no ads, no sign-up required for open-source solution & no donation button. Workflow software is not only free, but open-source ❣️
This post is longer than I anticipated, but I think it’s really important and I’ve tried to add as many screenshots and videos to make it easier to understand.I just don’t want to pay for any more $9 a month chatgpt wrappers.And I don’t think you do either..
Lots of folks were saying that one prompt alone cannot give you the quality you expect, so I kept experimenting and over the last 3 months of insane keyboard-tapping, I deduced a conversational-type experience is always the best.
I wanted to have these conversations, though, without actually having them... I wanted to automate the conversations I was already having on ChatGPT!
There was no solution, nor a free alternative to the giants (and the lesser giants who I know will disappear after the AI hype dies off), so I went ahead and made an OPEN-SOURCE (meaning free, and meaning you can see how it was made) solution called HeroML.
It’s essentially prompts chained together, and prompts that can reference previous responses for ❣️ context ❣️
There reason I wanted to make something like this is because I was seeing a lot of startups, for the lack of a better word, coming up with priced subscriptions to apps that do nothing more than chain a few prompts together, naturally providing more value than manually using ChatGPT, but ultimately denying you any customization of the workflow.
Let’s say you wanted to generate… an email! Here’s what that would look like in HeroML:
(BTW, each step is separated by ->>>>, so every time you see that, assume a new step has begun,the below example has 4 steps*)*
You are an email copywriter, write a short, 2 sentence email introduction intended for {{recipient}} and make sure to focus on {{focus_point_1}} and {{focus_point_2}}. You are writing from the perspective of me, {{your_name}}. Make sure this introduction is brief and do not exceed 2 sentences, as it's the introduction.
->>>>
Your task is to write the body of our email, intended for {{recipient}} and written by me, {{your_name}}. We're focusing on {{focus_point_1}} and {{focus_point_2}}. We already have the introduction:
Introduction:
{{step_1}}
Following on, write a short paragraph about {{focus_point_1}}, and make sure you adhere to the same tone as the introduction.
->>>>
Your task is to write the body of our email, intended for the recipient, "{{recipient}}" and written by me, {{your_name}}. We're focusing on {{focus_point_1}} and {{focus_point_2}}. We already have the introduction:
Introduction:
{{step_1}}
And also, we have a paragraph about {{focus_point_1}}:
{{step_2}}
Now, write a short paragraph about {{focus_point_2}}, and make sure you adhere to the same tone as the introduction and the first paragraph.
->>>>
Your task is to write the body of our email, intended for {{recipient}} and written by me, {{your_name}}. We're focusing on {{focus_point_1}} and {{focus_point_2}}. We already have the introduction:
Introduction:
{{step_1}}
We also have the entire body of our email, 2 paragraphs, for {{focus_point_1}} & {{focus_point_2}} respectively:
First paragraph:
{{step_2}}
Second paragraph:
{{step_3}}
Your final task is to write a short conclusion the ends the email with a "thank you" to the recipient, {{recipient}}, and includes a CTA (Call to action) that requires them to reply back to learn more about {{focus_point_1}} or {{focus_point_2}}. End the conclusion with "Wonderful and Amazing Regards, {{your_name}}
It may seem like this is a lot of text, and that you could generate this in one prompt in ChatGPT, and that’s… true! This is just for examples-sake, and in the real-world, you could have 100 steps, instead of the four steps above, to generate anything where you can reuse both dynamic variables AND previous responses to keep context longer than ChatGPT.
For example, you could have a workflow with 100 steps, each generating hundreds (or thousands) of words, and in the 100th step, refer back to {{step_21}}. This is a ridiculous example, but just wanted to explain what is possible.
I’ll do a quick deep dive into the above example.
You can see I use a bunch of dynamic variables with the double curly brackets, there are 2 types:
Variables that you define in the first prompt, and can refer to throughout the rest of the steps
{{your_name}}, {{focus_point_1}}, etc.
Step Variables, which are basically just variables that references responses from previous steps..
{{step_1}} can be used in Step #2, to input the AI response from Step 1, and so on.
In the above example, we generate an introduction in Step 1, and then, in Step 2, we tell the AI that "We have already generated an introduction: {{step_1}}"
When you run HeroML, it won’t actually see these variables (the double-curly brackets), it will always replace them with the real values, just like the example in the video above!
Please don’t hesitate to ask any questions, about HeroML or anything else in relation to this.
Free Library of HeroML Workflows
I have spent thousands of dollars (from OpenAI Grant money, so do not worry, this did not make me broke) to test and create a tonne (over 1000+) workflows & examples for most industries (even ridiculous ones). They too are open-source, and can be found here:
However, the Repo allows you or any contributor to make changes to these workflows (the .heroml) files, and when those changes are approved, they will automatically be merged online.
There are thousands of workflows in the Repo, but they are just examples. The best workflows are ones you create for your specific needs.
How to run HeroML
Online Playground
There are currently two ways to run HeroML, the first one is running it on Hero, for example, if you want to run the blog post example I linked above, you would simply fill out the dynamic variables, here:
0:00
0:10
Example of hero app playground
This method has a setback, it’s free (if you keep making new accounts so you don’t have to pay), and the model is gpt-3.5 turbo.. I’m thinking of either adding GPT4, OR allow you to use your OWN OpenAI keys, that’s up to you.
Also, I’m rate limited because I don’t have any friends in OpenAI, so the API token I’m using is very restricted, why might mean if a bunch of you try, it won’t work too well, which is why for now, I recommend the HeroML CLI (in your terminal), since you can use your own token! (I recommend GPT-4)
My favorite method is the one below, since you have full control.
Local Machine with own OpenAI Key
I have built a HeroML compiler in Node.js that you can run in your terminal. This page has a bunch of documentation.
Running HeroML example and Output
Here’s an example of how to run it and what do expect.
This is the script
simple HeroML script to generate colors, and then people’s names for each color.
This is how quick it is to run these scripts (based on how many steps):
0:00
0:19
using HeroML CLI with your own OpenAI Key
And this is the output (In markdown) that it will generate. (it will also generate a structured JSON if you want to clone the whole repo and build a custom solution)
Output in markdown, first line is response of first step, and then the list is response from second step. You can get desired output by writing better prompts 😊
Conclusion
Okay, that was a hefty post. I’m not sure if you guys will care about a solution like this, but I’m confident that it’s one of the better alternatives to what seems to be an AI-rug pull. I very much doubt that most of these “new AI” apps will survive very long if they don’t allow workflow customization, and if they don’t make those workflows transparent.
I also understand that the audience here is split between technical and non-technical, so as explained above, there are both technical examples, and non-technical deployed playgrounds.
Github Workflow Link is where to clone the app, or make edits to the workflow for the community.
Deployed Hero Playground is where you can view the deployed version of the link, and test it out. This is restricted to GPT3.5 Turbo, I’m considering allowing you to use your own tokens, would love to know if you’d like this solution instead of using the Hero CLI, so you can share and edit responses online.
Yes, I generated all the names with AI ✨, who wouldn’t?
Thank you for all your support in my last few posts ❣️
I’ve worked pretty exclusively on this project for the last 2 months, and hope that it’s at least helpful to a handful of people. I built it so that even If I disappear tomorrow, it can still be built upon and contributed to by others. Someone even made a python compiler for those who want to use python!
I’m happy to answer questions, make tutorial videos, write more documentation, or fricken stream and make live scripts based on what you guys want to see. I’m obviously overly obsessed with this, and hope you’ve enjoyed this post!
This project is young, the workflows are new and basic, but I won’t pretend to be a professional in all of these industries,but you may be*! So your contribution to these workflows (whichever whose industries you are proficient in) are what can make them unbelievably useful for someone else.*
Have a wonderful day, and open-source all the friggin way 😇
Generative AI tools like ChatGPT is helping workers make more money by automating time-consuming tasks and creating new income streams and full-time jobs.
Universal Music collaborates with Google on AI song licensing LINK
Universal Music Group is negotiating with Google to license artists’ voices and melodies for AI-generated songs, with Warner Music also participating.
Artists could opt out of the system, but the move could allow fans to create deepfakes of their favorite musicians.
While this might be lucrative for record labels, it poses challenges for artists who want to keep their voices free from AI-cloning.
AI’s role in reducing airlines’ contrail climate impact LINK
Contrails from airplanes trap heat in Earth’s atmosphere, leading to a net warming effect.
Pilots at American used Google’s AI predictions and Breakthrough Energy’s models to choose altitudes less likely to produce contrails.
After 70 test flights, satellite imagery revealed a 54% reduction in contrails, suggesting commercial flights can lessen their environmental impact.
Anthropic’s Claude Instant 1.2- Faster and safer LLM
Anthropic has released an updated version of Claude Instant, its faster, lower-priced yet very capable model which can handle a range of tasks including casual dialogue, text analysis, summarization, and document comprehension.
Claude Instant 1.2 incorporates the strengths of Claude 2 in real-world use cases and shows significant gains in key areas like math, coding, and reasoning. It generates longer, more structured responses and follows formatting instructions better. It has also made improvements on safety. It hallucinates less and is more resistant to jailbreaks, as shown below.
Why does this matter?
It looks like Claude Instant 1.2 is Anthropic’s safest AI model. However, it is an entry-level model intended to compete with similar offerings from OpenAI as well as startups such as Cohere. But with enhanced safety, skills, and context length same as Claude 2 (100K tokens), it can perhaps bring Anthropic a step closer to knowing how to challenge ChatGPT’s supremacy.
Google attempts to answer if LLMs generalize or memorize
LLMs can certainly seem like they have a rich understanding of the world, but they might just be regurgitating memorized bits of the enormous amount of text they’ve been trained on. How can we tell if they’re generalizing or memorizing?
In this research, Google examines the training dynamics of a tiny model and reverse engineers the solution it finds – and in the process provides an illustration of the exciting emerging field of mechanistic interpretability. It seems that LLMs start by generalizing reasonably well but then change towards memorizing things.
Why does this matter?
While there is no definitive conclusion from the research, it highlights the somewhat mysterious behavior of deep learning models, especially around the balance between memorization and generalization. It is also one step closer to understanding the exact dynamics of when and why certain models transition between these (and possibly back again).
Teams compete to best secure vital software systems from cyber risks.
Up to 20 teams advance from qualifiers to win $2 million each at DEF CON 2024.
Finalists eligible for more prizes, including $4 million top prize at DEF CON 2025.
Innovating Cybersecurity with AI
Competitors required to open source their AI systems for widespread use.
Collaboration from AI leaders like Anthropic, Google, Microsoft, and OpenAI.
Aims to push boundaries of AI for national cyber defense.
Previous Government Hacking Contests
Similar to 2014 DARPA Cyber Grand Challenge to develop automated cybersecurity.
Various prizes offered to drive innovation through competition.
Hopes AI can keep defense ahead of evolving threats.
The U.S. launched a $20 million AI hacking challenge to incentivize developing AI cybersecurity to protect critical infrastructure. It aims to push AI capabilities for national defense through collaboration and competition.
What Else Is Happening in AI on August 10th 2023
Amazon is testing a tool that uses AI to help sellers write descriptions for listings Link
Spotify and Patreon integrated, allowing Patreon-exclusive audio on Spotify, benefiting podcasters and sidestepping Spotify’s aversion to RSS feeds. LINK
National-level data doesn’t support negative wellbeing impacts of Facebook saturation, but overlooks specific vulnerable groups and children. LINK
Lyft aims to eliminate surge pricing due to abundant driver supply and rider dissatisfaction, resulting in reduced revenue but increased user numbers. LINK
AI-generated books falsely using Jane Friedman’s name surfaced on Amazon and Goodreads, sparking concerns over copyright and author identity verification. LINK
DARPA’s AI Cyber Challenge, supported by top tech firms, aims to enhance software security using AI, focusing on open source vulnerabilities and cyberdefense. LINK
Google research attempts to answer whether ML models memorize or generalize – While LLMs appear to have a rich understanding of the world, how do we know they’re not simply regurgitating from training data? In this new research, Google explores the phenomenon called grokking to learn more about how models learn.
IBM plans to make Meta’s Llama 2 available within its watsonx – It will host Llama 2-chat 70B model in the watsonx.ai studio, with early access available to select clients and partners. This will build on IBM’s collaboration with Meta on open innovation for AI, including work with open-source projects developed by Meta. This will also support IBM’s strategy of offering both third-party and its own AI models.
Amazon is testing a tool that uses AI to help sellers write product descriptions – This will be one of the first examples of Amazon integrating LLMs into its e-commerce business.
White House launches AI-based contest to secure government systems from hacks – It has launched a $27M cyber contest to spur the use of AI to find and fix security flaws in the US government infrastructure in the face of growing use of the technology by hackers for malicious purposes.
Microsoft partners with Aptos blockchain to marry AI and web3 – The collaboration allows Microsoft’s AI models to be trained using Aptos’ verified blockchain information.
OpenAI has a new update for free ChatGPT users – Custom instructions are now available to ChatGPT users on the free plan, except for in the EU & UK, where it will be rolling out soon.
Google’s redesigned Arts & Culture app includes AI-based features – A “Poem Postcards” feature that lets users send AI-generated postcards to friends. Other features include a new Play tab, a TikTok-like “Inspire” feed, and more.
Latest Tech News and Trends on August 10th 2023
A.I. can identify keystrokes by just the sound of your typing and steal information with 95% accuracy, new research shows. Researchers had artificial intelligence listen to the sounds of typing through a phone and over Zoom, with eerie results. Link
Wind-driven wildfires have claimed a minimum of six lives on Hawaii’s Maui island, causing extensive damage to hundreds of structures. Maui Mayor Richard Bissen Jr confirmed the fatalities, while rescue efforts continue.
Lawsuit filed after baby allegedly decapitated during delivery at metro Atlanta hospital. Link
6-year-old boy who shot his Virginia teacher said “I shot that b**** dead,” unsealed records show. Link
Paper exams, chatbot bans: Colleges seek to ‘ChatGPT-proof’ assignments. Link
36 confirmed dead: federal and local officials give updates on Maui fires. Link
Red Cross ends blood-donation restrictions that singled out gay and bisexual men. Link
Harry Kane transfer news: Bayern Munich agree deal in principle with Tottenham for striker. Link
Unraveling August 2023: August 09th 2023
Latest AI News and Trends on August 09th 2023
Step by Step Software Design and Code Generation through GPT
If you have used ChatGPT, or GPT in general, for software design and code generation, you might have noticed that for larger or trickier codes, it skips a lot of the implementation or misunderstands the design. That’s where tools like GPT Engineer and Aider come to help. However those tools for the most part keep the user out of the loop during the design. To explore the design space with GPT and be involved in decision making, you can use GPT-Synthesizer. GPT-synthesizer is a free and open-source tool which you can use for personal or commercial purposes. It uses LangChain to efficiently process larger codebases: https://github.com/RoboCoachTechnologies/GPT-Synthesizer
Collaboratively implement an entire software project with the help of an AI.
GPT-Synthesizer walks you through the problem statement and explores the design space with you through a carefully moderated interview process. If you have no idea where to start and how to describe your software project, GPT Synthesizer can be your best friend.
What makes GPT Synthesizer unique?
The design philosophy of GPT Synthesizer is rooted in the core, and rather contrarian, belief that a single prompt is not enough to build a complete codebase for a complex software. This is mainly due to the fact that, even in the presence of powerful LLMs, there are still many crucial details in the design specification which cannot be effectively captured in a single prompt. Attempting to include every bit of detail in a single prompt, if not impossible, would cause losing efficiency of the LLM engine. Powered by LangChain, GPT Synthesizer captures the design specification, step by step, through an AI-directed dialogue that explores the design space with the user.
GPT Synthesizer interprets the initial prompt as a high-level description of a programming task. Then, through a process, which we name “prompt synthesis”, GPT Synthesizer compiles the initial prompt into multiple program components that the user might need for implementation. This step essentially turns ‘unknown unknowns’ into ‘known unknowns’, which can be very helpful for novice programmers who want to understand an overall flow of their desired implementation. Next, GPT Synthesizer and the user collaboratively find out the design details that will be used in the implementation of each program component.
Different users might prefer different levels of interactivity depending on their unique skill set, their level of expertise, as well as the complexity of the task at hand. GPT Synthesizer distinguishes itself from other LLM-based code generation tools by finding the right balance between user participation and AI autonomy.
GPT Sythesizer is easy to use. It provides you with an intuitive AI assistant in your command-line interface. See our demo for an example of using GPT Synthesizer.
GPT Synthesizer uses OpenAI’s gpt-3.5-turbo-16k as the default LLM.
Setup your OpenAI API key: export OPENAI_API_KEY=[your api key]
Run:
Start GPT Synthesizer by typing gpt-synthesizer in the terminal.
Briefly describe your programming task and the implementation language:
Programming task: *I want to implement an edge detection method from live camera feed.*
Programming language: *python*
GPT Synthesizer will analyze your task and suggest a set of components needed for the implementation.
You can add more components by listing them in quotation marks: Components to be added: *Add 'component 1: what component 1 does', 'component 2: what component 2 does', and 'component 3: what component 3 does' to the list of components.*
You can remove any redundant component in a similar manner: Components to be removed: *Remove 'component 1' and 'component 2' from the list of components.*
After you are done with modifying the component list, GPT Synthsizer will start asking questions in order to find all the details needed for implementing each component.
When GPT Synthesizer learns about your specific requirements for each component, it will write the code for you!
You can find the implementation in the workspace directory.
Robots, computers, and algorithms are hunting for potential new therapies in ways humans can’t—by processing huge volumes of data and building previously unimagined molecules.
AT AN OLD biscuit factory in South London, giant mixers and industrial ovens have been replaced by robotic arms, incubators, and DNA sequencing machines. James Field and his company LabGenius aren’t making sweet treats; they’re cooking up a revolutionary, AI-powered approach to engineering new medical antibodies.
In nature, antibodies are the body’s response to disease and serve as the immune system’s front-line troops. They’re strands of protein that are specially shaped to stick to foreign invaders so that they can be flushed from the system. Since the 1980s, pharmaceutical companies have been making synthetic antibodies to treat diseases like cancer, and to reduce the chance of transplanted organs being rejected.
But designing these antibodies is a slow process for humans—protein designers must wade through the millions of potential combinations of amino acids to find the ones that will fold together in exactly the right way, and then test them all experimentally, tweaking some variables to improve some characteristics of the treatment while hoping that doesn’t make it worse in other ways. “If you want to create a new therapeutic antibody, somewhere in this infinite space of potential molecules sits the molecule you want to find,” says Field, the founder and CEO of LabGenius. Read more
NVIDIA Releases Biggest AI Breakthroughs
– Nvidia announced a new chip GH200, designed to run AI models. It has the same GPU as the H100, Nvidia’s current highest-end AI chip, but pairs it with 141 gigabytes of cutting-edge memory and a 72-core ARM central processor. This processor is designed for the scale-out of the world’s data centers.
– NVIDIA has announced new frameworks, resources, and services to accelerate the adoption of Universal Scene Description (USD), known as OpenUSD. Through its Omniverse platform and a range of technologies and APIs, including ChatUSD and RunUSD, NVIDIA aims to advance the development of OpenUSD, a 3D framework that enables interoperability between software tools and data types for creating virtual worlds.
– NVIDIA has introduced AI Workbench, a developer toolkit that simplifies the creation, testing, and customization of pretrained generative AI models. The toolkit allows developers to scale these models to various platforms, including PCs, workstations, enterprise data centers, public clouds, and NVIDIA DGX Cloud. This will speed up the adoption of custom generative AI for enterprises worldwide.
– NVIDIA and Hugging Face have partnered to bring generative AI supercomputing to developers. The integration of NVIDIA DGX Cloud into the Hugging Face platform will accelerate the training and tuning of large language models (LLMs) and make it easier to customize models for various industries. This partnership aims to connect millions of developers to powerful AI tools, enabling them to build advanced AI applications more efficiently.
75% of Organizations Worldwide Set to Ban ChatGPT and Generative AI Apps on Work Devices
Although ChatGPT currently has over 100 million users in June 2023, the concerns for its security and trustworthiness grow. AI cybersecurity pioneer, BlackBerry, calls for caution with consumer-grade Generative AI tools in the workplace.
Some impressive figures
– 75% of global organizations are either implementing or contemplating bans on ChatGPT and other Generative AI applications in their workplaces.
– 61% view these measures as long-term or permanent due to concerns over data security, privacy, and corporate reputation.
– 83% believe unsecured apps present a cybersecurity threat to their corporate IT systems.
– 80% of IT decision-makers believe organizations have the right to control applications used for business.
– 74% feel that such bans indicate “excessive control” over corporate and BYO devices.
As AI tools get better and rules are set, companies might change their rules. It’s important to have tools to watch and manage how these AI tools are used at work.
Research was conducted in June/July 2023 by OnePoll on behalf of BlackBerry, into 2,000 IT Decision Makers across North America (USA and Canada), Europe (UK, France, Germany and the Netherlands), Japan and Australia.
Google launches Project IDX, an AI-enabled browser-based dev environment.
– For building web and multiplatform apps. It currently supports frameworks like Angular, Flutter, Next.js, React, Svelte, and Vue, and languages like JavaScript and Dart. The project is based on Visual Studio Code and integrates with Codey, Google’s PaLM 2-based foundation model for programming tasks. – IDX offers features such as smart code completion, a chatbot for coding assistance, and the ability to add contextual code actions. Google plans to add support for more languages like Python and Go in the future.
It allows developers to work from anywhere, import existing projects, and preview apps across platforms. It supports frameworks like Angular, Flutter, Next.js, React, Svelte, Vue and languages like JavaScript and Dart. AI capabilities like smart code completion and contextual code actions are also included. Google plans to add support for more languages like Python and Go in the future. Additionally, Project IDX integrates with Firebase hosting for easy deployment of web apps.
Why does this matter?
By incorporating models like Codey, IDX offers tools like Studio Bot and Duet, Google IDX might revolutionize coding experiences in Android Studio and Google Cloud. Smart code completion, contextual actions, and an assistive chatbot can empower developers to write code more efficiently and maintain high standards.
Stability AI has released StableCode, an LLM generative AI product for coding.
– It aims to assist programmers in their daily work and provide a learning tool for new developers. StableCode uses three different models to enhance coding efficiency. The base model was trained on various programming languages, including Python, Go, Java, and more. It was then further trained on 560B tokens of code.
Hugging face launches tools for running LLMs on Apple devices.
– Hugging face have released a guide and alpha libraries/tools to support developers in running LLM models like Llama 2 on their Macs using Core ML.
Google AI is helping Airlines to reduce mitigate the climate impact of contrails.
– Google AI, American Airlines, and Breakthrough Energy collaborated to use AI and data analysis to develop contrail forecast maps. These maps help pilots choose routes that minimize contrail formation, reducing the climate impact of flights.
D-ID and ElevenLabs have announced a partnership to bring premium voices to D-ID’s
Creative RealityTM studio. This collaboration will allow users to create videos with more natural speech. The new features simplify the process and enable subscribers to add high-quality synthetic voices to their videos with one click. They offer AI-generated customized video narrators in 119 languages, making video creation easier and more cost-effective.
Google and Universal Music Group are in talks to license artists’ melodies and vocals for an AI-generated music tool.
– The tool would allow users to create AI-generated music using an artist’s voice, lyrics, or sounds. Copyright holders would be paid for the right to create the music, and artists would have the option to opt in.
Disney has formed a task force to explore the applications of AI across its entertainment conglomerate, despite the ongoing Hollywood writers’ strike.
– Disney currently has 11 job openings that require expertise in AI or machine learning, covering various departments such as Walt Disney Studios, engineering, theme parks, television, and advertising. The advertising team, in particular, is focused on building an AI-powered ad system for the future.
AI researchers claim 93% accuracy in detecting keystrokes over Zoom audio LINK
Researchers achieved over 90% accuracy in interpreting remote keystrokes by recording them and training a deep learning model on the unique sound profiles of individual keys.
Laptops, especially in quieter public places, are vulnerable to this kind of attack due to their consistent and non-modular keyboard acoustic profiles.
Previous methods achieved 74.3% to 91.7% accuracy in VoIP calls; the current research benefits from recent advancements in neural network technology, like self-attention layers, to enhance audio side channel attacks.
Researchers at the Massachusetts Institute of Technology (MIT) and the Dana-Farber Cancer Institute have discovered that the use of artificial intelligence (AI) could make it easier to determine the sites of origin for enigmatic cancers and enable doctors to choose more targeted treatments.[1]
Meta disbands protein-folding team in shift towards commercial AI.[2]
OpenAI has introduced GPTBot, a web crawler to improve AI models. GPTBot scrupulously filters out data sources that violate privacy and other policies.[3]
Disney has created a task force to study artificial intelligence and how it can be applied across the entertainment conglomerate, even as Hollywood writers and actors battle to limit the industry’s exploitation of the technology.[4]
Trending AI Tools
ChattyDocs: Create AI experts on selected topics using your documents. AI answers your or customers’ questions.
Social Bellow: AI-powered prompts to revolutionize content. Overcome writer’s block, craft engaging narratives, and personalize chatbots.
Neighborbrite Vision: Transform your yard with our AI app. Choose Desert, Modern, or Cottage vibes and get a custom design.
HomeStyler AI: Experience the future of interior design. Upload a picture and watch your space transform into a designer masterpiece.
CREaiD AI: Commercial Real Estate’s 1st AI chatbot. Connect with Verified Lenders and Primary Contacts, and get real-time insights.
Taylor AI: Fine-tune open-source LLMs in minutes. Focus on experimentation & building better models, not digging through libraries.
StudentAI: Every tool a student needs to save time and excel academically. From presentations to exams, we’ve got it covered.
liteLLM: Simplify using LLM APIs from OpenAI, Azure, Cohere, Anthropic, Replicate, and Google. Call all LLM APIs using chatGPT format.
Latest Tech News and Trends on August 09th 2023
GM’s EVs to offer vehicle-to-home charging by 2026 LINK
GM is introducing vehicle-to-home (V2H) bidirectional charging technology to its Ultium-based electric vehicles by 2026, allowing them to be used as backup power sources for homes.
The first models to feature this technology include the 2024 Chevrolet Silverado EV RST, GMC Sierra EV Denali Edition 1, and Cadillac Lyriq, among others.
This initiative is under GM Energy, a new business unit from GM launched in 2022, which offers various energy solutions including stationary storage and solar energy partnerships.
Norway imposes $100k daily fines on Meta over data harvesting LINK
Meta faces a new penalty from Norwegian regulators, amounting to 1 MILLION crowns (around $100,000) per day starting from August 14 due to privacy breaches.
Norway had previously announced a temporary ban on behavioural ads on Facebook and Instagram, and warned Meta of potential fines if violations were not addressed.
Despite Meta’s recent pledge to obtain EU user consent for personalized ads, Datatilsynet remains unimpressed and plans to continue daily fines until at least November 3, with the possibility of making them permanent.
The famously overworked visual effects workers behind the Marvel movies just voted to join a union. Link
Banks hit with $549 million in fines for use of Signal, WhatsApp to evade regulators’ reach. Link
Author discovers AI-generated counterfeit books written in her name on Amazon. Link
Wind-whipped wildfires in Hawaii forced hundreds of evacuations, overwhelmed hospitals in Maui and even sent some residents fleeing into the ocean.
9-year-old girl fatally shot by neighbor in front of her father after buying ice cream and riding her scooter, legal document says. Link
5 white nationalists sue Seattle man for allegedly leaking their identities. Link
Tory Lanez sentenced to 10 years for shooting Megan Thee Stallion in the foot. Link
Teenage cousin of Uvalde school shooter is arrested, accused of threatening to ‘do the same thing’ to a school. Link
Emergency rooms becoming the ‘dumping ground’ for mentally ill who often wait days for help. Link
Unraveling August 2023: August 08th 2023
Latest AI News and Trends on August 08th 2023
How to Leverage No-Code + AI to start a business with $0
Start your Business with $0
Need a Desinger — Use Canva Need a Website — Use Carrd Need Sales — Use Gumroad Need a Writer — Use Claude Need a CRM — Use Notion Need Marketing — Use Buffer Need Videos — Use CapCut
Remember, you don’t need to have the setup before starting a business Many successful businesses started w/ a notebook and an Excel sheet.
Leverage ChatGPT as Your Personal Finance Advisor
Are you an online business owner juggling numbers and financial decisions? With ChatGPT, you can gain insights and advice on managing your business’s finances more effectively.
Try the prompt below:
Act as a certified financial advisor. I'm the owner of an online business that's been operational for two years. Here's a simplified snapshot of my current financial situation:
Monthly revenue: $10,000
Monthly operating expenses (inventory, software, marketing, etc.): $6,000
Monthly net income: $4,000
Business savings: $20,000
Personal savings: $10,000
I'm looking to increase my savings, reduce expenses, and grow my business. Please provide a comprehensive financial plan which should include budgeting tips, investment advice, and strategies for improving my overall financial health. Remember to make your suggestions based on the best practices in personal finance and business finance management.
This prompt can be adjusted according to your unique financial circumstances. For example, if you’re more concerned about debt management, retirement planning, or making significant business investments, modify your request accordingly.
Note: ChatGPT can provide a helpful start in managing your finances, but it can’t be completely relied upon for professional financial advices. In addition, please be aware that sharing sensitive financial information online carries its own risks, even in a simulated conversation with AI.
Boosting in machine learning is a technique that trains algorithms to work better together, improving accuracy and reducing bias. Learn how boosting works.
New deep learning tool leverages chest x-rays and electronic health record data to flag elevated diabetes risk up to three years before diagnosis.
OpenAI launches a web crawler to train ChatGPT
Called GPTBot, the crawler will comb through the internet to train and enhance AI’s capabilities. It can be identified by the following user agent and string.
Web pages crawled with the GPTBot user agent may potentially be used to improve future models and are filtered to remove sources that require paywall access, are known to gather personally identifiable information (PII), or have text that violates our policies.
Moreover, OpenAI also revealed how websites can prevent GPTBot from accessing their sites, either partially or by opting out entirely.
Why does it matter?
GPTBot can help AI models become more accurate and improve their general capabilities and safety. However, OpenAI has often landed in hot waters for how it collects data. Blocking the GPTBot may be OpenAI’s first step to allow internet users to opt out of having their data used for training its LLMs.
Speech deepfakes are artificial voices generated by AI models. While studies investigating human detection capabilities are limited, a new experiment presented genuine and deep fake audio to individuals and asked them to identify the deep fakes. Listeners could correctly spot the deep fakes only 73% of the time.
The experiment was done in English and Mandarin to understand if language affects detection performance and decision-making rationale. However, there was no difference in detectability between the two languages.
Why does this matter?
As speech synthesis AI systems improve, it will become more difficult for humans to catch speech deepfakes. The study suggests the need for automated detectors to mitigate a human listener’s weaknesses. It also emphasizes that expanding fact-checking and detecting tools is a significant way to protect against deep fake threats by AI.
Microsoft has announced new Azure AI infrastructure advancements and availability to bring its customer closer to the transformative power of generative AI.
Azure OpenAI Service goes global: OpenAI’s most advanced models, including GPT-4 and GPT-35-Turbo, will now be available in multiple new regions and locations.
General availability of ND H100 v5 VMs for unprecedented AI processing and scale: -It also announced general availability of the ND H100 v5 Virtual Machine series, featuring the latest NVIDIA H100 Tensor Core GPUs and low-latency networking, propelling businesses into a new era of AI applications.
Why does it matter?
These enhancements will allow more customers to leverage the capabilities of generative AI, driving innovation and transformation across various industries. It will also empower their businesses with greater computational power with significantly faster AI model performance.
Erroneous results from ChatGPT seem to be leading many scholars and pundits to dismiss it as useless or even dangerous. That might make sense at first glance, but only if we see it as just another type of search engine.
In this article, Mark Humphries suggests if you focus solely on its errors, you need to think about it in a different way. The article discusses in detail how chatbots are different from search engines (even though they seem similar). It also points out why tools like ChatGPT were not intended to be used as search engines and what exactly makes them revolutionary.
Why does this matter?
In an era when we are racing to adopt generative AI, understanding the usefulness of models like ChatGPT despite their tendency to hallucinate sometimes requires examining how they work during these instances and why.
Google Search launched AI-powered grammar checker LINK
Google has introduced an AI-powered grammar check feature in its search bar, which is currently available only in English.
To use the feature, users can enter a sentence or phrase into Google Search, followed by “grammar check”, “check grammar” or “grammar checker”, and Google will indicate if the phrase is grammatically correct or suggest a correction if needed.
The grammar check tool is accessible on both Google desktop and mobile platforms.
Zoom can now train its AI using customer data LINK
Zoom’s updated Terms of Service in March gave the company the right to train AI on user data, but clarified in a recent blog post that they will not use audio, video, or chat content for AI training without customer consent.
The new terms sparked concern as Zoom customers must either agree to data use or leave a meeting if a call starts with generative AI features enabled; Zoom stated that customers decide whether to enable these AI features and share data for product improvement.
Zoom’s privacy track record is questionable, with a history of issues such as providing less secure encryption than claimed and sharing user data with Google and Facebook, leading to an $85 million settlement in 2021.
USEFUL AI TOOLS
Text Blaze: Eliminate repetitive typing with this Chrome extension that allows users to create customizable templates and shortcuts.
MakeLanding: Generate landing pages with ease using this AI-powered platform, which creates pages based on user-provided project descriptions, complete with custom copy, logos, and illustrations.
YouAI: Create and share your own AI tools for various purposes, including content creation, language learning, and professional services.
Storyboard AI: Reduce the time needed to create concepts, scripts, and full storyboards for video agencies and creators using AI.
AI Tools Database: Explore a comprehensive Notion database featuring 1350 useful AI tools curated by The Intelligo.
After last fall signaling its intentions to expand into cloud gaming, Netflix today launched a new app that will soon allow subscribers to play games on their TV. The app, dubbed “Netflix Game Controller,” lets you use your phone as a controller after pairing it with your TV in order to play…
X, the company formerly known as Twitter, appears to be hinting toward plans to move deeper into the music industry or collaborate with artists and labels with its recent takeover of the @music handle on the social network. The account was originally operated by software developer Jeremy Vaught…
Media companies including the Canadian Broadcasting Corp and Radio-Canada are requesting Competition Bureau to investigate Meta Platforms’ decision to block news content on its platforms in the country, the CBC reported on Tuesday.
Millions of Android phone users are being urged to check their mobile devices, after apps were discovered to drain batteries by running constant background adverts when the screen is off…
McAfee named TV/DMB Player, Music Downloader, News, and Calendar applications as some of the popular applications compromised. The adverts in these apps don’t secretly start popping up until a few weeks after an initial installation – which makes spotting the scam far more difficult.
McAfee is in-turn urging users to take care and conduct thorough research before downloading any new apps onto their mobile devices – scouring the permissions before hitting the big green install button. It’s also a wise move to check the performance of a device after installing new software – keeping an eye out for indicators like rapidly draining battery-life or slower operating systems.
ChatGPT is the most popular AI chatbot on the planet, and it’s not hard to see why. With its powerful natural language capabilities, remarkable versatility, and knack for engaging in […]
Want to Know How many Battery Cycles are Left on Android Device. Here is the Simple Guide to Check Battery Health on Android.
Check Battery Details Using the Settings App
In some devices, from the Settings app, you can check the battery health of the Android phone.
Open the Settings app > tap on Battery.
Tap on View Detailed Usage.
The above steps may vary a little depending on the model of the phone that you are using.
Dial a USSD Code to Know the Battery Health of your Android Device
Unstructured Supplementary Service Data, often abbreviated as USSD codes, are certain configurations of numerics and special symbols that return certain helpful information about your phone when dialed using the phone app.
To know the battery health of an Android device, there is a specific USSD code.
Launch the phone app and go to the phone keypad.
Dial the code *#*#4636#*#*.
Press the call button.
NOTE: The above USSD code may not work on all Android devices. However, you can try and see if it works or not. The above code is quite safe to dial and has no effect on your device or its data.
Install a Third-Party Battery Health Checker
It is always difficult to find a trusted third-party app to check the device’s battery health. There is the app AccuBattery which I tried on my Android device. It is quite simple to use and doesn’t ask for unnecessary permissions on your device.
Launch the app, and it will start calibrating your device’s battery.
Tap on Finish.
Let the battery charge drop to 15 percent.
Charge the Battery completely.
Next, tap on the tab Health to know your device’s battery health.
Check the Battery Health of Samsung Smartphones
If you have a Samsung smartphone, you can install a specific app called Samsung Members on the device. Using this app, you can deduce the overall battery usage and health of the battery present in the Samsung smartphone.
Install the Samsung Members app from the Play Store.
Launch the app > tap on Discover> tap on Phone Diagnostics.
Tap on Battery Status to run a quick test.
The battery status will display on the screen.
How to Increase the Battery Life of a Phone?
Here are a couple of tips to increase your smartphone’s battery life.
Always use the official power adapter of the phone. If the power adapter got damaged, then get another official charger. Avoid using cheap third-party adapters.
Always charge your device up to 80 Percent. Also, set the device on charge when the battery level is around 30 -35 percent.
Never set your phone on charging and do activities like playing games or making phone calls. That will cause overheating and, in the long run, will damage the Battery’s health.
Use power saver mode whenever possible to avoid losing battery power.
Promote Efficient Battery Performance on Android Devices
Now, I hope you know the different methods to check the battery health status on your Android device. Also, follow the above tips to manage the battery health and increase it for more prolonged use on your phone.
Nuclear fusion scientists achieve net energy gain LINK
U.S. scientists at the Lawrence Livermore National Laboratory in California have successfully recreated a fusion ignition reaction, yielding an even higher energy gain than the initial experiment announced in December.
The fusion experiment required 2 megajoules of energy and produced 3 megajoules, indicating a significant milestone where fusion reactions output more energy than they consume, traditionally a major challenge in fusion research.
Despite these successes, the development of fusion power stations is still likely decades away, but these breakthroughs show potential for the development of clean, laser-induced fusion energy on Earth.
PayPal launches first major U.S. dollar-backed stablecoin LINK
PayPal has announced the rollout of its stablecoin, PayPal USD (PYUSD), issued by Paxos Trust Company and backed by U.S. dollar deposits and similar cash equivalents, marking a first for a major U.S. financial institution.
Eligible U.S. PayPal customers can transfer PYUSD between PayPal and compatible external wallets, use it for person-to-person payments and purchases, and convert other supported cryptocurrencies to and from PYUSD.
As an ERC-20 token on the Ethereum blockchain, PYUSD is available to a growing community of external developers, wallets, and web3 applications, and Paxos will publish monthly reports detailing the assets backing PYUSD.
$5 billion Google lawsuit over ‘incognito mode’ tracking moves a step closer to trial
Judge Yvonne Gonzalez-Rogers denied Google’s push for a summary judgment in a lawsuit over the way it tracked internet activity even after users switched to “Incognito mode.” Link
The iPhone X, iPhone 8, iPhone 8 Plus, and older models won’t be eligible for iOS 17 updates. But any iPhone capable of running iOS 12 will still get Apple’s security updates – for two years.
Apple lets you share your account passwords with a bunch of trustworthy people in a group using your iPhone. We’ll walk you through the process.
The ability to share passwords with a group of people you trust is one of the best hidden features of iOS 17. With shared passwords, you get the convenience of sharing your online accounts with multiple users while minimizing the risk of data theft.
How to Add People to Your Shared Password Group on an iPhone
When you create a new shared password group, you have complete control over the passwords you share with other people in the group. You can add or remove members or even delete the entire group anytime.
This feature can come in handy if you already use Family Sharing on your iPhone to share apps and subscriptions, as not all services support this feature, and you might need to share credentials with your family members.
Here’s how you can make a new shared password group and add people to it:
Launch the Settings app on your iPhone and select Passwords.
Enter your passcode or unlock it with Face ID for verification.
Tap the blue Get Started button and hit Continue
Enter the name of the group and tap Add People.
Search the name of the person you want to invite and tap Add in the top-right corner.
Tap Create and choose the passwords and passkeys you want to share.
Press the Move button.
After this, you’ll get a prompt asking if you want to notify the person. If so, press the Notify via Message and send an invitation. Else tap Not now.
Once you’ve successfully created a shared password group, you can easily add more people whenever you like. Go to your shared group, tap Manage, and repeat the steps you followed to add your contacts.
How to Add Passwords to Your Shared Group on an iPhone
If you want to add more passwords to your shared group, here’s what you need to do:
Go to Settings > Passwords and select the group.
Tap the plus (+) icon in the top-right corner and select Move Passwords to Group. You can also manually add a new password to the group by selecting New Password.
Sharing Your Wi-Fi Password With Another Apple Device
Apple is known for easy interoperability between its devices. That’s why many people say Apple is a walled garden—once you’re in the Apple ecosystem, it’s tough to get out because you’ll miss the convenience of owning Apple products.
For instance, it’s easy to share Wi-Fi passwords on your iPhone with another iPhone or even another Apple device like your Mac. As long as you have each other’s iCloud email addresses in the Contacts app, you can just bring your iPhone close to other Apple devices, and the one connected to Wi-Fi will automatically ask if you want to share the password. Here are the steps:
If the device that needs to connect is an iPhone or iPad, go to Settings > Wi-Fi. If it’s a Mac, go to System Settings > Wi-Fi. Then, tap on the desired network.
Now, bring the Wi-Fi-connected iPhone close to the device that needs to connect.
A Wi-Fi Password prompt will then appear on the Wi-Fi-connected iPhone, asking if its owner wants to share the Wi-Fi password.
Tap Share Password. Your iPhone will get the password and connect to the Wi-Fi network.
Hours after the US women’s national team crashed out of the World Cup, former US President Donald Trump seized the opportunity to blame the loss on star player Megan Rapinoe and the country’s “woke” path under President Joe Biden.
Large brawl in Alabama as people defend Black riverboat worker against white assailants. Link
Campbell will acquire Rao’s premium sauces parent company for $2.7 billion. Link
Texas hiker died at Utah national park while scattering father’s ashes. Link
Global child sexual abuse probe that was launched after two FBI agents were killed leads to almost 100 arrests. Link
NYC doctor sexually assaulted unconscious patients and filmed himself doing it, prosecutors say. Link
Appeals court upholds Josh Duggar’s conviction for downloading child sex abuse images. Link
Mother who was accused by Southwest of trafficking her biracial daughter files federal discrimination suit. Link
Latest Football Soccer Sport News on August 08th 2023:
The OncoNPC machine-learning model may help identify sites of origin for cancers whose origins are unknown, which could enable targeted tumor treatments.
AI facial recognition falsely identifies pregnant woman as felon
Detroit police wrongly arrested a pregnant woman based on incorrect facial recognition, the latest in a string of false identifications by law enforcement AI tools.
The Wrongful Arrest:
Porcha Woodruff was arrested for a robbery she didn’t commit due to AI facial recognition.
An 8-year-old photo led to her false identification by the AI system.
She’s now suing Detroit over the arrest that saw her jailed while pregnant.
A Systemic Issue:
At least 6 wrongful arrests linked to facial recognition AI have occurred.
All wrongly identified have been black people so far.
Critics argue it leads police to shoddy, biased investigations.
AI Accountability:
Powerful AI requires meticulous training and testing to avoid mistakes.
False arrests raise real concerns over reliance on imperfect technology.
Legal, ethical, and financial liabilities will pile up if issues persist.
TL;DR: Detroit police falsely arrested a pregnant woman based on incorrect facial recognition AI identification, prompting a lawsuit. Critics argue reliance on imperfect technology leads police to biased, shoddy investigations as wrongful arrests mount.
OpenAI CEO Sam Altman expressed concerns about generative AI’s potential impacts on future elections, particularly with hyper-targeted synthetic media.
AI-generated media has already been used in American campaign ads for the 2024 election and has sometimes caused misinformation to spread.
Altman acknowledges the risks of the technology he’s helping develop and emphasizes the importance of raising awareness about its implications.
Apple bets big on AI
Apple’s CEO Tim Cook has stated that AI and ML are embedded in every company product. This comes after concerns were raised about Apple’s lack of discussion on its AI plans while competitors have been actively incorporating the technology into their products. He also emphasized that AI is central to the design of Apple’s products, contradicting suggestions that the company has not yet integrated the technology.
Cook reassured that Apple has invested in AI for years, and this year’s Research & Development spending has hit $22.61 billion. They are also hiring dozens of AI jobs in the US, France, and China, looking to fill roles that could help build Gen AI tools.
Why does this matter?
This move signifies the potential for enhanced personalization and contextual relevance in user interactions, leading to a more intuitive and tailored experience within the Apple ecosystem. The seamless integration of AI may also pave the way for groundbreaking applications in health, home automation, and more. Ultimately redefining how users interact with and benefit from Apple’s ecosystem of products and services.
Jupyter AI is a tool that brings generative AI to Jupyter notebooks, allowing users to explore and work with AI models. It offers an %%ai magic command that turns the notebook into a reproducible generative AI playground, a native chat UI for working with generative AI as a conversational assistant, and support for various generative model providers.
Jupyter AI is compatible with JupyterLab, with version 1.x supporting JupyterLab 3.x, and version 2.x supporting JupyterLab 4.x. The main branch of Jupyter AI focuses on the newest supported version of JupyterLab, with features and bug fixes backported to JupyterLab 3 if deemed valuable.
(Here is an example of how to use ChatGPT to generate working code within the notebook cells.)
Why does this matter?
Integrating advanced AI chat-based assistance directly into Jupyter’s environment may improve coding, summarization, error correction, and content generation tasks. And with support for leading LLMs like OpenAI, AI21, Anthropic, Cohere, and even local models, JupyterAI empowers users with a powerful toolset to streamline coding workflows and obtain accurate answers.
ChatGPT’s emotional awareness is more than humans’. What?
A study found that ChatGPT has higher emotional awareness than humans. The machine was subjected to a standardized test measuring human emotional awareness and scored significantly higher. The test required participants to show empathy in fictional scenarios.
ChatGPT outperformed humans in all categories, achieving an overall score of 85 compared to 56 for men and 59 for women. The researchers suggest that ChatGPT could be helpful in psychotherapy, cognitive training, and diagnosing mental illness. Previous studies have shown that people perceive ChatGPT’s responses as more empathetic than medical professionals.
Why does this matter?
This upgrade underscores AI’s ability to comprehend emotions and could help with therapy, mental health diagnosis, and making healthcare interactions more empathetic. This shows how AI can learn emotions and improve how it interacts with people.
Promptpack: How to build a second-brain (featuring AI)
This Promptpack by Chantal Smith and Azeem Azhar explores how to build a second brain using AI-powered tools. It discusses the use of knowledge bases and the role of generative AI in research and knowledge processing. The author shares their experience using Notion as a smart knowledge repository and tools like Perplexity and Elicit to enhance search capabilities.
They also highlight ChatGPT as their favorite AI tool. The article emphasizes the importance of natural language processing and reasoning in the evolving data and knowledge management landscape.
Why does this matter?
This article explores how AI tools like Notion, Perplexity, and ChatGPT enhance knowledge management and research. Also highlights how these tools facilitate efficient information gathering, processing, and storage, emphasizing their relevance in leveraging natural language as a primary interface for data-driven reasoning.
Salesforce introduces Einstein Studio to train AI models using Data Cloud. – This new feature allows enterprises to connect and train their own AI models on proprietary data within Salesforce. Once trained, these models can be used to power various applications within Salesforce. The offering has already been tested by multiple enterprises and is now available for all users of Salesforce’s Data Cloud.
Rapper Lupe Fiasco collabs with Google for the launch of AI Tool TextFX! – Now AI will assist rappers in their songwriting process by generating alternate meanings and phrases for chosen words. Google’s Creative Technologist, Aaron Wade, credits Fiasco with taking their vision for TextFX to another level, as he wanted a tool to explore the possibilities that can arise from words and concepts, rather than having an A.I. write lyrics for him.
Azure ChatGPT supporting GPT-4 is launched! (Link)
Salesforce introduces Einstein Studio to train AI models using Data Cloud. (Link)
White Castle wants to roll out AI-enabled voices to over 100 drive-thrus. (Link)
Rapper Lupe Fiasco collabs with Google for the launch of AI Tool TextFX! (Link)
Zoom’s new terms of service allow AI training on user content, no opt-out. (Link)
Latest Tech News and Trends on August 07th 2023
X will pay legal bills of people punished for posting on platform LINK
Elon Musk commits that his social media platform, X (formerly Twitter), will cover legal expenses for users “unfairly treated” by employers due to their site activity.
Musk’s declaration on X ensures there will be “no limits” to the financial support for legal bills.
In addition to funding legal battles, Musk promises to make these lawsuits “extremely loud” and to target the boards of directors of offending companies.
Apple explores lip-reading capabilities for Siri LINK
Apple has filed a patent for lip-reading technology using motion sensors, aiming to improve Siri’s speech recognition and battery life.
While this technology could enhance user privacy, it raises data protection concerns due to the potential collection of personal information.
Though the patent showcases Apple’s R&D efforts, it doesn’t confirm the actual implementation of the technology, and its primary focus remains uncertain.
MIT finds potential energy storage method in cement LINK
MIT researchers have developed a supercapacitor using cement, carbon black, and water, potentially allowing energy storage in a building’s foundation.
The cement-based material, when combined with a special salt solution, can act as a powerful supercapacitor, offering rapid energy delivery.
While the technology is promising, questions about its durability and long-term viability remain.
Startup crafts a high-speed tube propelling items to orbitnLINK
Longshot Space CEO Mike Grace is developing a hypersonic launch system that aims to provide a cheaper alternative to rockets for sending payloads into space.
The “Longshot” accelerator uses compressed gas to propel objects through very long concrete tubes, with the goal of achieving speeds up to Mach 25 to 30.
Despite its simplicity and the accompanying challenges, the project has backing from significant figures like OpenAI’s Sam Altman and Draper VC.
‘LK-99’ trend sparks superconductors market frenzy LINK
A team of scientists from South Korea and Virginia claim to have created a superconductor, called LK-99, that can transmit electrical currents without resistance at room temperature, which could result in significant advances in fields like computing and energy.
The claim has led to viral interest and significant stock market activity, particularly for companies with perceived connections to superconductors, though the scientific community remains skeptical and is actively working to verify the findings.
Even if LK-99 is confirmed as a viable room-temperature superconductor, substantial work will be needed to figure out how to implement it into commercial products, underscoring that the technology remains in early stages.
Google’s narrowing legal battlefield in antitrust case LINK
Federal Judge Amit Mehta has dismissed certain claims in an antitrust lawsuit against Google, ruling that the plaintiffs, including the Department of Justice, have not proven that Google is maintaining a monopoly by favoring its own products in search results.
The judge also dismissed antitrust allegations related to Android’s compatibility, Google Assistant, and certain other aspects of Google’s operations.
However, the DOJ can proceed with other arguments in the case, such as claims that Google abuses its power through deals requiring Android manufacturers to pre-load Google apps and make Google the default search engine on their devices.
SoftBank’s $150M claim against IRL for creating fake users LINK
SoftBank is suing defunct social app IRL, which it had previously invested in, alleging fraud and seeking $150 million in damages, after an internal investigation revealed 95% of the app’s users were fake.
IRL had claimed significant user numbers, including that it was downloaded by 25% of US teens and was growing at a 400% annual rate, figures SoftBank alleges were misrepresented and inflated using bots and a secret firm to skew data.
IRL is also under investigation by the SEC to ascertain whether the app violated security laws by misleading investors, with SoftBank’s complaint implicating IRL CEO Abraham Shafi and several of his family members in the alleged fraud.
Google’s $99 on-campus hotel offer to push hybrid work LINK
Google is running a summer promotion allowing full-time staff to book stays at the Bay View campus’ hotel for $99 per night to ease the transition to a hybrid workplace, thereby eliminating commuting for those who choose to stay.
While the offer may align with some apartment rental costs, it necessitates employees to pay for their stay, potentially leading to additional costs if they maintain a separate home, and the benefit is limited to those working at the Bay View campus.
This move coincides with increasing pressure from Google on remote workers to return to the office, amidst rising tensions, including a complaint lodged by YouTube contractors alleging the misuse of return-to-office policies to suppress labor organization.
Tesla jailbreak enabled by unpatchable hardware flaw LINK
Researchers from Technische Universität Berlin have reportedly jailbroken Tesla vehicles, unlocking features usually available through in-car purchases, and are set to present their findings at the 2023 Black Hat USA conference.
The jailbreak could potentially allow hackers to access hardware-protected keys used by Tesla for vehicle authentication and decrypt a vehicle’s internal storage, gaining access to personal user data.
The vulnerability is tied to an unpatchable flaw in each Tesla’s AMD processor, and the researchers used a voltage fault injection attack to manipulate the power flow and gain root privileges, a technique they have previously used to bypass AMD’s firmware TPM in PCs.
Developing AI with emotions, desires, and the ability to learn and grow, raises many philosophical and ethical questions. Such AI may mimic human behavior to a certain extent, but the essence of being human—rooted in our unique biological and experiential nature—could remain distinct.
With AI technology advancing rapidly, the possibility of AI initiating unprompted conversations might be within reach. However, these advancements also underline the need for stringent ethical guidelines to ensure respectful and beneficial human-AI interaction.
AI has revolutionized therapy by providing round-the-clock emotional support. As AI therapists become more sophisticated, they’re enhancing mental health care accessibility, yet also raising important questions about empathy and the human touch in therapy.
Generative AI tools are providing businesses with unprecedented capabilities, from designing new products to automating content creation. However, as these tools evolve, it’s critical for businesses to understand and manage their ethical implications.
Creating an AI that can convert 2D images into 3D models presents a complex challenge, but strides are being made in this area. While no perfect solution exists yet, researchers are continually exploring alternative methods to solve this problem.
OpenAI is rolling out new updates to improve ChatGPT
OpenAI is shipping out a bunch of small updates over the next week to improve the ChatGPT experience. Here’s a tl;dr
1. Prompt examples: At the beginning of a new chat, you will now see examples to help you get started.
2. Suggested replies: ChatGPT will suggest relevant ways to continue your conversation.
3. GPT-4 by default: When starting a new chat as a Plus user, ChatGPT will remember your previously selected model – no more defaulting back to GPT-3.5.
4. Upload multiple files: Now, ChatGPT can analyze data and generate insights across multiple files.
5. Stay logged in: You’ll no longer be logged out every 2 weeks!
6. Keyboard shortcuts: Work faster with shortcuts, like ⌘ (Ctrl) + Shift + ; to copy last code block. Try ⌘ (Ctrl) + / to see the complete list.
OpenAI has recently filed a Trademark application with the US Patent and Trademark Office for GPT-5. The application was filed on 18-07-2023 and is currently awaiting examination.
The trademark is intended to cover categories of:
Downloadable computer programs and software related to language models
The AI of human speech and text, NLP, ML-based language, and speech processing
Translation of text or speech and sharing datasets for ML
Conversion of audio data into text, voice, and speech recognition
Creating and generating text and developing and implementing artificial neural networks.
The application relates to Software as a Service (SaaS) in these areas.
Google DeepMind has introduced Robotic Transformer 2 (RT-2), a first-of-its-kind vision-language-action (VLA) model that learns from both web and robotics data. It then translates this knowledge into generalized instructions for robotic control. This helps robots more easily understand and perform actions– in both familiar and new situations
The approach results in very performant robotic policies and, more importantly, leads to a significantly better generalization performance and emergent capabilities due to web-scale vision-language pretraining. Thus, internet-scale text, image, and video data can now be used to help robots develop better common sense.
ChatGPT to Bard– Researchers find a way to turn AI chatbots evil
LLMs today undergo extensive fine-tuning to ensure they do not produce harmful content in their responses. However, new research has introduced an approach that automatically produces adversarial suffixes to prompt the models, which results in affirmative responses for objectionable queries.
Unlike traditional jailbreaks, these are built in an entirely automated fashion, allowing one to create virtually unlimited number of such attacks. Although built to target open-source LLMs, the strings easily transfer to many closed-source, publicly-available chatbots too, like ChatGPT, Bard, and Claude.
Together AI has released LLaMA-2-7B-32K, a 32K context model built using Meta’s Position Interpolation and Together AI’s data recipe and system optimizations, including FlashAttention-2. You can fine-tune the model for targeted, long-context tasks– such as multi-document understanding, summarization, and QA. Here’s the model in Playground completing a book:
Upon evaluation, the model achieves comparable quality than the original LLaMA-2-7B base model.
Threads, a Twitter rival developed by Meta, had a record-breaking launch, reaching 100 million users within days, but its daily active user count has since declined by 82%.
Users are spending much less time on the app, with usage dropping from nearly 20 minutes per day at launch to barely three minutes per day now.
Despite the decline, Meta’s CEO, Mark Zuckerberg, remains optimistic about Threads and plans to focus on retention and improving the app’s features.
Apple’s third quarter shows mixed results: iPhone sales down, but subscriptions growing LINK
Apple’s third-quarter earnings for 2023 surpassed analyst expectations, but hardware revenue declined compared to the previous year.
iPhone, Mac, and iPad sales were down by 2%, 7%, and 20% respectively, while the “Other Products” category, including wearables, grew by 2%.
The highlight of the earnings report was Apple’s services division, which saw an 8% year-over-year growth, with more than 1 billion paying users in various subscription services, generating $21.21 billion in Q3 2023.
Alphabet sells 90% of its stake in struggling Robinhood LINK
Alphabet, the parent company of Google, reduced its stakes in several publicly traded firms, including Robinhood, 23andMe, and Duolingo.
The company sold nearly 90% of its stake in Robinhood and also trimmed significant positions in Duolingo and 23andMe.
Robinhood, which saw a surge of users during the pandemic, reported stronger-than-expected earnings but still faces challenges with depressed monthly active users.
FCC issues a record $300 million fine against largest robocall scam LINK
The FCC issued a record-breaking fine of $300 million to an international network of companies responsible for making over five billion illegal robocalls to more than 500 million phone numbers, including violating federal spoofing laws.
Phone companies were told to block the numbers used by the callers, resulting in a 99% decrease in calls.
The FCC described it as the largest illegal robocall operation ever investigated, and they are determined to stop the scammers behind these calls.
Bitfinex hackers who stole billions in crypto plead guilty LINK
Ilya Lichtenstein and Heather Morgan, the couple involved in the 2016 Bitfinex hack, have pleaded guilty in court.
Lichtenstein used advanced hacking tools to gain access to Bitfinex and moved 119,754 bitcoins to his own wallets, while Morgan helped him move and launder the stolen funds.
The couple set up false identities, used darknet markets and crypto exchanges, and purchased physical gold coins with the stolen money. Lichtenstein faces up to 20 years in prison, while Morgan could be sentenced to up to five years.
World’s First Tooth Regrowth Medicine Enters Clinical Trials — ‘Every Dentist’s Dream’ Could Be A Life-Changing Reality. Link
Frackers can use dangerous chemicals without disclosure due to “Halliburton loophole”. Link
The Reddit Protest Is Finally Over. Reddit Won. Link
NASA has reestablished full communications with Voyager 2. Link.
Latest News and Trends on August 05th 2023
A judge has ruled Texas’ abortion ban is too restrictive for women with pregnancy complications. Link
FDA approves first postpartum depression pill in the US. Link
Appeals court throws out conviction after judge says Black man ‘looks like a criminal to me’. Link
In a unique feat of AI, ‘Elvis’ has been brought back to life, in a manner of speaking, to perform a humorous rendition of a modern classic. The technology behind this achievement demonstrates how AI is becoming an increasingly powerful tool in music generation and other creative fields.
Meta has released AudioCraft, an open-source suite of AI audio tools, marking a significant contribution to the AI audio technology sector. These tools are expected to facilitate advancements in audio synthesis, processing, and understanding.
Researchers have discovered a method to manipulate AI into displaying prohibited content, revealing potential vulnerabilities in these systems. This research underscores the importance of ongoing studies into the reliability and integrity of AI, as well as measures to safeguard against misuse.
Meta is planning to deploy AI-powered chatbots as part of a strategy to boost user numbers on their social media platforms. This approach signifies the growing influence of AI in enhancing user interaction and engagement on digital platforms.
Despite its evident potential, some businesses and industry leaders are maintaining a cautious distance from AI, hesitating to embrace its advantages to their operations.
Summary: Google’s Search Generative Experiment (SGE) is stepping up its AI game. Not only does it offer AI-powered results, but now also related images and videos, making searches easier and engaging. (source)
Key Points:
Google’s SGE now displays images and videos related to your search query.
Publication dates are also shown, so you know how fresh the information is.
Why It Matters: This update takes Google’s AI search to a new level, providing a richer and more dynamic user experience. Getting information from searches will become easier than ever.
Tutorial: Craft Your Marketing Strategy with ChatGPT
Whether you’re a seasoned marketer or a startup founder, creating a comprehensive marketing strategy that captures the attention of your target audience can be a complex task. ChatGPT can serve as a sounding board, providing suggestions based on historical marketing knowledge and best practices.
Try the prompt below:
You are a seasoned marketing strategist. I'm launching a new eco-friendly skincare line targeted at millennials and Gen Z. I need a comprehensive marketing strategy that covers product positioning, branding, pricing, distribution channels, and promotion strategies. This should include social media campaigns, influencer partnerships, SEO, and content marketing tactics. Your recommendations should consider current marketing trends and consumer behaviors.
You can modify this prompt to suit your specific marketing needs. Whether you’re promoting a physical product, a digital service, or a personal brand, you can ask ChatGPT for tailored advice.
Karim Lakhani is a professor at Harvard Business School who specializes in workplace technology and particularly AI. He’s done pioneering work in identifying how digital transformation has remade the world of business.
Justin Stewart left for Mount Chimborazo in August 2022 to collect fungal samples from the Ecuadoran volcano at an elevation of 4,000 meters, or about 13,000 feet.
DeepSpeed-Chat: Affordable RLHF training for AI
New Microsoft research has introduced DeepSpeed-Chat, a novel system that makes complex RLHF (Reinforcement Learning with Human Feedback) training fast, affordable, and easily accessible to the AI community (open-sourced). It has three key capabilities:
Easy-to-use Training and Inference Experience for ChatGPT Like Models
A DeepSpeed-RLHF pipeline that replicates the training pipeline from InstructGPT
A robust DeepSpeed-RLHF system that combines various optimizations for training and inference in a unified way
The system delivers unparalleled efficiency and scalability, enabling training of models with hundreds of billions of parameters in record time and at a fraction of the cost. Here’s how it compares to two other frameworks (Colossal-AI and HuggingFace DDP) for accelerating RLHF training on a single NVIDIA A100-40G commodity GPU.
Why does it matter?
The current landscape lacks an accessible, efficient, and cost-effective end-to-end RLHF training pipeline for powerful models like ChatGPT, particularly when training at the scale of billions of parameters. DeepSpeed-Chat paves the way for broader access to advanced RLHF training, thereby fostering innovation and further development in the field of AI.
OpenAI is rolling out new updates to improve ChatGPT
OpenAI is shipping out a bunch of small updates over the next week to improve the ChatGPT experience. Here’s a tl;dr
1. Prompt examples: At the beginning of a new chat, you will now see examples to help you get started.
2. Suggested replies: ChatGPT will suggest relevant ways to continue your conversation.
3. GPT-4 by default: When starting a new chat as a Plus user, ChatGPT will remember your previously selected model – no more defaulting back to GPT-3.5.
4. Upload multiple files: Now, ChatGPT can analyze data and generate insights across multiple files.
5. Stay logged in: You’ll no longer be logged out every 2 weeks!
6. Keyboard shortcuts: Work faster with shortcuts, like ⌘ (Ctrl) + Shift + ; to copy last code block. Try ⌘ (Ctrl) + / to see the complete list.
Why does it matter?
These improvements make ChatGPT more user-friendly and streamline human-AI interactions, making it a more user-friendly and powerful tool overall. It will set the stage for improved and advanced AI applications as ChatGPT is today’s leading LLM.
Latest versions of Vicuna, based on the open LLaMA-2
The latest Vicuna v1.5 series based on Llama 2 features 4K and 16K context lengths (has extended context length via positional interpolation by Meta), and have improved performance on almost all benchmarks. Vicuna 1.5 tl;dr
7B & 13B parameter versions
4096 and 16384 token context window
trained on 125k ShareGPT conversations
Commercial use
Evaluated with standard benchmarks, human preference, and LLM-as-a-judge
Why does this matter?
Since its release, Vicuna has been one of the most popular chat LLMs. It has enabled pioneering research on multi-modality, AI safety, and evaluation. Since the latest versions are based on the open-source Llama-2, they can be an open LLM alternative to ChatGPT/GPT-4.
Today at the Society of NeuroInterventional Surgery’s (SNIS) 20th Annual Meeting, researchers discussed a smartphone app created that reliably recognizes patients’ physical signs of stroke with the power of machine learning.
In the study, “Smartphone-Enabled Machine Learning Algorithms for Autonomous Stroke Detection,” researchers from the UCLA David Geffen School of Medicine and multiple medical institutions in Bulgaria used data from 240 patients with stroke at four metropolitan stroke centers. Within 72 hours of the start of the patients’ symptoms, researchers used smartphones to record videos of patients and test their arm strength in order to detect patients’ facial asymmetry, arm weakness, and speech changes-;all classic stroke signs.
To evaluate facial asymmetry, the study authors used machine learning to analyze 68 facial landmark points. To test arm weakness, the team used data from a smartphone’s standard internal 3D accelerometer, gyroscope, and magnetometer. To determine speech changes, researchers used mel-frequency cepstral coefficients, a typical sound recognition method that translates sound waves into images, to compare normal and slurred speech patterns. They then tested the app using neurologists’ reports and brain scan data, finding that the app was sensitive and specific enough to diagnose stroke accurately in nearly all cases.
The world of finance is no stranger to the nefarious activities of money laundering, a global menace that has proven to be a tough nut to crack for financial institutions and regulatory bodies. However, the advent of Artificial Intelligence (AI) and Machine Learning (ML) is heralding a new frontier in global anti-money laundering efforts, offering promising solutions to this age-old problem.
Money laundering, the process of making illegally-gained proceeds appear legal, is a complex and sophisticated crime. It often involves multiple transactions, used to disguise the origin of financial assets so that they appear to have originated from legitimate sources. Traditional methods of detecting and preventing money laundering have often fallen short, due to the sheer volume of financial transactions that occur daily and the clever tactics employed by money launderers.
Enter AI and ML, two technological advancements that are revolutionizing various sectors, including finance. These technologies are now being harnessed to combat money laundering, and early indications suggest they could be game-changers.
AI, with its ability to mimic human intelligence, and ML, a subset of AI that involves the science of getting computers to learn and act like humans, are being used to analyze vast amounts of financial data. They can sift through millions of transactions in a fraction of the time it would take a human, identifying patterns and anomalies that could indicate suspicious activity.
Moreover, these technologies are not just faster; they are also more accurate. Traditional anti-money laundering systems often generate a high number of false positives, leading to wasted time and resources. AI and ML, on the other hand, can learn from past data and improve their accuracy over time, reducing the number of false positives and allowing financial institutions to focus their resources on genuine threats.
The use of AI and ML in anti-money laundering efforts is not without its challenges. For one, these technologies require vast amounts of data to function effectively. This raises privacy concerns, as financial institutions must balance the need for effective anti-money laundering measures with the need to protect their customers’ personal information. Additionally, the use of AI and ML requires significant investment in technology and skilled personnel, which may be beyond the reach of smaller financial institutions.
Meta has introduced AudioCraft, a new family of generative AI models built for generating high-quality, realistic audio & music from text. AudioCraft is a single code base that works for music, sound, compression & generation — all in the same place. It consists of three models– MusicGen, AudioGen, and EnCodec.
Meta is also open-sourcing these models, giving researchers and practitioners access so they can train their own models with their own datasets for the first time. AudioCraft is also easy to build on and reuse. Thus, people who want to build better sound generators, compression algorithms, or music generators can do it all in the same code base and build on top of what others have done.
Why does it matter?
AudioCraft is a significant step forward in generative AI research. It opens up unprecedented possibilities for creating unique audio/music– whether for video games, merchandise promos, YouTube content, educational purposes, etc. Moreover, the open-source initiative will further help advance the field of AI-generated audio and music.
AudioCraft is for musicians what ChatGPT is for content writers.
LLaMA2-Accessory is an advanced open-source toolkit for pre-training, fine-tuning, and deployment of Large Language Models (LLMs) and multimodal LLMs. Its repository is mainly inherited from LLaMA-Adapter with more advanced features.
Thus, it supports more datasets, tasks, visual encoders, and efficient optimization methods. (LLaMA-Adapter is a lightweight adaption method to efficiently fine-tune LLaMA into an instruction-following model).
Why does this matter?
It will allow to easily and quickly experiment with and build upon state-of-the-art language models, saving time and resources in the development process. Moreover, its open-source nature democratizes access to advanced AI tools, enhancing engagement and progress toward groundbreaking AI solutions across various industries and domains.
In a cutting-edge collaborative study between Google and Osaka University, scientists have revealed a novel artificial intelligence (AI) system capable of producing music reminiscent of songs that individuals were listening to while undergoing brain scans.
The research team built an AI-based pipeline, called Brain2Music, that utilized functional magnetic resonance imaging (fMRI) data to recreate music corresponding to snippets of songs that subjects listened to. The fMRI technique observes oxygen-rich blood flow in the brain to determine the most active regions.
The collected brain scans were from five participants who listened to 15-second clips of various genres, such as blues, classical, hip-hop, and pop.
While there have been previous studies on reconstructing sounds like human speech or bird songs from brain activity, attempts to recreate music from brain signals have been rare.
The process began by training an AI program to associate features of music, such as genre, rhythm, mood, and instrumentation, with participants’ brain signals. The mood of the music was labeled by researchers with descriptive terms such as happy, sad, or exciting.
The AI was then customized for each participant, establishing connections between individual brain activity patterns and diverse musical elements.
Upon training, the AI could translate unseen brain imaging data into a format representing the musical elements of the original song clips. This information was fed into another AI model developed by Google, known as MusicLM, which was initially created to generate music from text descriptions.
MusicLM utilized this information to generate musical clips that fairly and accurately resembled the original song snippets, achieving an agreement level of about 60% in terms of mood. The genre and instrumentation in both the reconstructed and original music matched considerably more often than what could be attributed to chance.
Timo Denk, a software engineer at Google in Switzerland and the study’s co-author, emphasized that the method was robust across subjects, hinting at its likely effectiveness if applied to new individuals.
The underlying goal of the research is to enhance understanding of how the brain processes music. The team observed that listening to music activated specific brain regions, such as the primary auditory cortex and the lateral prefrontal cortex. The latter seems to be crucial for interpreting the meaning of songs, but more research is needed to validate this finding.
Intriguingly, the team also hopes to explore the possibility of reconstructing music that individuals are merely imagining, rather than actually hearing.
The study, published on July 20 in the preprint database arXiv, awaits peer review. The generated musical clips can be listened to online, showcasing a remarkable advancement in AI’s capabilities to bridge the gap between human cognition and machine interpretation.
One of the most comprehensive trial of its kind has found that using AI in breast cancer screening is safe and can significantly reduce the workload of radiologists. It’s also revealed that AI-supported screening can detect cancer at a similar rate to standard double reading without increasing false positives, thereby possibly easing the pressure on medical professionals.
AI’s effectiveness in screening is found to be on parwith two radiologists working together, providing a new tool in early detection.
The technology almost halves the workload for radiologists, greatly improving the efficiency.
No increase in the false-positive rate, with 41 more cancers detected with AI support.
The study, results, and future implications
The study was a randomised controlled trial involving over 80,000 women, primarily from Sweden, comparing AI-supported screening with standard care.
Interim analysis considers AI use in mammography safe, with the potential to reduce radiologists’ workload by 44%. The lead author calls for further understanding, trials, and evaluations to assess the full potential and implications of AI…
What Else Is Happening in AI?
Instagram is working on labels for AI-generated content (Link)
Google’s generative search feature now shows related videos and images (Link)
Tinder tests AI photo selection feature to help users build profiles (Link)
Alibaba rolls out open-sourced AI model to take on Meta’s Llama 2 (Link)
IBM and NASA announced the availability of the watsonx.ai geospatial foundation model on (Link)
As generative AI enters the mainstream, the crowdfunding platform Kickstarter has struggled to formulate a policy that satisfies parties on all sides of the debate.
Germany have been knocked out of the Women’s World Cup after a 1-1 draw with South Korea in Brisbane, combined with Morocco’s win over Colombia which saw them join the South Americans in the last 16.
Morocco reach the knockout stage in their first ever Women’s World Cup
Knockout Stage Bracket for 2023 Women’s World Cup
Tom Brady invests in Birmingham City and joins the advisory Board
Golden Boot race for the Women’s World Cup after the group stage
Latest World and USA News on August 03 2023
‘Cancer-killing pill’ that appears to ‘annihilate’ solid tumours is now being tested on humans. Link
Body found in floating border barrier between Texas and Mexico. Link
DeSantis-controlled Disney World district gets rid of all diversity, equity and inclusion programs and staffers. Link
Federal court sides with Indiana trans schoolchildren on bathroom access. Link
A-listers including Oprah Winfrey, Meryl Streep, Leonardo DiCaprio donate $1 million each to SAG-AFTRA relief fund. Link
Federal jury acquits Louisiana trooper caught on camera pummeling Black motorist. Link
Atlantic orcas ‘learning from adults’ to target boats off Spain’s coast. Link
Google and DeepMind have introduced Med-PaLM M, a multimodal biomedical AI system that can interpret diverse types of medical data, including text, images, and genomics. The researchers curated a benchmark dataset called MultiMedBench, which covers 14 biomedical tasks, to train and evaluate Med-PaLM M.
The AI system achieved state-of-the-art performance across all tasks, surpassing specialized models optimized for individual tasks. Med-PaLM M represents a paradigm shift in biomedical AI, as it can incorporate multimodal patient information, improve diagnostic accuracy, and transfer knowledge across medical tasks. Preliminary evidence suggests that Med-PaLM M can generalize to novel tasks and concepts and perform zero-shot multimodal reasoning.
Why does this matter?
It brings us closer to creating advanced AI systems to understand and analyze various medical data types. Google DeepMind’s MultiMedBench and Med-PaLM M show promising performance and potential in healthcare applications. It means better healthcare tools that can handle different types of medical information, ultimately benefiting patients and healthcare providers.
Meta, the owner of Facebook, is developing chatbots with different personalities to increase engagement on its platforms. These chatbots, known as “personas,” will mimic human conversations and may include characters like Abraham Lincoln or a surfer. The chatbots are expected to launch early in September and will provide users with search functions, recommendations, and entertainment.
The move is aimed at retaining users and competing with platforms like TikTok. However, there are concerns about privacy, data collection, and the potential for manipulation.
Why does this matter?
Meta’s move to develop AI-powered chatbots with different personas comes in response to competition from rivals like TikTok and Snap. TikTok has been gaining popularity and challenging established platforms like Facebook. Meanwhile, Snap has already launched its “My AI” feature, an experimental chatbot that has engaged 150 million users. Meta is also challenging companies like OpenAI, which launched ChatGPT. By introducing these chatbots, Meta aims to attract and retain users while staying at the forefront of AI innovation in social media.
An Asian woman asked AI to improve her headshot and it turned her white
An Asian-American MIT grad used an AI image generator to make her headshot more professional but was shocked to find it altered her appearance to look white. The incident led to discussions about racial bias in AI, eliciting reactions from the CEO and highlighting concerns over the technology’s imperfections.
What happened and the reactions
Rona Wang, an Asian-American MIT grad, used Playground AI’s image editor to make her headshot look more professional, only to find that it lightened her skin and altered her race.
Wang expressed disbelief and concern over the incident, wondering if the AI assumed that she needed to be white to appear professional.
The incident quickly caught public attention, and both the CEO of Playground AI, Suhail Doshi, and media outlets reacted to it.
CEO’s response was evasive…
Suhail Doshi, the CEO of Playground AI, responded to the Boston Globe’s interview but did not directly address the concerns about racial bias.
He used a metaphor involving rolling a dice to question whether the incident was indicative of a systemic issue.
… which leads to the broader issue of racial bias in AI
Wang’s experience brought attention to the recurring problem of racial bias, a concern she had previously expressed.
Her evolving views on the AI’s bias and her struggles with AI photo generators highlight ongoing challenges in the industry.
The incident serves as a stark reminder of the imperfections in AI and raises questions about the haste to integrate such technology in various sectors.
How China Is Using AI In Schools To Improve Education & Efficiency
1. AI Headband: Headbands measure how focused students are. Teachers and parents get this information on their computers.
2. Robots: Robots in classrooms look at students’ health and how involved they are in lessons.
3. Tracking Uniforms: Students wear special uniforms with chips that show where they are.
4. Surveillance Cameras: Cameras watch how often students look at their phones or yawn in class.
These efforts are part of a big experiment to use AI to make education in China better and more efficient.
Could this be the future of education worldwide?
Top 4 AI models for stock analysis/valuation?
– Boosted.ai – AI stock screening, portfolio management, risk management – Danielfin – Rates stocks and ETFs with an easy-to-understand global AI Score – JENOVA – AI stock valuation model that uses fundamental analysis to calculate intrinsic value – Comparables.ai – AI designed to find comparables for market analysis quickly and intelligently
What Machine Learning Reveals About Forming a Healthy Habit. Link
Contrary to popular belief, behaviors don’t become habits after a “magic number” of days. Wharton’s Katy Milkman shares what machine learning is teaching scientists about habit formation.
“There’s this widely spread rumor that it takes 21 days to form a habit. You may have also heard it takes 90 days to form a habit. There are popular books that tout these numbers that don’t have a sound basis in research. What we find is there is no such magic number,” said Katy Milkman, a Wharton professor of operations, information and decisions.
What Else Is Happening in AI on August 02nd 2023
Uber is creating a ChatGPT-like AI bot, following competitors DoorDash & Instacart. (Link)
YouTube testing AI-generated video summaries. (Link)
AMD plans AI chips to compete Nvidia and calls it an opportunity to sell it in China. (Link)
Kickstarter needs AI projects to disclose model training methods. (Link)
UC hosting AI forum with experts from Microsoft, P&G, Kroger, and TQL. (Link)
AI employment opportunities are open at Coca-Cola and Amazon. (Link)
Latest Tech News on August 02nd 2023
Meta is so unwilling to pay for news under a new Canadian law that it’s starting to block it on Facebook and Instagram in that country. Meta permanently ending news availability on its platforms in Canada starting today. Link
Uber CEO balks after a reporter tells him the cost of his 2.9-mile Uber ride: ‘Oh my God. Wow.’ Link
Reddit beats film industry, won’t have to identify users who admitted torrenting. Link
Superconductor Breakthrough Replicated, Twice, in Preliminary Testing. Link
Tech experts are starting to doubt that ChatGPT and A.I. ‘hallucinations’ will ever go away: ‘This isn’t fixable’. Link’
Latest Football/Soccer/Sports News on August 02nd 2023
Women World Cup: France 6-3 Panama; Brazil 0-0 Jamaica; Argentina 0 – 2 Sweden; South Africa 3- Italy 2; Brazil and Argentina are out. Link
France 6-3 Panama; Brazil 0-0 Jamaica; France go through to the last 16 as group winners. The result confirms Brazil’s elimination. Jamaica are through in second place.
South Africa 3- Italy 2 South Africa are into the last 16 after claiming their first Women’s World Cup win with a thrilling 3-2 victory over Italy in Wellington.
Argentina 0 – 2 Sweden; Sweden beat Argentina to make it three wins from three at the Women’s World Cup, clinching top spot in Group G and a mouth-watering last-16 clash with the USA.
Italian goalkeeper Gianluigi Buffon has retired from football, the 45-year-old announced on Wednesday, after 28 years in which he reached exceptional milestones for club and country.
Forward who trained as a youngster at Arsenal with women and boys is justifying the excitement over her talent with England
“UEFA or FIFA Must Find Solutions” – Liverpool Boss Jurgen Klopp Complains About Saudi Arabia’s Transfer Deadline. Link
Arsenal agree terms with Brentford keeper Raya; Dembele to PSG is done; Chelsea sign Rennes midfielder Ugochukwu; Mane leaves Bayern to join Ronaldo at Al-Nassr; Will Haaland continue breaking records. Link
Arsenal agree terms with Brentford keeper Raya – Gossip
Dembele to leave Barcelona for PSG – Xavi
Chelsea sign Rennes midfielder Ugochukwu
Mane leaves Bayern to join Ronaldo at Al-Nassr
Will Haaland continue breaking records
Erling Haaland broke all, well most of, the Premier League goalscoring records in his first season in England – so what can he do this season?
The Norway forward scored a record 36 goals in 35 league games to win the Golden Boot – and netted 52 goals, a record for a Manchester City player, in 53 games in all competitions.
He never looked back after his opening games, where he smashed many of the records for fast goalscoring starts in the Premier League that had been set back in 1992-93 by Coventry City’s Mick Quinn.
Haaland also helped City win the Treble of Premier League, Champions League and FA Cup.
What records can be break in 2023-24?
Katie Ledecky makes swimming history with major world championship wins. Link
Jalin Hyatt reported to have broken NFL record for the fastest speed at 24 MPH. Link
Guo Jincheng obliterates 50m world record, breaking :30 in S5 free at Para Swim Worlds. Link
Google Street View car evades police at 100 mph, crashes into creek, Indiana cops say. Link
Unraveling August 2023: Latest News on August 02nd 2023
Trump charged by Justice Department for efforts to overturn his 2020 presidential election loss. Link
FBI finds 200 sex trafficking victims, 59 missing children in two-week sweep. Link
Woman accused of killing bride in DUI golf cart crash must remain in custody, S.C. judge orders. Link
U.S. ban on popular lightbulb goes into effect. Link
The Pittsburgh synagogue gunman will be sentenced to death for the nation’s worst antisemitic attack. Link
Unraveling August 2023: August 01st 2023
Latest AI News on August 01st 2023
News Corp Leverages AI to Produce 3,000 Local News Stories per Week
News Corp Australia now uses AI to generate 3,000 local news articles each week. The executive chair, Michael Miller, disclosed this at the World News Media Congress in Taipei.
A team of four, known as the Data Local unit, utilizes AI to create localized news stories that span across various topics, including weather, fuel prices, and traffic reports. Peter Judd, News Corp’s data journalism editor, leads the team (he is also the credited author for many of these AI-generated stories).
News Corp’s AI supplements the work of reporters covering stories for the company’s 75 “hyperlocal” mastheads spread across Australia, from Penrith to Cairns. AI-generated content such as “Where to find the cheapest fuel in Penrith” is supervised by journalists. However, there is currently no indication within the articles that they are AI-assisted.
These thousands of AI-generated “articles” are more service-information-oriented, according to a News Corp spokesperson. They emphasized that the automated updates on local fuel prices, court lists, traffic, weather, and other areas are all overseen by the Data Local team’s journalists.
Miller revealed that a majority of their new subscribers sign up for the local news, but stay for the national, world, and lifestyle news. He also disclosed that 55% of all subscriptions are spurred by hyperlocal mastheads. Amidst the shift to digital platforms and local digital-only titles, News Corp seems to be harnessing the power of AI to enhance its hyperlocal news offerings.
The success of News Corp’s AI usage in journalism suggests a trend that other newsrooms in Australia, like ABC and Nine Entertainment, may be considering. As media companies explore AI applications, the question becomes how to use it effectively to enhance content accessibility, personalization, and more.
Workers are spilling more secrets to AI than to their friends
A new study reveals that workers are more open to sharing company secrets with AI tools than with friends. The research also highlights both the popularity of AI tools in workplaces and the potential security risks, with an emphasis on the growing challenges related to cybersecurity.
Here’s thesource, which I summarized in a few main points:
Workers’ positive attitudes towards AI, especially in the US
A third of workers from the US and UK would continue using AI tools even if banned by their companies.
69% believe the benefits of AI tools outweigh the risks, with US workers being the most optimistic (74%).
Widespread use of AI in the workplace and lack of awareness about dangers
Half of the respondents use AI for tasks like research, copywriting, and data analysis.
CybSafe’s report emphasizes that businesses are not informing employees about risks, leading to potential threats like phishing scams.
Challenges in cybersecurity and distinguishing human from AI-generated content
64% of US workers have entered work-related information into AI tools, and 93% are potentially sharing confidential data with AI.
60% of respondents claim they can accurately distinguish human from AI content, yet the blurring line poses risks for cybercrime.
Google’s AI will auto-generate ads
Google Ads has introduced a new feature that uses AI to generate advertisements on its platform automatically. The feature utilizes Large Language Models and generative AI to create campaign workflows based on prompts from marketers.
Google Ads can analyze landing pages, successful queries, and approved headlines to generate new creatives. The company also highlighted its commitment to privacy and introduced enhanced privacy features like Privacy Sandbox.
Why does this matter?
Using LLMs and Generative AI, this AI tool for auto-generated ads will save time, ensure privacy, and empower small businesses to leverage AI. Integrating generative AI in content creation also promises exciting possibilities beyond advertising.
Meta prepares AI chatbots with personas to try to retain users
Meta is preparing to launch AI chatbots with distinct personalities, in an effort to retain users on its platforms. This move aims to capitalize on the growing enthusiasm for AI technology and present a challenge to rivals like OpenAI, Snap, and TikTok.
If you want to stay up to date on the latest in AI and tech, look here first.
Thearticle(Financial Times) is paywalled, so here’sa recap of the article’s main points:
Meta’s strategy for engaging users through chatbots
Meta is developing chatbots that exhibit distinct personalities, such as those of historical figures and characters, to create a more engaging and personalized user experience.
The company is targeting a launch as early as September, aiming to enhance user interaction with new search functions, recommendations, and entertaining experiences with these persona-driven chatbots.
Competitive landscape and user engagement
Meta’s aim is to boost engagement and keep pace with competitors like TikTok
They will introduce “personas” to provide search functions, recommendations, and entertainment
Finally, they plan to use these chatbots to collect user data for more relevant content targeting
Addressing challenges and ethical concerns
Unraveling August 2023: LLMs to think more like a human for answer quality
This research introduces “Skeleton-of-Thought” (SoT), a method to decrease the generation latency of large language models. SoT guides LLMs first to generate the skeleton of the answer and then complete the contents of each skeleton point in parallel.
This approach provides significant speed-up (up to 2.39x across 11 different LLMs) and can potentially improve answer quality regarding diversity and relevance. SoT is an initial attempt at optimizing LLMs for efficiency and encouraging them to think more like humans for better answers.
Research by: Microsoft Research And Department of Electronic Engineering, Tsinghua University.
Why does this matter?
By emulating human-like thinking processes, LLMs can deliver more natural and contextually appropriate answers, enhancing their practical applications across various domains, such as NLP, customer support, and information retrieval. This advancement brings us closer to creating AI systems that can interact with users more effectively, making them more valuable tools in our everyday lives.
Summary:UCLA researchers have discovered that GPT-3 matches or outperforms undergrad students in solving reasoning problems typically found on exams like the SAT. (source)
Key points:
In a study conducted at the University of California, Los Angeles, GPT-3 was tested against 40 undergraduates to solve reasoning problems. The AI scored an impressive 80%, leaving the humans trailing with a below 60% average.
In the SAT “analogy” questions that were unpublished online – GPT-3 outperformed the average human score yet again.
GPT-3 had a tougher time matching a piece of text with a short story that had the same message. But GPT-4, the next version of the model, did a better job.
While researchers recognize GPT-3’s advancements, they also acknowledge its limitations in social interactions, mathematical reasoning, and spatial problem-solving.
Why it Matters: Picture AI and humans, inching closer in a problem-solving marathon. This isn’t about robots stealing jobs, no. It’s about reshaping the way we learn and do business with AI.
Unraveling August 2023: ToolLLM masters 16k+ real-word APIs
ToolLLM is a framework that enhances the tool-use capabilities of open-source LLMs by training them to follow human instructions to use external tools (APIs). The framework includes a dataset called ToolBench, which contains instructions for using over 16,000 real-world APIs.
A depth-first search-based decision tree (DFSDT) is used to improve the planning and reasoning capabilities of the LLMs. An automatic evaluator called ToolEval is also developed to assess the performance of the LLMs. The results show that the trained LLM, ToolLLaMA, can execute complex instructions and generalize to unseen APIs, performing comparably to closed-source LLMs like ChatGPT.
Why does this matter?
ToolLLM, can execute complex instructions and perform comparably to closed-source models like ChatGPT. And it bridges the gap between language models and practical tool usage, making them more versatile and valuable for various applications.
GMPlus : GMPlus is a chrome extension that makes your email writing easier by providing a shortcut anytime you write an email. No need to switch to other tabs. It helps you compose high-quality emails in minutes.
NanoNets AI email autoresponder : It’s free no login AI email writer that helps you write an effective email copy in minutes. With this tool, you can automate your email responses and create compelling email copies.
Rytr : Rytr AI is an AI-powered writing tool that helps users generate high-quality content quickly and easily. It is easy to use and requires very little effort to generate email copy that converts.
Smartwriter AI : It is an AI email marketing tool that helps generate personalized emails that can get positive replies faster and cheaper. It automates email outreach, so you don’t have to research constantly.
Copy AI : It’s an easy to use copy-generating tools that can help you write copy real quick. It can generate copy for Instagram captions, nurturing email subject lines, cold outreach pitches. Thoughts ? More useful resources in this guide.
Tutorial: ChatGPT Prompt to Enhance Your Customer Service
In the evolving landscape of online businesses, excellent customer service remains pivotal. ChatGPT can play a vital role in elevating your customer service quality. In this tutorial, we will explore how you can utilize ChatGPT to ensure your customers feel valued, and their concerns are promptly addressed. Here’s a customized prompt you can try with ChatGPT to streamline your customer service approach.
Try the prompt below:
Assume the role of a customer service expert. I run an online store selling tech gadgets and I'm receiving an increasing volume of customer inquiries and complaints. I need a comprehensive plan to improve my customer service. This should include strategies for effectively managing and responding to customer inquiries, handling complaints, providing after-sales service, and turning negative experiences into positive ones. Your recommendations should be based on the latest best practices in customer service and consider the specific challenges of an online business.
This prompt can be customized according to your business’s specific needs. Whether you’re struggling with a high volume of inquiries, dealing with complex complaints, or looking to improve your overall customer satisfaction, you can seek advice from ChatGPT.
Daily AI Update News from Google DeepMind, Together AI, YouTube, Capgemini, Intel, and more
DoNotPay, an AI lawyer bot known as ChatGPT4, is transforming how users handle legal issues and save money. In under two years, this innovative robot has successfully overturned more than 160,000 parking tickets in cities like New York and London. Since its launch, it has resolved a total of 2 million related cases.
Microsoft hints Windows 11 Copilot with third-party AI plugins is almost here.
In an analyst note on Tuesday, the financial services arm of Swiss banking giant UBS raised its guidance for long-term AI end-demand forecast from 20% compound annual growth rate (CAGR) from 2020 to 2025 to 61% CAGR between 2022 to 2027.
The next generation of the successful OpenAI language model is already on the way. It has been discovered that the North American company has filed a registration application for the GPT-5 mark with the United States Patent and Trademark Office.
Dell and Nvidia join hands for Gen AI solutions – The Dell Generative AI solutions portfolio builds on the initial Project Helix announcement made in May, which involved a close collaboration with Nvidia. The portfolio includes new validated designs to help enterprises deploy AI workloads on-premises. This partnership aims to assist customers in navigating the generative AI landscape and provide them with the necessary tools to successfully implement AI solutions in their businesses.
Google will update Assistant with similar tech like ChatGPT – Google is planning to update its Assistant with features powered by generative AI, similar to ChatGPT and Bard. The company has already started exploring a “supercharged” Assistant powered by large language models. The team has begun working on this update, starting with mobile.
ChatGPT Android app is now available in all countries and regions where it is supported.
Incredible response to Meta’s Llama 2, 150K+ downloads in just a week! – In just one week, they received over 150,000 download requests, showcasing the excitement and interest from the community. They are eagerly looking forward to seeing how developers and users utilize these models in their projects and applications.
Google DeepMind introduces AI model to control robots – It has introduced Robotic Transformer 2 (RT-2), a first-of-its-kind vision-language-action (VLA) model that learns from both web and robotics data. It then translates this knowledge into generalized instructions for robotic control. This helps robots more easily understand and perform actions– in both familiar and new situations. – The approach results in very performant robotic policies and, more importantly, leads to a significantly better generalization performance and emergent capabilities due to web-scale vision-language pretraining.
ChatGPT to Bard; researchers find a way to turn AI chatbots evil – New research has introduced an approach that automatically produces adversarial suffixes to prompt language models, which results in affirmative responses for objectionable queries. – Unlike traditional jailbreaks, the approach is built in an entirely automated fashion, allowing one to create virtually unlimited number of such attacks. Although built to target open-source LLMs, the strings easily transfer to many closed-source, publicly-available chatbots too, like ChatGPT, Bard, and Claude.
Together AI extends Llama-2 to 32k context – It has released LLaMA-2-7B-32K, a 32K context model built using Meta’s Position Interpolation and Together AI’s data recipe and system optimizations, including FlashAttention-2. You can fine-tune the model for targeted, long-context tasks– such as multi-document understanding, summarization, and QA.
Forget subtitles; YouTube now dubs videos with AI-generated voices – It is using Aloud, a free tool that automatically dubs videos using synthetic voices.
Capgemini will invest 2Bn euro in AI and double AI teams – The Paris-based IT firm will invest 2 billion euro in AI and plans to double its data and AI teams in the next three years.
Intel plans to build AI into its every product – Intel CEO Pat Gelsinger was very bullish on AI during the company’s Q2 2023 earnings call, telling investors that Intel plans to “build AI into every product that we build.”
GPT-4 passes first Harvard semester in humanities and social sciences – In an experiment, a Harvard student had GPT-4 write seven essays on topics such as economic concepts, presidentialism in Latin America, and a literary analysis of a passage from Proust. GPT-4 earned a respectable 3.57 GPA.
AI Knowledge Nugget: Large Language Models and Nearest Neighbors
This thoughtful article by Sebastian Raschka, PhD explores using nearest-neighbor methods in the context of large language models. He highlights the beauty of simple techniques like nearest neighbor algorithms and discusses their potential for making significant contributions based on foundational or classic approaches. Nearest neighbor algorithms, though not as popular as before, are still widely used in practice, and the k-Nearest Neighbor algorithm is recommended as a benchmark for predictive performance in classification projects.
(A k-nearest neighbor classifier with k=5.)
The article also provides additional resources on improving computational performance for nearest-neighbor methods.
Why does this matter?
This article showcases a simple yet effective method. It demonstrates that foundational techniques can still be competitive in low-resource scenarios and highlights the potential of alternative approaches.
Unraveling August 2023: Latest Sport News on August 01st 2023
Bayern Munich are prepared to break their club-record 80m euro (£68m) fee to sign 30-year-old England striker Harry Kane;
Tuesday’s gossip: Kane, Mbappe, Johnson, Lukaku, Vlahovic, Kolo Muani, Colwill, Verratti, Osimhen, Virgil van Dijk named new Liverpool captain, Trent Alexander-Arnold vice-captain.
Chelsea are now back in talks again with Juventus. Swap deal between Romelu Lukaku & Dušan Vlahović has been discussed again.
Bayern Munich are prepared to break their club-record 80m euro (£68m) fee to sign 30-year-old England striker Harry Kane from Tottenham.(Sky Sports)
Tottenham and Bayern held talks in London on Monday and are about £25m apart in their valuation of Kane. (Athletic – subscription)
Tottenham could use the money raised by Kane’s sale to bring in Barcelona’s Ivory Coast midfielder Franck Kessie, 26, and 28-year-old France defender Clement Lenglet. (Mundo Deportivo – in Spanish)
Tottenham are eyeing Nottingham Forest’s £50m-rated Wales forward Brennan Johnson, 22, if Kane is sold. (Mail)
Chelsea co-owner Todd Boehly faces competition from Barcelona in offering a player-plus-cash deal to Paris St-Germain for 24-year-old France forward Kylian Mbappe. (Independent)
Chelsea are exploring a potential swap deal involving Belgium striker Romelu Lukaku, 30, and Juventus’ 23-year-old Serbia forward Dusan Vlahovic. (Fabrizio Romano)
PSG have rekindled their interest in Eintracht Frankfurt’s 24-year-old France forward Randal Kolo Muani. (L’Equipe – in French)
Chelsea’s 20-year-old English defender Levi Colwell has agreed to sign a new six-year contract. (Guardian)
Man United are expected to announce decision regarding Mason Greenwood’s future opening PL game of the season on August 14.
Lauren James produced a sensational individual performance as England entertained to sweep aside China and book their place in the last 16 of the Women’s World Cup as group winners. Source: BBC
27 years ago today Alan Shearer signed for Newcastle for a world record £15 million. Measuring football inflation as changes in PL total income, that figure today would be £235 million
Unraveling August 2023: Latest Tech News on August 01st 2023
Scientists Create New Material Five Times Lighter and Four Times Stronger Than Steel.
Researchers from the University of Connecticut and colleagues have created a highly durable, lightweight material by structuring DNA and then coating it in glass. The resulting product, characterized by its nanolattice structure, exhibits a unique combination of strength and low density, making it potentially useful in applications like vehicle manufacturing and body armor. (Artist’s concept.)
First U.S. nuclear reactor built from scratch in decades enters commercial operation in Georgia
Nuclear power now makes up about 25% of the generation of Georgia Power, the largest unit of Atlanta-based Southern Co.
A.I. is on a collision course with white-collar, high-paid jobs — and with unknown impact
About 1 in 5 American workers have a job with “high exposure” to artificial intelligence, according to Pew Research Center. It’s unclear if AI would enhance or displace these jobs.
Workers with the most exposure to AI like ChatGPT tend to be women, white or Asian, higher earners and have a college degree, Pew found.
Technology has led some to “lose out” in the past, largely when their job is substituted by automation, one expert said.
Amazon rolls out its virtual health clinic nationwide:
Amazon is expanding its virtual clinic service nationwide.
The company launched Amazon Clinic last November as a way for patients to connect with telemedicine providers to help receive treatment for common conditions such as acne and hair loss.
Amazon has been trying to break into the health-care industry for years with mixed success.
YouTube will pay artists and rights holders for AI-generated music used on the platform. This aims to balance creative innovation and fair compensation.
MidJourney introduced a new AI feature, ‘Vary’. (Link)
Fintech giant Paytm invests in AI to develop an Artificial General Intelligence software stack. (Link)
India using AI to bring voice-activated mobile payments, RBI’s new plan. (Link)
Developers exploring AI to create Text-to-Music Apps. (Link)
Chinese firm launches WonderJourney satellite with AI-powered ‘brain.’ (Link)
Unraveling August 2023: August 21st, 2023
Latest AI News and Trends on August 21st, 2023
OpenCopilot- AI sidekick for everyone
OpenCopilot allows you to have your own product’s AI copilot. With a few simple steps, it takes less than 5 minutes to build.
It integrates with your underlying APIs and can execute API calls whenever needed. It uses LLMs to determine if the user’s request requires calling an API endpoint. Then, it decides which endpoint to call and passes the appropriate payload based on the given API definition.
Why does this matter?
Shopify has an AI-powered sidekick, while Microsoft (Windows Copilot, Bing Copilot) and GitHub (GitHub Copilot) have copilots. The above innovation empowers every SaaS product to have its own AI copilots tailored for its unique products.
LLMs are already good at synthesizing text, but personalized text generation can unlock even more. New Google research has proposed an approach inspired by the practice of writing education for personalized text generation using LLMs. It has a multistage and multitask framework consisting of multiple stages: retrieval, ranking, summarization, synthesis, and generation.
In addition, they introduce a multitask setting that further helps the model improve its generation ability, which is inspired by the observation that a student’s reading proficiency and writing ability are often correlated. When evaluated on three public datasets, each covering a different and representative domain, the results showed significant improvements over various baselines.
Why does this matter?
Customizing style is essential for many domains like personal communication, dialogue, marketing copies, stories, etc., which is hard to do via pure prompt engineering or custom instructions. The research attempts to address this and highlights how we can take inspiration from how humans achieve tasks to apply it to LLMs.
For businesses, local LLMs offer competitive performance, cost reduction, dependability, and flexibility. This article by ScaleDown provides practical guidance on setting up and running LLMs locally using a user-friendly project.
Moreover, Llama-2 and its variants are the go-to models, and the community continually refines them. The article highlights some things to note when running Llama models locally, including memory and model loader challenges.
Why does this matter?
This helps make AI accessible to individuals and businesses while avoiding limitations and high expenses associated with commercial APIs. Locally deploying LLM also helps businesses have more over the model, customize it, integrate with existing systems, and enable full utilization of its capabilities.
AI creates lifelike 3D experiences from your phone video
Luma AI has introduced Flythroughs, an app that allows one-touch generation of photorealistic, cinematic 3D videos that look like professional drone captures. Record like you’re showing the place to a friend, and hit Generate– all on your iPhone. No need for drones, lidar, expensive real estate cameras, and a crew.
Flythroughs is built on Luma’s breakthrough NeRF and 3D generative AI and a brand new path generation model that automatically creates smooth dramatic camera moves.
Why does this matter?
This marks a significant leap in democratizing 3D content creation with AI and making it cost-efficient. It opens up new possibilities for storytelling and crafting stunning digital experiences for users across various industries.
Genetic Algorithm Optimized Neural Network Model for Malicious URL Detection
URL Genie is a web application implementing a Multilayer Perceptron Neural Network optimized using genetic algorithms. Detect whether a domain name or URL is malicious by inputting a URL.
– Boosted.ai – AI stock screening, portfolio management, risk management
– JENOVA – AI stock valuation model that uses fundamental analysis to calculate intrinsic value
– Danielfin – Rates stocks and ETFs with an easy-to-understand global AI Score
– Comparables.ai – AI designed to find comparables for market analysis quickly and intelligently
Daily AI Update News from OpenCopilot, Google, Luma AI, AI2, and more
AI Copilot for your own SaaS product – OpenCopilot allows you to have your own product’s AI copilot. It integrates with your underlying APIs and can execute API calls whenever needed. It uses LLMs to determine if the user’s request requires calling an API endpoint. Then, it decides which endpoint to call and passes the appropriate payload based on the given API definition.
Teach LLMs to Personalize – New Google research has proposed a general approach for personalized text generation using LLMs, inspired by the practice of writing education. Upon evaluation, the results showed significant improvements over a variety of baselines.
Introducing Flythroughs, an app that creates lifelike 3D experiences from your phone video – It allows one-touch generation of photorealistic, cinematic videos that look like professional drone captures. No need for drones, lidar, expensive real estate cameras, and a crew. Record like you’re showing the place to a friend, and hit Generate; all on your iPhone.
Big brands are increasingly using AI-generated ads, including Nestlé and Mondelez – More brands see generative AI as a means to make creating an ad less painful and costly. However, there are concerns over whether to let people know it’s AI-generated, whether AI ads can receive copyright protections, and security risks when using AI.
AI2 drops the biggest open dataset yet for training language models – Language models like GPT-4 and Claude are powerful and useful. Still, the data on which they are trained is a closely guarded secret. The AI2’s (Allen Institute for AI) new, huge text dataset, Dolma, is free to use and open to inspection.
Ex-Machine Zone CEO launches BeFake, an AI-based social media app – Alias Technologies has introduced BeFake, a social media app for digital self-expression. Now available on both the App Store and Google Play, it aims to offer a refreshing alternative to the conventional reality portrayed on existing social media platforms.
Some of the world’s biggest advertisers, from food giant Nestle to consumer goods multinational Unilever, are experimenting with using generative AI software like ChatGPT and DALL-E to cut costs and increase productivity.
The New York Times may sue OpenAI over its AI chatbot ChatGPT, which uses the newspaper’s stories to generate text. The paper is unhappy that OpenAI is not paying for the use of its content and is also worried that ChatGPT could reduce its online traffic by providing answers based on its reporting.
Mantella allows you to have natural conversations with NPCs in Skyrim using your voice by leveraging Whisper for speech-to-text, ChatGPT for text generation, and xVASynth for text-to-speech. NPCs also have memories of your previous conversations and have awareness of in-game events.
British Prime Minister Rishi Sunak is set to spend 100 million pounds ($130 million) to buy thousands of computer chips to power artificial intelligence amid a global shortage and race for computing power.
Imagine a world where you reside in a luxurious home, an architectural marvel adorned with every comfort and amenity that one could possibly fathom. But it doesn’t stop there; your creativity knows no bounds, and you envision entire universes with their own laws of physics, teeming with diverse civilizations.
As you journey through life, your passions take an intriguing turn, guiding you towards the realm of digital marketing.
Yet, amid this shift in interests, a captivating question continues to linger in your mind like an enigmatic riddle: “If I possessed the power to design anything in the world, what wondrous creation would spring forth from my imagination?”
As your knowledge expands and your expertise in digital marketing deepens, you become acquainted with the remarkable world of graphic design software. Herein lies the key to unlock the gateway to your wildest ideas and aspirations.
With the vast array of possibilities that graphic design software offers, you come to realize that you can bring to life virtually anything your mind can conceive – and that realization holds true for anyone daring enough to venture into this realm.
While some graphic design software tools are tailored to cater to specialized fields, such as web design software that masters the dynamic nature of webpages or CAD software that focuses on technical drawings, at its core, graphic design software is an all-encompassing and versatile tool. It empowers individuals to transform their creative visions into tangible realities.
Within the confines of this article, we shall embark on a journey exploring the finest AI design software tools currently available. These cutting-edge tools are poised to revolutionize the design process and elevate your artistic capabilities to unprecedented heights.
By leveraging the power of artificial intelligence, these tools open up new horizons, enabling you to streamline and automate your design workflow like never before.
So, fasten your seatbelts and prepare to delve into the realm of limitless creativity. In the following sections, we shall uncover the potentials of AI-driven design software and how they stand as testaments to the boundless human imagination.
It’s time to manifest your artistic dreams into reality – let the voyage commence!
When it comes to harnessing the power of AI for creating mesmerizing visual graphics, few tools can rival the prowess of Adobe Photoshop CC. Renowned across the globe, this software stands as a beacon of creativity and innovation, empowering artists, designers, and digital enthusiasts to bring their imaginations to life in the most astonishing ways.
At the heart of Adobe Photoshop CC lies an impressive array of features that cater to every aspect of design. Whether you aim to craft captivating illustrations, design stunning artworks, or manipulate photographs with unprecedented precision, this software has got you covered.
With its user-friendly interface and intuitive controls, even those new to the world of digital design can quickly find themselves delving into the realm of endless possibilities.
One of the standout strengths of Photoshop lies in its ability to produce highly realistic and detailed images. From refining minute details in portraits to creating breathtaking landscapes, the software’s tools and filters enable artists to achieve a level of precision that defies belief.
The result is a visual masterpiece that captures the essence of the creator’s vision with unparalleled fidelity.
But Photoshop is not merely limited to polishing existing images; it opens the gates to boundless creativity by allowing users to remix and combine multiple images seamlessly. Whether it’s composing fantastical scenes or crafting surreal montages, the software’s blending capabilities grant designers the freedom to construct their own visual universes.
What truly sets Adobe Photoshop CC apart from the rest is its ingenious integration of artificial intelligence. The inclusion of AI-driven features elevates the design process to a whole new dimension.
Dull and lackluster photographs transform into jaw-dropping works of art with just a few clicks, as the software’s AI algorithms intelligently enhance colors, textures, and lighting, breathing life into every pixel.
Adobe’s suite of creative tools, including the likes of Adobe Illustrator and others, work in seamless harmony with Photoshop. This synergy empowers designers to amplify their creative potential even further.
Whether you’re crafting a logo, designing a website, or creating intricate vector graphics, the integration of these tools allows you to transcend the boundaries of imagination.
Planner 5D stands as an ingenious AI-powered solution, offering you the gateway to realize your long-cherished dream of a perfect home or office space. With its cutting-edge technology, this software empowers you to dive into the realm of architectural creativity and interior design like never before.
The first remarkable feature that sets Planner 5D apart is its AI-assisted design capabilities. Imagine describing your ideal home or office, and watch as the AI effortlessly translates your vision into a stunning 3D representation. From grand entrances to cozy corners, the AI understands your preferences, ensuring that every aspect of your dream space aligns with your desires.
Gone are the days of struggling with pen and paper to create floor plans. Planner 5D streamlines the process, enabling you to effortlessly design detailed and precise floor plans for your dream space.
Whether you seek an open-concept layout or a series of interconnected rooms, this software provides the tools to bring your architectural visions to life.
But that’s not all – Planner 5D goes above and beyond to cater to every facet of interior design. With an extensive library of furniture and home décor items at your disposal, you can furnish and decorate your space with ease.
From stylish sofas and elegant dining tables to enchanting wall art and lighting fixtures, the possibilities are limitless.
The user-friendly 2D/3D design tool within Planner 5D is a testament to the software’s commitment to simplicity and innovation. Whether you’re an aspiring designer or a seasoned professional, navigating through the interface is a breeze, allowing you to create the perfect space for yourself, your family, or your business with utmost ease and precision.
For those seeking a more hands-off approach, Planner 5D also offers the option to hire a professional designer through their platform. This feature is a boon for individuals who desire a polished and expertly curated space but prefer to leave the intricate details to the experts.
By collaborating with skilled designers, you can rest assured that your dream home or office will become a reality, tailored to your unique taste and requirements.
Uizard emerges as a game-changing tool that holds the power to transform the creative process for founders and designers alike. This innovative software enables you to breathe life into your ideas by swiftly converting your initial sketches into high-fidelity wireframes and stunning UI designs.
Gone are the days of spending endless hours painstakingly crafting wireframes and prototypes manually. With Uizard, the transformation from a low-fidelity sketch to a polished, high-fidelity wireframe or UI design can occur within mere minutes.
The speed and efficiency afforded by this cutting-edge technology empower you to focus on refining your concepts and iterating through ideas at an unprecedented pace.
Whether your vision encompasses web apps, websites, mobile apps, or any digital platform, Uizard stands as a reliable companion, streamlining the design process with its versatility. You no longer need to possess extensive design expertise, as the tool is intuitively designed to cater to users of all backgrounds and skill levels.
From tech-savvy founders to aspiring entrepreneurs, Uizard ensures that the creative journey remains accessible and enjoyable for everyone.
The user-friendly interface of Uizard opens up a realm of possibilities, allowing you to bring your vision to life with ease. Its intuitive controls and extensive feature set empower you to craft pixel-perfect designs that align with your unique style and brand identity.
Whether you’re a solo founder or part of a dynamic team, Uizard fosters seamless collaboration, enabling you to share and iterate on designs effortlessly.
One of the most significant advantages of Uizard lies in its ability to gather invaluable user feedback on your designs. By sharing your wireframes and UI designs with stakeholders, clients, or potential users, you can gain insights and refine your creations based on real-world perspectives.
This not only accelerates the decision-making process but also ensures that your final product resonates with your target audience.
Enter the extraordinary realm of 3D animation with Autodesk Maya, a software that transcends conventional boundaries to grant you the power to breathe life into expansive worlds and intricate characters. Whether you’re an aspiring animator, a seasoned professional, or a visionary storyteller, Maya provides the tools to transform your creative visions into stunning reality.
Imagination knows no bounds with Maya, as its powerful toolsets empower you to embark on a journey of endless possibilities. From the grandest of cinematic tales to the most whimsical of animated adventures, this software serves as your creative canvas, waiting for your artistic touch to shape it.
Complexity is no match for Maya’s prowess, as it deftly handles characters and environments of any intricacy. Whether you seek to create lifelike characters with nuanced emotions or craft breathtaking landscapes that transcend the boundaries of reality, Maya’s capabilities rise to the occasion, ensuring that your artistic endeavors know no limits.
Designed to cater to professionals across various industries, Maya stands as the perfect companion for crafting high-quality 3D animations for movies, games, and an array of other purposes. Its versatility makes it a go-to choice for animators, game developers, architects, and designers alike, unleashing the potential to tell stories and visualize concepts with stunning visual fidelity.
The heart of Maya lies in its engaging animation toolsets, each one carefully crafted to nurture the growth of your virtual world. From fluid character movements to dynamic environmental effects, Maya opens the doors to your creative sanctuary, enabling you to weave intricate tales that captivate audiences across the globe.
But the journey doesn’t end there – with Autodesk Maya, you are the architect of your digital destiny. As you explore the depths of this software, you discover its seamless integration with other creative tools, expanding your capabilities even further.
The synergy between Maya and its counterparts unlocks new avenues for innovation, granting you the freedom to experiment, iterate, and refine your creations with ease.
Aimed at architects, engineers, and a myriad of other professionals, this cutting-edge tool serves as a gateway to bring your imaginative designs to life with astonishing realism.
Architects find solace in Autodesk 3Ds Max as it empowers them to create stunningly realistic models of their architectural wonders. Engineers, too, discover the power of this software in crafting intricate and precise 3D models of their mechanical and industrial designs.
The software becomes a haven for creative professionals seeking to visualize and communicate their concepts with exceptional clarity and visual fidelity.
Beyond the realms of architecture and engineering, Autodesk 3Ds Max knows no bounds. Its versatility allows you to explore various dimensions of creativity, from crafting intricate product prototypes to fashioning enchanting animated characters.
Whatever your design aspirations may be, this software stands as a reliable companion, ensuring that your visions manifest into awe-inspiring digital realities.
In the fast-paced world of business and design, having a tool that caters to multiple purposes becomes invaluable. Autodesk 3Ds Max stands tall as a versatile and adaptable solution, making it a coveted asset for businesses and individuals alike.
Its potential to enhance the visual storytelling capabilities of any venture unlocks a new era of creativity and communication.
One of the most cherished qualities of Autodesk 3Ds Max lies in its user-friendly interface, fostering a seamless and intuitive design process. With this tool at your disposal, iteration becomes a breeze, allowing you to refine your designs effortlessly.
Each new iteration becomes a steppingstone towards perfection, ensuring that your final creation exudes excellence.
With Foyr Neo at your disposal, you can witness the transformation of your design ideas into reality in as little as a fifth of the time it takes with other software tools.
Gone are the days of grappling with complex design interfaces and spending endless hours on a single project. Foyr Neo streamlines the journey from a floor plan to a finished render, presenting you with a user-friendly interface that simplifies every step of the design process.
With its intuitive controls and seamless functionality, the software becomes an extension of your creative vision, ensuring that your ideas manifest into remarkable designs with utmost ease.
To further elevate your experience, Foyr Neo provides a thriving community and comprehensive training resources. This collaborative ecosystem allows you to connect with fellow designers, share insights, and gain inspiration from the collective creative pool.
Additionally, the abundance of training materials and support ensures that you can unlock the full potential of the software, mastering its capabilities and expanding your design horizons.
Bid farewell to the hassle of juggling multiple tools to complete a single project – Foyr Neo serves as the all-in-one solution to cater to your design needs. By integrating various design functionalities within a single platform, the software streamlines your workflow, saving you precious time and effort.
This seamless experience fosters uninterrupted creativity, enabling you to focus on the art of design without the burden of managing disparate software tools.
With this cutting-edge software, you can witness a remarkable increase in image resolution of up to 16 times, all without sacrificing an ounce of quality.
Gone are the days of tedious manual editing, spending hours painstakingly enhancing your images pixel by pixel. Let’s Enhance simplifies the process, offering a swift and efficient solution to elevate your photos’ quality with ease.
Whether you’re a professional photographer seeking crisper images for print or a social media enthusiast aiming to enlarge your visuals, this software promises to deliver the perfect shot every time.
Let’s Enhance’s proficiency in improving image resolution, colors, and lighting automatically alleviates the burden of post-processing. By entrusting this task to the intelligent algorithms of the software, you gain more time to focus on the core aspects of your business or creative endeavors.
Embrace the art of delegation and allow Let’s Enhance to handle the technicalities while you concentrate on perfecting your artistic vision.
The applications of Let’s Enhance are vast and diverse. Photographers, designers, artists, and marketers alike can benefit from this powerful tool. Imagine effortlessly preparing your images for print, knowing they’ll boast impeccable clarity and sharpness.
Envision your social media posts grabbing attention with larger-than-life visuals, thanks to Let’s Enhance’s seamless enlargement capabilities.
But Let’s Enhance doesn’t stop at resolution enhancement. It also becomes a reliable ally in refining color palettes, breathing new life into dull or faded images, and balancing lighting for picture-perfect results.
Whether it’s subtle adjustments or dramatic transformations, the software empowers you to create visuals that captivate audiences and leave a lasting impression.
6 AI Text to Video compared (updated August 2023 ) Link
Runway Features
– Text-to-video feature – Automatic prompt suggestions – The option to upload an image for reference – Different previews to choose from before generating a video – Free plan to test the tool out
Pros
– Best of AI text-to-video research – Comprehensive set of tools for video editing – Available as both a desktop and mobile app
Cons
– Gen-2 has limitations in generating intricate details, like fingers – Gen-2 video generation is limited to 4 seconds per video – The tool does not offer text-to-speech capabilities
Synthesia AI Features
– 120+ voices and accents – 140+ diverse AI avatars – 60+ video templates designed by professional designers – The option to have a custom avatar created
In today’s world, messaging apps are becoming increasingly popular, with WhatsApp being one of the most widely used. With the help of artificial intelligence, chatbots have become an essential tool for businesses to improve their customer service experience. Chatbot integration with WhatsApp has become a necessity for businesses that want to provide a seamless and efficient customer experience. ChatGPT is one of the popular chatbots that can be integrated with WhatsApp for this purpose. In this blog post, we will discuss how to integrate ChatGPT with WhatsApp and how this chatbot integration with WhatsApp can benefit your business.
The site uses openAI to generate trivia on anything and everything you want ! You can then revisit trivia you or others have made and replay them at anytime.
Solo & real time multiplayer, daily challenge, infinite playability and is getting updates daily !
Current feature roadmap :
jeopardy mode ( multiple topics and large question count )
email / sms notifications for new daily challenges etc.
public lobbies / multiplayer against random players
IBM’s study indicates that 40% of the global workforce, or 1.4 billion people, will need to reskill in the next three years due to AI’s rise.
While AI technologies, such as generative models, might shift job responsibilities, 87% of surveyed executives believe AI will augment jobs rather than replace them.
The focus in job skills has shifted from technical STEM skills (most important in 2016) to people skills like team management and adaptability (most important in 2023).
Meta did it first… Generative AI for producers
Generative AI is revolutionizing this decade’s technology, breaking into the realm of creativity once reserved for humans. Jobs are shifting, with some roles being replaced and others benefiting from AI assistance.
Content creators, take note! Meta just revealed that platforms like Facebook and Instagram will employ AI to produce music. This means no more copyright issues or losing business. Simply choose a genre, provide a sample, and the AI crafts tailor-made music for your videos.
Facebook’s music library becomes obsolete as Meta leads the way, while YouTube and TikTok will likely follow suit. As a content creator, AI eliminates rights concerns. However, creators of original music may face challenges.
AI’s impact extends to various fields, affecting writers, musicians, artists, and photographers. While some might feel the pinch, the creative economy as a whole benefits, making custom content creation easier.
Imagine conceiving, designing, and animating with AI—a reality that even big players like Disney face. This emerging world is thrilling and transformative.
To prepare, embrace AI. Integrate it into your work wherever possible. If you want to stay ahead and not fall behind to AI, leverage its capabilities.
Trustworthy LLMs: A survey and guideline for evaluating LLMs’ alignment
Ensuring alignment, which refers to making models behave in accordance with human intentions, has become a critical task before deploying LLMs in real-world applications. This new research has proposed a more fine-grained taxonomy of LLM alignment requirements. It not only helps practitioners unpack and understand the dimensions of alignments but also provides actionable guidelines for data collection efforts to develop desirable alignment processes.
It also thoroughly surveys the categories of LLMs that are likely to be crucial to improve their trustworthiness and shows how to build evaluation datasets for alignment accordingly.
NVIDIA’s tool to curate trillion-token datasets for pretraining LLMs
Most software/tools made to create massive datasets for training LLMs are not publicly released or scalable. This requires LLM developers to build their own tools to curate large language datasets. To meet this growing need, Nvidia has developed and released the NeMo Data Curator– a scalable data-curation tool that enables you to curate trillion-token multilingual datasets for pretraining LLMs. It can scale the following tasks to thousands of compute cores.
The tool curates high-quality data that leads to improved LLM downstream performance and will significantly benefit LLM developers attempting to build pretraining datasets.
Despite AI, Bing’s market share remains static LINK
Microsoft’s Bing AI Chat has failed to significantly boost Bing’s search engine market share despite its new capabilities.
Statistics from different sources show minimal changes in Bing’s global search engine share and web traffic.
Microsoft claims success based on its internal data, reporting over 100 million daily active users and strong growth since the launch of Bing Chat.
Microsoft-DataBricks collab may hurt OpenAI
Microsoft is reportedly planning to sell a new version of Databricks software, It helps customers create AI applications for their businesses. This move could potentially harm OpenAI, as Databricks allows companies to develop AI models from scratch or repurpose open-source models instead of licensing OpenAI’s proprietary ones.
Microsoft has been aggressively investing in AI services and integrating AI functionality across its products. Neither Microsoft nor Databricks have commented on the report.
What else happened in AI this week of August 12-20?
Google appears to be readying new AI-powered tools for ChromeOS
Zoom rewrites policies to make clear user videos aren’t used to train AI
Anthropic raises $100M in funding from Korean telco giant SK Telecom
Modular, AI startup challenging Nvidia, discusses funding at $600M valuation
California turns to AI to spot wildfires, feeding on video from 1,000+ cameras
FEC to regulate AI deepfakes in political ads ahead of 2024 election
Google’s AI search offers AI-generated summaries, definitions, and coding improvements.
Google Photos introduce a new AI feature, ‘Memories view’!
Amazon using AI to enhance product reviews.
WhatsApp test beta upgrade with new feature ‘custom AI-generated stickers’.
Google is testing an AI assistant that will give you Life Advice.
Robomart adopts “store-hailing” for self-driving stores delivered to customers.
OpenAI acquires AI design studio Global Illumination to work on core products, ChatGPT
The Associated Press releases guidelines for Generative AI to its journalists
Consulting giant McKinsey unveils its own generative AI tool for employees: Lilli
Opera for iOS will now include Aria, its browser AI built in collaboration with OpenAI
UK is using AI road safety cameras to detect potential driver offenses in passing vehicles
Adobe Express with AI Firefly app, now out of beta, is available worldwide
Ex-Google Brain researchers have started an AI research company called Sakana AI in Tokyo.
Runway, a gen AI video startup, has launched a new ‘Watch’ feature.
Research shows AI bots beat CAPTCHA and humans.
ML startup Arthur launched an open-source tool to help find the best LLM.
Buildbox has launched a new tool called StoryGames.AI!
Latest Tech News and Trends on August 20th, 2023
Major concerns after Cruise robotaxi incidents LINK
Following a recent collision between a Cruise robotaxi and a fire truck in San Francisco, the California DMV requested Cruise to halve its robotaxi fleet in the city.
The state agency is investigating “recent concerning incidents” with Cruise vehicles, emphasizing the need to ensure the safety of the public sharing the road with these autonomous vehicles.
This specific accident saw a Cruise Chevy Bolt EV hit by an emergency vehicle at an intersection, resulting in passenger injuries; it adds to a series of issues potentially affecting Cruise’s future operations.
As wildfires spread, Canadian leaders ask Meta to reverse its news ban LINK
The Canadian government demands that Meta lift its ban on domestic news sharing, citing its impact on sharing information about wildfires.
Meta blocked news on Facebook and Instagram due to a new law requiring payment for news articles, but this move hampers access to crucial information.
Officials and citizens express concerns, urging Meta to reinstate news sharing for safety and emergency information during the wildfire crisis.
Elon Musk suggests that Twitter’s block feature, except for direct messages, may be removed, causing concern among users.
Blocking is currently used to restrict interactions and visibility of accounts, while mute only hides posts; users value blocking for spam control and harassment prevention.
Musk’s statement prompts backlash and uncertainty about whether the feature will actually be removed.
Discover the OpenAI code interpreter, an AI tool that translates human language into code. Learn about its functions, benefits and drawbacks in this guide.
The basics of OpenAI code interpreter
OpenAI, a leading entity in the field of artificial intelligence, has developed OpenAI code interpreter, a specialized model trained on extensive data sets to process and generate programming code.
OpenAi code interpreter is a tool that attempts to bridge the gap between human language and computer code, offering myriad applications and benefits. It represents a significant step forward in AI capabilities. It is grounded in advanced machine learning techniques, combining the strengths of both unsupervised and supervised learning. The result is a model that can understand complex programming concepts, interpret various coding languages, and generate human-like responses that align with coding practices.
New Generations of People Are Becoming More and More Indistinguishable from AI
One of the most concerning aspects of this trend is the way that new generations are rewriting previous information. In the past, people would typically come up with their ideas and opinions. However, today, it is much more common for people to simply rewrite information that they have found online. This is a trend that is being exacerbated by the rise of large language models (LLMs), which can generate text that is nearly indistinguishable from human-written text. Article: new-generations-of-people-are-becoming-more-and-more-indistinguishable-from-ai/
Neolithics, an agritech company based in Israel, is using artificial intelligence and machine learning to reduce food waste and ensure food safety and quality through its optical sensing AI-powered solution known as Crystal.eye™. This technology, which can be mounted and configured in various ways, automates and upgrades quality control for fresh produce, in order to maximize utilization and reduce waste.
While the normal spectrum of visible light has 3 colors – red, green, and blue, Crystal.eye™ uses hyperspectral imaging, with over 400 spectra of light. This light can penetrate deep into a fruit or vegetable and allows the device to scan even inside the sample, eliminating the need to cut it open or grind it.
The images produce a unique fingerprint, which is then analyzed by Neolithics’ food scientists to identify various characteristics, such as firmness, moisture content, sugar content, acidity, and many more. The data is then fed to an AI machine learning engine, allowing the system to scan and analyze a large batch of samples in a matter of seconds.
The outcomes of the inspections are then instantly displayed on a digital dashboard and can be delivered as reports, tailored to each customer’s unique requirements. For example, french fry makers need to know how much dry matter is contained in the potatoes they process, while winemakers take into account the grapes’ acidity and sweetness to obtain the flavor profile they desire.
Using Crystal.eye™ allows growers and distributors to greatly expand their sampling, from the usual 1% to around 30% to 40%. This ensures greater accuracy and significantly reduces the chance of produce being discarded due to not meeting the customers’ requirements.
According to Wayne Nathanson, the company’s VP for Global Development, knowledge in food science is Neolithics’ main differentiator. While there are other companies that make the hardware to move around and sort fruits and vegetables, he says that usually these technologies work on exterior qualities, and aren’t able to analyze the produce’s interior. Most companies do not have a team of expert food scientists to fully harness the information gathered from the produce like Neolithics, he adds.
Currently, Crystal.eye™ can check the content or defects of produce, providing customers with various external or internal attributes. This solution has been launched and is being used by an increasing number of growers, distributors, and food processing companies. At the end of this year, Neolithics expects to update the technology with the capability to assess the produce’s maturity cycle, allowing customers to identify how long it will take before it spoils. The company is also working on being able to identify traces of pesticides and other banned chemicals on the produce, with release estimated for next year.
“Sustainability is very important to Neolithics, and our mission is to reduce food waste and improve food safety. Knowing how much food is wasted daily is a major motivator for making a difference. We want to eliminate food wastage across the supply chain, including removing the need to destroy the produce when it’s being inspected. We also want to get more edible quality produce to the consumer, by helping the various links of the supply chain distribute it better. There are 1.3 billion tons of wasted food annually, and there are roughly a billion people in the world experiencing hunger. We believe there’s an opportunity to feed more people with the food that is thrown out. This becomes more and more critical, the closer the world population gets to the 10 billion mark,” Nathanson says.
The new AI programming jobs that require only very basic programming skills
There has never been a more exciting and promising time to get into AI development. Forbes reports that job listings for ChatGPT-related positions increased 21 times since last November:
They need both prompt engineers and programmers. But because of Copilot and other advances in AI programming they are looking for people with some basic programming skills but who mainly excel in advanced critical analysis and reasoning skills.
They basically need people who know how to think so for people with IQs above 130, (in the genius range) this could be a dream career. But really it’s not so much about IQ as it is about the ability to think rather than just mostly learn and remember. In fact programming courses must already be teaching this brand new kind of prompt engineering and programming.
I imagine that computer programming instruction is going through very rapid evolution right now as teaching fundamental programming skills more and more gives way to teaching how to most quickly and intelligently prompt AIs to do whatever programming is needed.
If incumbent programming schools are not changing fast enough they risk losing a substantial market share to startups that begin teaching much more marketable skills.
Many businesses today want to start using AIs but they don’t know how to go about it. Computer programmers and prompt engineers who can explain all of this to them have a ready and rapidly growing job market.
Yeah there could never be a better time to get into computer programming!
The importance of making superintelligent small LLMs
Google’s Gemini will set a new standard in AI largely because of the massive data set that it is trained on.
If you’re not familiar with Gemini yet, watch this amazingly intelligent 8-minute YouTube video:
The next step would be for Google to train that stronger intelligence to shift from relying on data to relying on principles for its logic and reasoning.
Once AI’s intelligence is based on principles, subsequent iterations will no longer require massive data for their training.
That achievement will level the playing field so that Gemini is much sooner joined by competitive or stronger models.
Once that happens, everything will get very intelligent.
As Hollywood strikes, 96% of entertainment companies are boosting generative AI spend
As the Hollywood strike continues, 96% of entertainment companies are ramping up their investments in generative AI, revealing a shift in the industry’s approach to content creation and potential concerns for its workforce.
If you want to stay ahead of the curve in AI and tech, look here first.
The rise in AI spending amidst the Hollywood strike
The Hollywood writer’s strike underscores a shift in the entertainment industry’s investment strategy.
Lucidworks’ research, one of the largest of its kind, shows 96% of executives prioritize generative AI investments.
Countries like China, the UK, France, India, and the U.S. have companies heavily investing in this technology.
AI’s potential impact on Hollywood content creation
Generative AI can produce content, virtual environments, and images, posing a potential disruption to traditional methods.
Predictions suggest that by 2025, up to 90% of Hollywood content could be influenced by AI.
There’s a growing concern among Hollywood writers about the rapid integration of AI and its effect on their careers.
The future of the entertainment industry with generative AI
The emergence of synthetic actors could revolutionize the way movies and shows are produced.
AI-driven actors don’t strike, age, or demand pay raises, presenting potential benefits for studios but challenges for human actors.
Microsoft is reportedly planning to sell a new version of Databricks software, It helps customers create AI applications for their businesses. This move could potentially harm OpenAI, as Databricks allows companies to develop AI models from scratch or repurpose open-source models instead of licensing OpenAI’s proprietary ones.
Microsoft has been aggressively investing in AI services and integrating AI functionality across its products. Neither Microsoft nor Databricks have commented on the report.
Why does this matter?
Microsoft’s reported intention to introduce an AI-focused Databricks software version carries implications for OpenAI. This software empowers businesses to craft AI solutions without relying on OpenAI’s proprietary models, potentially impacting OpenAI’s market.
Meta and CMU Robotics Institute’s New Robotics research: RoboAgent. It is a universal robotic agent that can efficiently learn and generalize a wide range of non-trivial manipulation skills. It can perform 12 skills across 38 tasks, including object manipulation and re-orientation, and adapt to unseen scenarios involving different objects and environments.
The development of the RoboAgent was made possible through a distributed robotics infrastructure, a unified framework for robot learning, and a high-quality dataset. The agent also utilizes a language-conditioned multi-task imitation learning framework to enhance its capabilities. Meta is open-sourcing RoboSet, a large, high-quality robotics dataset collected with commodity hardware, to support and accelerate open-source research in robot learning.
Why does this matter?
RoboAgent has the potential to accelerate automation, manufacturing, and daily tasks as the end users can enjoy more capable and helpful robots at home. Industries can streamline operations with efficient automation, technology could push AI and robotics boundaries, and innovation might surge across sectors.
Meta challenges OpenAI with code-gen free software
Meta is set to release Code Llama, an open-source code-generating AI model that competes with OpenAI’s Codex. The software builds on Meta’s Llama 2 model and allows developers to automatically generate programming code and develop AI assistants that suggest code.
Llama 2 disrupted the AI industry by enabling companies to create AI apps without relying on proprietary software from major players like OpenAI, Google, or Microsoft. Code Llama is expected to launch next week, further challenging the dominance of existing code-generating AI models in the market.
Why does this matter?
Meta’s Code Llama is set to rival OpenAI’s Codex; this open-source AI model is an update of Meta’s Llama 2. This tool challenges giants like OpenAI, Google, and Microsoft, giving developers more control and reducing dependence on their proprietary tools.
The Associated Press has established standards for the use of generative AI in its newsroom, emphasizing that AI is not a replacement for human journalists and cautioning against creating publishable content with AI-generated text or images.
AP journalists are directed to treat AI-generated content as “unvetted source material” and apply editorial judgment and sourcing standards before considering it for publication.
The organization warns about the potential for AI to spread misinformation and advises its journalists to exercise caution, skepticism, and verify sources when dealing with AI-generated content.
A significant portion of scientific researchers using X have reduced their usage or left the platform altogether, with over 47% decreasing usage and nearly 7% quitting, according to a survey by Nature.
About 47% of polled researchers have turned to alternative platforms, with Mastodon being the most popular, followed by LinkedIn and Instagram.
The change in researcher behavior on X is attributed to the platform’s evolving dynamics, increased content prioritization, and limited accessibility of its API for researchers.
Starting from October 1st, third-party merchants on Amazon who ship their own packages will be required to pay a 2% fee per product sold.
This new fee is in addition to other charges Amazon already receives from merchants, including selling plan costs and referral fees based on product categories.
The fee comes as Amazon’s marketplace is under scrutiny, with the FTC planning to file an antitrust lawsuit over allegations that Amazon rewards third-party merchants using its logistics services while penalizing those fulfilling their own orders.
New York City is banning TikTok from government devices within 30 days, with immediate prohibition on downloading and usage by employees.
The NYC Cyber Command cited TikTok as a security threat to the city’s technical networks, prompting the decision.
While some states have broadly banned TikTok, most have restricted its use on government-owned tech, amid ongoing debates about the app’s security risks.
Unraveling August 2023: August 17th, 2023
Latest AI News and Trends on August 17th, 2023
You can now write one sentence to train an entire ML model.
How does it work?
You just describe the ML model you want…a chain of AI systems will take that sentence…it generates a dataset based on that sentence…and it trains a model for you…in ten minutes 😳
What does that mean?
Custom models in AI just got a whole lot easier. You can go from an idea (“a model that writes Python functions”) to a fully trained custom Llama-2 model in minutes 😮
Why should I care?
If you aren’t thinking about the impact of change in your industry, start now. It’s not linear and continuous, it’s exponential with step functions. 3 out of 4 C-suite executives believe that if they don’t scale artificial intelligence in the next five years, they risk going out of business entirely.
What should I do about it?
Further proof that AI is changing our work processes rapidly. You need to build a team and org that’s first and foremost, ready for change. And if you haven’t started pulling together an AI working group to get cracking on your AI usage principles and first AI use case, do it.
GPT-4 Code Interpreter masters math with self-verification
OpenAI’s GPT-4 Code Interpreter has shown remarkable performance on challenging math datasets. This is largely attributed to its step-by-step code generation and dynamic solution refinement based on code execution outcomes.
Expanding on this understanding, new research has introduced the innovative explicit code-based self-verification (CSV) prompt, which leverages GPT4-Code’s advanced code generation mechanism. This prompt guides the model to verify the answer and then reevaluate its solution with code.
The approach achieves an impressive accuracy of 84.32% on the MATH dataset, significantly outperforming the base GPT4-Code and previous state-of-the-art methods.
Why does this matter?
The study provides the first systematic analysis of the role of code generation, execution, and self-debugging in mathematical problem-solving. This highlights the importance of code understanding and generation capabilities in LLMs. Plus, the ideas presented can help build high-quality datasets that could potentially help improve the mathematical capabilities in open-source LLMs like Llama-2.
Multi-level machine learning models for estimating the risk of delay between cancer diagnosis and treatment initiation in a large group of cancer patients.
Study significance
Machine learning models that incorporate multi-level data sources can effectively identify cancer patients who are at a greater risk of experiencing treatment delays of more than 60 days after their initial cancer diagnosis.
Although neighborhood-level social determinants of health are incorporated in the study model as contributing variables, no significant impact of these factors was observed on the model performance. Furthermore, the model exhibits lower predictive effectiveness in vulnerable populations.
Future studies should include a higher proportion of vulnerable populations and more relevant social variables to improve the model performance.
Journal reference:
Frosch Z. A. K., Hasler, J., Handorf, E., et al. (2023). Development of a Multilevel Model to Identify Patients at Risk for Delay in Starting Cancer Treatment. JAMA Network Open. doi:10.1001/jamanetworkopen.2023.28712, https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2808249
Discover the top AI jobs shaping the future. Navigate the booming AI industry with insights on rewarding careers.
1. AI product manager
An AI product manager is similar to other program managers. Both jobs require a team leader to develop and launch a product. In this case, it is an AI product, but it’s not much different from any other product in terms of leading teams, scheduling and meeting milestones.
AI product managers need to know what goes into making an AI application, including the hardware, programming languages, data sets and algorithms, so that they can make it available to their team. Creating an AI app is not the same as creating a web app. There are differences in the structure of the app and the development process.
2. AI research scientist
An AI research scientist is a computer scientist who studies and develops new AI algorithms and techniques. They develop and test new AI models, collaborate with other researchers, publish research papers and speak at conferences. So, programming is only a small portion of what a research scientist does.
The tech industry is extremely open to self-taught and non-formally trained programmers, but it makes an exception for AI research scientists. They need to have a strong understanding of computer science, mathematics and statistics. Typically, they need graduate degrees.
3. Big data engineer
AI works with large data sets and so does its precursor, big data. A big data engineer is similar to an AI engineer because they are responsible for designing, building, testing and maintaining complex data processing systems that work with large data sets. But, instead of working with GPT or LaMDA, they work with big data tools, like Hadoop, Hive, Spark and Kafka.
Like AI researchers, big data engineers often have advanced degrees in mathematics and statistics. These degrees are necessary for designing, maintaining and building data pipelines based on massive data sets.
Business intelligence (BI) is also a data-driven discipline that predates the modern AI rush. Like big data and AI, BI also relies on large data sets. BI developers use data analytics platforms, reporting tools and visualization techniques to turn raw data into meaningful insights to help organizations make informed decisions.
BI developers work with a variety of coding languages and tools from major vendors, including SQL, Python, Tableau from Salesforce and Power BI from Microsoft. They also need to have a strong understanding of business processes to help improve them through data insight.
5. Computer vision engineer
A computer vision engineer is a developer who specializes in writing programs that utilize visual input sensors, algorithms and systems. These systems see the world around them and act accordingly, such as self-driving and self-parking cars and facial recognition.
They use languages like C++ and Python, along with visual sensors, such as Mobileye from Intel. Examples of use cases include object detection, image segmentation, facial recognition, gesture recognition and scenery understanding.
6. Data scientist
A data scientist is a technology professional who collects, analyzes and interprets data to solve problems and drive decision-making within the organization. They are not necessarily programmers, although many do write their own applications. Mostly, they use data mining, big data and analytical tools.
Their use of business insights derived from data enables businesses to improve sales and operations; make better decisions; and develop new products, services and policies. They use predictive modeling to forecast future events, such as customer churn, and data visualization to display research results visually. Some also use machine learning to build models to automate these tasks.
7. Machine learning engineer
A machine learning engineer is responsible for developing and implementing machine learning training algorithms and models. Training is the demanding side of machine learning and is the most processor- and computation-intensive aspect of machine learning. Therefore, it requires the highest level of skill and training.
Because of the need for advanced math and statistics skills, most machine learning engineers have advanced degrees in computer science, math or statistics. They often continue training through certification programs or a master’s degree in machine learning, deep learning or neural networks.
8. Natural language processing engineer
A natural language processing (NLP) engineer is a computer scientist who specializes in the development of algorithms and systems that understand and process natural human language input.
One of the big differentiators between traditional search engines and generative AI interfaces, such as ChatGPT, is that search engines use keywords and gather information from large amounts of existing online data. Generative AI creates new content based on other examples and patterns, and it answers queries in a chat-type format.
Like machine learning engineers, NLP engineers are not necessarily programmers first. They need to understand linguistics as much as they need to understand programming. NLP projects require machine translation, text summarization, answering questions and understanding context.
9. Robotics engineer
A robotics engineer is a developer who designs, develops and tests software for running and operating robots. Robotics has advanced significantly in recent years, such as automated home cleaners and precision cancer surgery equipment. Robotics engineers may also use AI and machine learning to boost a robotic system’s performance.
As a result, robotics engineers are typically designing software that receives little to no human input but instead relies on sensory input. Therefore, a robotics engineer needs to debug the software and the hardware to make sure everything is functioning as it should.
Robotics engineers typically have degrees in engineering, such as electrical, electronic or mechanical engineering.
10. Software engineer
A software engineer can cover various activities in the software development chain, including design, development, testing and deployment. Engineering professionals are needed at all points of software development. The demands are so high that it’s rare to find someone well versed in all of them. Most engineers tend to specialize in one discipline.
We spoke with MIT CSAIL head Daniela Rus about the emerging technology of liquid networks and implications for robotics.
The initial research papers date back to 2018, but for most, the notion of liquid networks (or liquid neural networks) is a new one. It was “Liquid Time-constant Networks,” published at the tail end of 2020, that put the work on other researchers’ radar. In the intervening time, the paper’s authors have presented the work to a wider audience through a series of lectures.
Ramin Hasani’s TEDx talk at MIT is one of the best examples. Hasani is the Principal AI and Machine Learning Scientist at the Vanguard Group and a Research Affiliate at CSAIL MIT, and served as the paper’s lead author.
“These are neural networks that can stay adaptable, even after training,” Hasani says in the video, which appeared online in January. When you train these neural networks, they can still adapt themselves based on the incoming inputs that they receive.”
The “liquid” bit is a reference to the flexibility/adaptability. That’s a big piece of this. Another big difference is size. “Everyone talks about scaling up their network,” Hasani notes. “We want to scale down, to have fewer but richer nodes.” MIT says, for example, that a team was able to drive a car through a combination of a perception module and liquid neural networks comprised of a mere 19 nodes, down from “noisier” networks that can, say, have 100,000.
“A differential equation describes each node of that system,” the school explained last year. “With the closed-form solution, if you replace it inside this network, it would give you the exact behavior, as it’s a good approximation of the actual dynamics of the system. They can thus solve the problem with an even lower number of neurons, which means it would be faster and less computationally expensive.”
The concept first crossed my radar by way of its potential applications in the robotics world. In fact, robotics make a small cameo in that paper when discussing potential real-world use. “Accordingly,” it notes, “a natural application domain would be the control of robots in continuous-time observation and action spaces where causal structures such as LTCs [Liquid Time-Constant Networks] can help improve reasoning.”
AI reconstructs song from brain activity
Neuroscientists recorded electrical activity from areas of the brain (yellow and red dots) as patients listened to the Pink Floyd song “Another Brick in the Wall, Part 1.” Using AI software, they were able to reconstruct the song from the brain recordings. This is the first time a song has been reconstructed from intracranial electroencephalography recordings.
Why does this matter?
By capturing the musicality of speech through neural signals, this research presents an innovative application of AI that could redefine how we interact and communicate, particularly for those who struggle with traditional modes of communication.
Saudi Arabia and UAE join the race for scarce Nvidia chips
Saudi Arabia has purchased at least 3,000 of Nvidia’s H100 chips at $40,000 apiece, while UAE has ordered a fresh batch of semiconductors to power its LLM. This signals the Gulf states’ intention to become major players in AI by buying up thousands of Nvidia’s GPUs which are vital in powering the boom in generative AI that has swept markets this year.
Why does this matter?
This makes them the latest to join the ever-growing queue of buyers for Nvidia chips to power AI ambitions. But will Nvidia be able to produce enough GPUs to meet the massive demand? It was reported in June that Nvidia GPUs are already in short supply (and very expensive).
Snapchat’s AI chatbot creates unexpected chaos LINK
Snapchat users reported an unexpected video posted on the My AI chatbot’s Story, which some interpreted as showing a corner between a ceiling and a wall.
The unexpected post led to concerns and fears among users, with some believing the AI feature had become sentient or evolved, prompting some to delete the app.
Snapchat described the event as a “temporary outage”, which has since been resolved, and the AI chat feature temporarily stopped responding during this period.
Exploring the Power of Mojo Programming Language
Mojo is a new programming language that combines the usability of Python with the performance of C. It is designed to be the perfect language for developing AI models and applications. Mojo is fast, efficient, easy to use, and open source. Mojo is based on the LLVM (Low Level Virtual Machine) compiler infrastructure, which is one of the most advanced compiler frameworks in the world right now. Mojo uses a new type of system that allows for better performance and error checking. Mojo has a built-in autotuning system that can automatically optimize your code for the specific hardware that you are using.
Genmo is an artificial intelligence-driven video generator that takes text beyond the two dimensions of a page. Algorithms from natural language processing, picture recognition, and machine learning are used to adapt written information into visual form. It can turn text, pictures, symbols, and emoji into moving images. Background colors, characters, music, and other elements are just some of how the videos can be personalized. The movie will include the text and any accompanying images that you provide. The videos can be shared on many online channels like YouTube, Facebook, and Twitter. Videos made by Genmo’s AI can be used for advertising, instruction, explanation, and more. It’s a fantastic resource for companies, groups, and people who must rapidly and cheaply make interesting movies.
D-ID is a video-making platform powered by artificial intelligence that makes producing professional-quality videos from text simple and quick. Using Stable Diffusion and GPT-3, the company’s Creative RealityTM Studio can effortlessly create videos in over a hundred languages. D-ID’s Live Portrait function makes short films out of still images, and the Speaking Portrait function gives a speech to written or spoken text. Its API has been refined with the help of tens of thousands of videos, allowing it to generate high-quality visuals. Digiday, SXSW, and TechCrunch have all recognized D-ID for their ability to help users create high-quality videos at a fraction of the expense of traditional approaches.
The LeiaPix Converter is a web-based, no-cost service that changes regular photographs into 3D Lightfield photographs. It employs AI to turn your images into lifelike, immersive 3D environments. Select the desired output format and upload your picture to LeiaPix Converter. The converted file can be exported in several forms, including the Leia Image Format, Side-by-Side 3D, Depth Map, and Lightfield Animation. The LeiaPix Converter’s output is great quality and straightforward to use. It’s a fantastic way to give your pictures a new feel and make unique visual compositions. It does a 3D Lightfield conversion from a 2D image. Leia Image Format, Side-by-Side 3D, Depth Map, and Lightfield Animation are only a few of the supported export formats that bring about excellent outcomes. Depending on the size of the image, the conversion procedure could take a while. The quality of your original photograph will affect the final conversion outcomes. Because the LeiaPix Converter is currently in beta, it may include problems or have functionality restrictions.
A new open-source framework called instaVerse makes building your dynamic 3D environments easy. The background can be generated in response to AI cues, and players can then create their avatars to explore it. The first step in making a world in InstaVerse is picking a premade layout. Forests, cities, and even spaceships are just some of the many premade options available. After selecting a starter document, an AI assistant will guide you through the customization process. A forest with towering trees and a flowing river are just one of the many landscapes instaVerse may create at your command. Characters can also be generated in your universe. Humans, animals, and even robots are all included in the instaVerse cast of characters. Once a character has been created, you can use the keyboard or mouse to direct its actions. While InstaVerse is still in its early stages, it shows great promise as a robust platform for developing interactive 3D content. It’s simple to pick up and use and lets you make your special universes.
Sketch is a web app for turning sketches into GIF animations. It’s a fun and easy method to make unique stickers and illustrations to share on social media or use in other projects. Using Sketch is as easy as posting your drawing online. Then, you may utilize the drawing tools to give your work some life with some animation. Objects can be repositioned, recolored, and given custom sound effects. You can save your finished animation as a GIF after you’re satisfied. Sketch is a fantastic program for both young and old. It’s a terrific opportunity to show off your imagination and get a feel for the basics of animation simultaneously. In terms of ease of use, Sketch is excellent. Sketch makes it easy to create beautiful animations, even if you have no prior experience with the medium. With Sketch’s many tools, you can design elaborate and intricate animations. You can save your finished animation as a GIF after you’re satisfied. After that, your animation is ready for sharing or further use.
NeROIC can reconstruct 3D models from photographs as an element of AI technology. NeROIC, created by a reputable tech company, has the potential to transform our perceptions and interactions with three-dimensional objects radically. NeROIC can create a 3D model of the user’s intended message using an approved image. The video-to-3D capabilities of NeROIC are comparable to its image-to-3D capability. This means a user can create an interactive 3D setting from a single video. Because of this, creating 3D scenes is faster and easier than ever.
The discipline of computer science concerned with creating 3D models from 2D photographs is advancing quickly. Deep learning-based techniques may be used to train point clouds and 3D meshes to depict real-world scenes better. A potential method, DPT Depth Estimation, employs a deep convolutional network to read depth data from a picture and generate a point cloud model of the 3D object. DPT Depth Estimation uses monocular photos to input a deep convolutional network pre-trained on data from various scenes and objects. Following data collection, the web will use the information to create a point cloud from which 3D models can be made. When compared to conventional techniques like stereo-matching and photometric stereo, DPT’s performance can surpass a human’s. Because of its fast inference time, DPT is a promising candidate for real-time 3D scene reconstruction.
RODIN is quickly becoming the go-to 2D-to-3D generator in artificial intelligence. The creation of 3D digital avatars is now drastically easier and faster than ever before, thanks to this breakthrough. Creating a convincing 3D character based on a person’s likeness has always been more difficult. RODIN is an artificial intelligence-driven technology that can generate convincing 3D avatars using private data such as a client’s photograph. Customers are immersed in the action by seeing these fabricated avatars in 360-degree views.
That part at least seems pretty clear beyond a shadow of a doubt: Generative Enhanced Multimodal Intelligent Network Interface.
The word “Gemini” comes from Latin and means “twins” in German. Some possible meanings in the context of Google’s AI system:
Gemini combines two components: Text and image processing. It is, in a sense, a “twin system.”
Gemini could refer to the „twins“ Sergey Brin and Larry Page, the founders of Google.
Astrology assigns communication strength and flexibility to the zodiac sign Gemini. Gemini as an AI assistant aims to adapt linguistically and situationally.
The name suggests a dual strength or ability. Gemini aims to unite Google’s text and image AI to outperform the competition.
As a twin system, Gemini combines different perspectives and approaches, similar to different human characters. So the name is both an allusion to the system’s integrative capabilities and a promising indication of Google’s ambitions with this AI product.
Why is Google superior?
To do that, you have to understand WHAT treasure trove of data Google is actually sitting on. Here are a few facts:
Google, through its various services such as Google Search, YouTube and others, has an enormous amount of data that is very useful for developing AI systems.
On YouTube alone, over 500 hours of video material are uploaded every day, according to Statista. The total video database is over 30 million hours of video. The subtitles and transcripts of these videos give Google a gigantic text dataset for training language models.
According to a report by ARK Invest, Google owns over 130 exabytes of data. For comparison, 1 exabyte is equal to 1 billion gigabytes. This means that the entire data set comprises more than 130,000,000,000,000,000 bytes of information.
Google Search accounts for a large part of this data. Google says it processes over 40,000 search queries per second. That’s over 3.5 trillion search queries per year. From these queries and the clicked results, Google gains further insights.
Overall, it shows that Google has virtually inexhaustible data resources for AI research. Both the breadth of different types of data and the sheer volume should give Google a significant edge in the AI field.
Google – The Research Giant
In 2020, Google published over 1300 artificial intelligence research papers, according to the Papers with Code database. In 2021, Google increased the number of publications significantly again to over 2000 papers on AI and machine learning.
Topics included:
Computer Vision (image recognition)
Natural Language Processing (NLP)
Speech Recognition
reinforcement learning
Robotics
Multimodal AI
Recommender Systems
Applications in medicine
With over 3300 AI publications in 2020 and 2021, Google has greatly expanded its research output in artificial intelligence. The company is one of the most active players in this research field. This intensive work over the past few years is now being incorporated into the development of Gemini.
According to the AI publication database Papers with Code, Google published more than 1,500 artificial intelligence research papers in 2022 alone. That’s far more than other tech corporations like Meta or Microsoft.
This is a partial selection of Google’s most groundbreaking developments in AI in recent years. The list shows the enormous range of research from machine learning and computer vision to robotics and autonomous systems.
AlphaGo: Go game AI that defeated world champion Lee Sedol in 2016.
BERT (Bidirectional Encoder Representations from Transformers): breakthrough language model for NLP from 2018.
PaLM (Pathways Language Model): enormous language model with 540 billion parameters from 2022
PaLM-SayCan: variant of PaLM that can carry on human-like conversations
Imagen: image generation AI for realistic and creative images
MusicLM: AI for music composition and production
RLHF (Reinforcement Learning with Human Feedback): Reinforcement learning with human feedback
Model Based RL: reinforcement learning with explicit models of the environment
RobustFit: Robust neural network against data noise
T5: Text-to-text transfer transducer for various NLP tasks
ViT (Vision Transformer): Image recognition with Transformer architecture
WAYMO: Autonomous driving and robot cab service
ProteinFold: Protein structure prediction with Deep Learning
FLOOD: AI for flood prediction and prevention
SLIDE: pixel-level image segmentation
Switch Transformers: efficient architecture for very large transformers
MuZero: reinforcement learning without environmental model in games
Meena: conversational AI from 2020
DALL-E & DALL-E 2: text-to-image generation.
When you look at the sheer amount of data Google has collected over the years, it initially makes you dizzy. Over 500 hours of video footage are uploaded to YouTube every day. The total video database is over 30 million hours. Add to that countless search queries, texts, images and conversations. It’s an almost unimaginable amount of data.
Coupled with intensive research activity in the AI field, it adds up to enormous potential. In recent years, Google has produced groundbreaking innovations such as the BERT language model, the AlphaGo Go AI, and the DALL-E image generator. When you put all these puzzle pieces together, things take on almost frightening proportions.
Project: Google Gemini
With the new Gemini AI system, Google now seems to have bundled the essence of these years of data aggregation and research. If the company succeeds in combining all of its AI developments and treasure trove of data in this system, it would be a demonstration of the sheer power of innovation. It will be interesting to see whether Gemini can deliver on this promise. In any case, the expectations are huge – here what we know and what the rumors say:
Facts Google Gemini
There are already some facts from the Google Blog:
Gemini is supposed to be released this fall
Gemini combines text and image generation
Can create contextual images based on text generation
Has been trained with YouTube transcripts
Google lawyers are monitoring the training to avoid copyright issues
Gemini is said to have multiple modalities, e.g., text, image, audio, video
Sergey Brin is involved in development
Rumors
From Reddit and countless other sources on the web, there could be other features as well:
Gemini is said to be capable of AI image understanding and modification
Is said to combine text capabilities like GPT-4 with image generation
Has been developed from the ground up as a multimodal model
Could handle audio, video, 3D renderings, graphics, etc.
Shall learn with user interactions and thus become effective AGI
Architecture could enable lifelong learning
There are concerns about privacy and information leaks between users
Google Gemini and the (then new) AI market:
The AI market situation is likely to change significantly with the introduction of Google Gemini:
For OpenAI:
Strong new competitor for ChatGPT and DALL-E.
Google has significantly more resources and data
OpenAI could lose market share and come under pressure
For Anthropic:
Claude must stand up to Google Assistant with Gemini
Advantage due to focus on security and control
Risk of falling behind
For Microsoft:
Partnership with OpenAI important to compete with Google
Microsoft must further develop Azure AI services
Advantage due to strong cloud infrastructure
For others:
Startups could have a very hard time against Google
Consolidation in the market possible
Significantly higher innovation speed
Overall, competitive pressure in the AI market will increase sharply. With its resources, Google is in a very good starting position to take a leading role with Gemini. It will be more difficult for other providers to keep pace with Google. It remains to be seen whether the high expectations for Gemini are justified.
Google Gemini Conclusion
Google Gemini seems to be a very ambitious AI project that should give the company a competitive edge. The combination of different modalities in one model is new and could improve AI capabilities tremendously. However, there are still many unanswered questions regarding the specific capabilities and data security. The release this fall will show whether Google can deliver on its promise to outperform the competition. Much is still speculation, but expectations are high.
#ai #ki #google #gemini #text #image #multimodal
Artificial intelligence steps in to assist dementia patients with ‘AI Powered Smart So