[AI WEEKLY NEWS RUNDOWN] The White-Collar Purge, Slack Data Sales, and the Space Infrastructure War (April 12-19 2026)

DjamgaMind: Audio Intelligence for the C-Suite (Daily AI News, Energy, Healthcare, Finance)

Full-Stack AI Intelligence. Zero Noise.The definitive audio briefing for the C-Suite and AI Architects. From Daily News and Strategic Deep Dives to high-density Industrial & Regulatory Intelligence—decoded at the speed of the AI era. . 👉 Start your specialized audio briefing today at Djamgamind.com

AI Jobs and Career

I wanted to share an exciting opportunity for those of you looking to advance your careers in the AI space. You know how rapidly the landscape is evolving, and finding the right fit can be a challenge. That's why I'm excited about Mercor – they're a platform specifically designed to connect top-tier AI talent with leading companies. Whether you're a data scientist, machine learning engineer, or something else entirely, Mercor can help you find your next big role. If you're ready to take the next step in your AI career, check them out through my referral link: https://work.mercor.com/?referralCode=82d5f4e3-e1a3-4064-963f-c197bb2c8db1. It's a fantastic resource, and I encourage you to explore the opportunities they have available.

• Full Stack Engineer [$150K-$220K]
• Software Engineer, Tooling & AI Workflow, Contract [$90/hour]
• DevOps Engineer, India, Contract [$90/hour]
• More AI Jobs Opportunitieshere

Job Title	Status	Pay
Full-Stack Engineer	Strong match, Full-time	$150K - $220K / year
Developer Experience and Productivity Engineer	Pre-qualified, Full-time	$160K - $300K / year
Software Engineer - Tooling & AI Workflows (Contract)	Contract	$90 / hour
DevOps Engineer (India)	Full-time	$20K - $50K / year
Senior Full-Stack Engineer	Full-time	$2.8K - $4K / week
Enterprise IT & Cloud Domain Expert - India	Contract	$20 - $30 / hour
Senior Software Engineer	Contract	$100 - $200 / hour
Senior Software Engineer	Pre-qualified, Full-time	$150K - $300K / year
Senior Full-Stack Engineer: Latin America	Full-time	$1.6K - $2.1K / week
Software Engineering Expert	Contract	$50 - $150 / hour
Generalist Video Annotators	Contract	$45 / hour
Generalist Writing Expert	Contract	$45 / hour
Editors, Fact Checkers, & Data Quality Reviewers	Contract	$50 - $60 / hour
Multilingual Expert	Contract	$54 / hour
Mathematics Expert (PhD)	Contract	$60 - $80 / hour
Software Engineer - India	Contract	$20 - $45 / hour
Physics Expert (PhD)	Contract	$60 - $80 / hour
Finance Expert	Contract	$150 / hour
Designers	Contract	$50 - $70 / hour
Chemistry Expert (PhD)	Contract	$60 - $80 / hour

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience HERE.

#DjamgaMind #AIUNRAVELED

Summary: In this mid-April 2026 weekly edition, we analyze a massive structural shift in the modern enterprise. We deconstruct Meta’s plans to fire 8,000 employees to fund AI automation, and the shocking new trend of bankrupt startups selling their former employees’ private Slack messages as AI training data. On the corporate warfare front, we examine Amazon’s $11.5 billion acquisition of Globalstar, the internal boardroom drama at OpenAI (where investors doubt Sam Altman’s leadership), and the leaked memo accusing Anthropic of inflating its revenue by $8 billion. Finally, we discuss Stanford’s confirmation that China has caught up to the US in AI, while testing deep-sea internet cable cutters.

This episode is made possible by our sponsor:

• DjamgaMind: High-Fidelity Intelligence for the C-Suite. Strategic audio forensics in Enterprise Tech, Defense, and Finance. Visit https://DjamgaMind.com .

Important Topics Covered:

• The White-Collar Purge: Meta prepares to lay off 8,000 employees (10% of staff) to fund its “Applied AI” coding agents, while Snap cuts 16% of its workforce. Algorithmic displacement is no longer theoretical.

• Data Capitalism & Privacy: Defunct startups are using new services like “Asset Hub” to sell their former employees’ private Slack messages and emails for hundreds of thousands of dollars to train AI models.

• The Space & Hardware War: Amazon buys satellite firm Globalstar for $11.5B to rival Elon Musk’s Starlink network (which recently caused a blackout for US Navy drones).

• Internal Friction (OpenAI vs Anthropic): Investors question Sam Altman’s leadership at OpenAI. Meanwhile, an internal memo from OpenAI’s CRO accuses Anthropic of deceptive accounting practices, inflating its run-rate by $8 billion via gross cloud-partner sales.

  

• The Stanford Report: The 2026 AI Index confirms China has closed the technological gap with the US, while Nvidia’s CEO publicly panics about the “DeepSeek” model running on Chinese Huawei chips.

• Hardware & Daily AI: A RAM shortage forces Microsoft to hike Surface PC prices by $500; the Internet Archive is blocked by the NYT; and Apple’s $599 MacBook Neo is sold out until May.

🛠️ The AI Executive Toolkit: Stop collecting PDFs. Deploy real infrastructure. Get the hand-picked, forensic-vetted implementation stack built for the C-Suite. 👉 Get the Toolkit: https://DjamgaMind.com/Toolkit

• ElevenLabs: Transform lengthy compliance bulletins into high-fidelity “Audio Intelligence” for your team to consume on the go. (https://try.elevenlabs.io/4z7r3skyymar)

• Google Workspace: Professionalize your firm’s infrastructure with secure, cloud-based collaboration and branded communication. (https://referworkspace.app.goo.gl/Q371)

⚗️ PRODUCTION NOTE: We Practice What We Preach.

AI Unraveled is produced using a hybrid “Human-in-the-Loop” workflow.

OpenAI investors doubt Sam Altman can lead IPO

• Some OpenAI investors are questioning whether Sam Altman should lead the company’s IPO, given his growing focus on side projects and his own admission that he is not excited about running a public company.

• Altman recently asked OpenAI’s board to fund Helion Energy, a nuclear fusion startup he partly owns, and also pushed the company to back rocket maker Stoke Space, even floating an acquisition.

• The current favorite to replace Altman as CEO is OpenAI board chair Bret Taylor, a former Salesforce co-CEO who co-created Google Maps and is known for focusing on one role at a time.

Meta plans to cut 8,000 jobs next month

• Meta is preparing to lay off around 8,000 employees next month, which amounts to roughly 10% of its nearly 79,000 global workforce, according to three sources familiar with the plans.

• The company is planning further layoffs in the second half of the year, though the date and size are not yet settled, and executives may adjust based on developments in artificial intelligence.

• Meta has also reorganized teams in Reality Labs and moved engineers into a new “Applied AI” organization focused on building AI agents that can write code and carry out complex tasks.

China humanoid robot half-marathon to showcase technical leaps

• China is holding its second humanoid robot half-marathon this Sunday in Beijing, with over 300 robots from more than 70 teams racing across 21 km of paved slopes and parkland.

• Almost 40% of the robot participants will navigate the course autonomously this year, a big change from last year when all entrants were remotely controlled by human operators.

• Experts say the skills on display do not translate to widespread commercialisation, noting that Chinese robotics firms still struggle to develop AI software enabling humanoids to match human factory workers.

World partners with Zoom and Tinder for human verification

• World, the verification project co-founded by Sam Altman, announced partnerships with Zoom, Tinder, and DocuSign to bring its “proof of human” technology into dating apps, video calls, and digital signatures.

  Smartphone 101 - Pick a smartphone for me - android or iOS - Apple iPhone or Samsung Galaxy or Huawei or Xaomi or Google Pixel
  

• Tinder will expand its World ID verification integration to global markets including the U.S., after a pilot in Japan, adding an emblem to profiles that confirms a user is a real person.

• A new Concert Kit feature lets artists reserve tickets for World ID-verified humans through Ticketmaster and Eventbrite, aiming to block scalper bots, with 30 Seconds to Mars and Bruno Mars signed on.

Failed startups sell old Slack chats to train AI

• Startups going out of business are selling their internal Slack messages, emails, and other company data for up to hundreds of thousands of dollars to help train AI models.

• A shutdown service called SimpleClosure launched Asset Hub this week, letting failing startups license workplace data, source code, and internal communications after removing personally identifiable information.

• Privacy advocates warn that selling employee Slack messages raises serious concerns, since those internal communications contain identifiable people and were never meant to become AI training data.

Nvidia CEO warns DeepSeek on Huawei chips threatens US

• Nvidia CEO Jensen Huang said on the Dwarkesh Podcast that DeepSeek optimizing its AI models for Huawei’s Ascend chips instead of American hardware would be “a horrible outcome” for the United States.

• DeepSeek has been rewriting its core code to work with Huawei’s CANN framework, moving away from Nvidia’s CUDA ecosystem and breaking the software-hardware dependency that gave America a second layer of control.

• DeepSeek’s upcoming V4 multimodal foundation model is expected to run on Huawei’s Ascend 950PR processor, which would show competitive AI can be built without any American chips in the supply chain.

Anthropic launches Claude Design to challenge Figma

• Anthropic has launched Claude Design, a new product that lets users create visual designs, interactive prototypes, and slide decks through text prompts, directly competing with Figma, Adobe, and Canva.

• The tool is powered by Claude Opus 4.7, released the same day, which supports images up to 2,576 pixels on the long edge and scored 98.5% on XBOW’s visual-acuity benchmark versus 54.5% for its predecessor.

• Anthropic’s chief product officer Mike Krieger resigned from Figma’s board on April 14, just days before the launch, complicating a partnership where Figma had recently built features around Claude Code integration.

OpenAI launches new AI model for life sciences research

• OpenAI has released an early version of GPT-Rosalind, a new AI model designed to help researchers speed up the process of discovering new drugs and turning scientific studies into patient treatments.

  
  
  

• The model is built for life sciences research, including analyzing large volumes of data, and is available as a research preview to select business customers like Amgen, Moderna, and the Allen Institute.

• OpenAI joins a growing number of tech companies trying to show that AI can lead to scientific breakthroughs, with GPT-Rosalind representing its first dedicated push into health-care and drug discovery.

MacBook Neo sells out for April amid high demand

• Apple’s MacBook Neo has completely sold out for April deliveries, with new orders from the company’s website now showing estimated ship dates starting in early May.

• More than a month after its March 11 release, the $599 laptop continues to sell faster than Apple can produce it, and CEO Tim Cook called it the best Mac launch week ever for first-time buyers.

• Retailers like Amazon and Walmart still have some MacBook Neo stock available sooner than Apple, with Amazon offering a 30-day low price of $589.99 and Walmart shipping select colors overnight.

Physical Intelligence says its robot brain learns untaught tasks

• Physical Intelligence published research showing its new model, called π0.7, can direct robots to perform tasks they were never explicitly trained on, a result the company’s own researchers say surprised them.

• In one test, π0.7 figured out how to use an air fryer despite having only two loosely related episodes in its training data, and after step-by-step verbal coaching it successfully cooked a sweet potato.

• The startup has raised over $1 billion at a $5.6 billion valuation and is reportedly in talks for a new round that would nearly double that figure to $11 billion.

New undersea cable cutter risks Internet’s backbone

• A Chinese research ship has tested a new deep-sea device that can cut through submarine data cables at depths of up to 13,123 feet, raising security concerns about undersea infrastructure.

• The device relies on an electro-hydrostatic actuator that powers a diamond-coated grinding wheel strong enough to slice cables armored with layers of steel, rubber, and polymer.

• Chinese military and civilian organizations have filed multiple patents for cable-cutting tools in recent years, while Chinese-registered ships have been linked to damage to subsea cables worldwide.

Anthropic reveals Claude Opus 4.7

• Opus 4.7 improves on Opus 4.6 in advanced software engineering, with particular gains on the most difficult coding tasks, handling complex long-running work with rigor while devising ways to verify its own outputs.

• Though less broadly capable than Claude Mythos Preview — which remains Anthropic’s most powerful and best-aligned model — Opus 4.7 beats Opus 4.6 across a range of benchmarks in coding, vision, and document reasoning.

• Anthropic trained Opus 4.7 with efforts to differentially reduce its cyber capabilities below Mythos Preview, and is using its deployment to test safeguards before an eventual broad release of Mythos-class models.

Musk asks chipmakers to move at “light speed”

• Musk’s team asked chipmaking equipment suppliers like Applied Materials, Tokyo Electron, and Lam Research to move at “light speed,” seeking fast price quotes and delivery timelines for gear needed to build the Terafab project.

• The Terafab joint venture between Tesla and SpaceX plans to start with a pilot line processing 3,000 wafers per month, with a goal to begin silicon manufacturing by 2029, though no fixed orders have been placed yet.

• Samsung proposed giving Tesla more capacity at its planned Texas factory instead of directly supporting Terafab, while Intel said it would join the initiative after Musk visited its Santa Clara office.

Apple sends Siri engineers to AI coding bootcamp

• Apple is sending a group of fewer than 200 Siri engineers to a multi-week bootcamp where they will learn to code using AI tools, according to a report from The Information.

• AI coding tools like Anthropic’s Claude Code and OpenAI’s Codex have already taken off in other parts of Apple, with some teams allocating large budgets specifically for Claude Code.

• The bootcamp comes ahead of WWDC26 on June 8, where Apple is expected to announce a revamped Siri powered by Google’s Gemini models after multiple delays and leadership changes.

Spotify just won $322 million from music pirates it can’t find

• Spotify and the three major labels won a $322 million default judgment against Anna’s Archive after the pirate activist group scraped 86 million songs from Spotify’s platform.

• Judge Jed Rakoff awarded Spotify $300 million and UMG, WMG, and Sony Music $22.2 million collectively, plus a permanent injunction requiring internet service providers to block the website.

• Enforcing the ruling will be difficult because the operators behind Anna’s Archive remain unknown, and the group has previously relaunched on new domain names to avoid shutdowns.

Google and Pentagon discuss classified AI deal

• Google is in talks with the Pentagon to strike a classified AI deal that would let the Department of Defense run its Gemini models in classified settings for all lawful uses.

• During negotiations, Google has proposed contract language that would block its AI from being used for domestic mass surveillance or autonomous weapons without appropriate human control.

• The deal would help Alphabet grow its government ties as the US pushes to embed artificial intelligence into its processes to cut costs and speed up administrative work.

Starlink outage disrupts US Navy drone tests

• A global Starlink outage last August knocked out communications for two dozen unmanned Navy vessels during a drone test off the California coast, halting operations for nearly an hour.

• Internal Navy documents show multiple test disruptions tied to SpaceX’s satellite network, including an April 2025 incident where Starlink struggled under the high data load of controlling several drones.

• The incidents highlight Pentagon risks from depending heavily on a single company, SpaceX, which holds a near-monopoly in low-earth orbit communications and space launches for the US military.

OpenAI launches GPT-5.4 Cyber

• OpenAI is releasing a new AI model called GPT-5.4-Cyber only to vetted security partners, starting with hundreds of organizations and expanding to thousands more in the coming weeks.

• The move comes one week after Anthropic took a similar step by limiting access to its own model, Claude Mythos Preview, over concerns about cybersecurity misuse.

• GPT-5.4-Cyber is designed to identify security holes in software, but like other cybersecurity tools throughout history, it can be used to both attack and defend networks.

Allbirds ditches sneakers for AI, stock surges 373%

• Allbirds, the sneaker company popular in Silicon Valley, sold its shoe brand for $39 million and is now pivoting to AI under the new name NewBird AI, sending its stock up 373 percent.

• NewBird AI plans to become a GPU-as-a-Service and AI-native cloud solutions provider, and it secured a $50 million investment from an undisclosed institutional investor through a convertible financing facility.

• The pivot echoes Long Island Iced Tea’s 2017 rebrand to blockchain, which saw a 275 percent stock jump before NASDAQ delisted it the following year after Bitcoin fever died down.

Snap cuts 16% of workforce, saying AI has reduced need for repetitive work

• Snap is laying off about 1,000 people, roughly 16% of its full-time employees, and closing over 300 open roles, partly because AI tools have reduced the need for repetitive work.

• The company expects these cuts to reduce its annualized cost base by more than $500 million by the second half of 2026, helping establish a clearer path to net-income profitability.

• Affected U.S.-based employees will receive four months of severance, healthcare coverage, and equity vesting, along with career transition support, while international staff will get comparable local packages.

Google Chrome adds reusable one-click Gemini Skills

• Google Chrome now lets users save Gemini prompts as “Skills,” which are reusable one-click shortcuts that can be accessed while browsing on the desktop version of the browser.

• Previously, users had to manually type or copy-paste a prompt each time they wanted Gemini to do something in Chrome, but Skills remove that repeated effort entirely.

• Saved Skills sync across devices through your Google account and can be triggered by typing a forward slash or clicking the plus button, running instantly in the current tab.

Amazon launches AI tool for drug discovery

• Amazon Web Services has launched Amazon Bio Discovery, an AI tool designed to help researchers speed up the process of finding and developing new drug molecules.

• The tool gives researchers access to a library of biological foundation models that can generate and evaluate drug molecules, plus an AI agent that helps set parameters and interpret results.

• AWS vice president Rajiv Chopra said the service is meant to augment scientists, not replace them, addressing a bottleneck created by the growing demand for computational biologists in drug discovery.

‘Google app for desktop’ launches on Windows

• Google has officially launched its “Google app for desktop” on Windows for English-speaking users worldwide, after testing began in September with a Google Labs sign-up requirement.

• The app opens with an Alt + Space shortcut and lets you search the web, Google Drive, local files, and installed apps in a Mac Spotlight-like experience.

• It includes AI Mode, Google Lens, screen sharing for contextual questions, and a floating results window, while a separate Gemini team is working on a macOS version.

Amazon buys Globalstar for $11.6B to rival Starlink

• Amazon announced a deal to buy satellite company Globalstar for $11.5 billion, paying $90 per share, in a move that strengthens its position against Elon Musk’s SpaceX in the satellite internet market.

• Amazon recently rebranded its satellite internet service from Project Kuiper to Leo and has launched more than 240 low Earth orbit satellites since last April with partners like United Launch Alliance and SpaceX.

• The company first announced plans six years ago to build a constellation of thousands of satellites offering high-speed, low-latency internet to consumers, corporations, and governments through square-shaped terminals.

Meta overtakes Google in global ad revenue

• Meta is forecast to pass Google in global ad revenue by the end of 2026, generating US$243 billion compared to Google’s US$239.5 billion, according to Emarketer’s latest outlook.

• Meta’s growth rate is expected to hit 24.1% in 2026, nearly double Google’s 12%, while Google’s share of global digital ad spend has been declining since 2021.

• Emarketer analysts credit tools like Advantage+ and AI-generated ads for pulling more ad dollars to Meta, saying advertisers follow performance rather than reacting to legal risks.

Microsoft is working on OpenClaw-like agent

• Microsoft is building an agent with features similar to the open source OpenClaw tool and plans to integrate it into its existing Microsoft 365 Copilot product for enterprise customers.

• The company already offers cloud-based agents like Copilot Cowork and Copilot Tasks, but the new project may run locally on a user’s computer, much like OpenClaw does.

• Microsoft told The Information the agent would be an always-working version of 365 Copilot that completes multistep tasks over long periods, and it may appear at Build in June.

OpenAI accuses Anthropic of inflating revenue by $8 billion

• OpenAI’s chief revenue officer Denise Dresser accused Anthropic in an internal memo of overstating its $30 billion run rate by about $8 billion by booking gross revenue from cloud-partner sales through AWS, Azure, and Google Cloud.

• Both gross and net accounting methods comply with US GAAP, but OpenAI says its own net reporting of Microsoft sales is closer to public-company standards, a distinction that could matter as both firms prepare for dual IPOs.

• Ramp corporate-card data shows Anthropic at 30.6 percent of enterprise AI-paying customers versus OpenAI’s 35.2 percent, with a projected crossover within two months, adding pressure behind the timing of Dresser’s memo.

Google Search flags ‘back button hijacking’ as spam

• Google Search has updated its spam policy to treat back button hijacking — where sites prevent users from returning to the previous page — as a violation that can lower a site’s ranking.

• Pages caught using this deceptive practice will face manual spam actions or automated demotions, with Google noting it has seen a rise of this type of behavior across the web recently.

• Site owners have until June 15, 2026, to remove any scripts or advertising platform code responsible for back button hijacking, including third-party libraries that may cause the problem without their knowledge.

Microsoft raises Surface PC prices by $500 amid RAM shortage

• Microsoft has increased prices on all Surface PCs by up to $500, with the flagship Surface Laptop 7 and Surface Pro 11 now costing far more than their original 2024 launch prices.

• Sources say the price increases are driven by recent spikes in RAM and component costs, and the Surface Laptop 7 at $1,499 is now $400 more expensive than the new MacBook Air.

• Microsoft’s upcoming wave of refreshed Surface PCs, expected over spring and summer, will likely maintain or even increase these starting prices if component pricing does not improve soon.

Apple tests four smart glasses designs

• Apple is currently testing four different smart glasses designs and could sell its first pair in 2027, with a possible unveiling later this year, according to Bloomberg’s Mark Gurman.

• The four designs include large and slim rectangular frames, plus larger and smaller oval or circular frames, with color options like black, ocean blue, and light brown.

• The glasses won’t have any displays but will let users take photos and videos, answer phone calls, play music, and interact with the long-promised Siri upgrade.

Meta is building an AI clone of Mark Zuckerberg

• Meta is developing an AI clone of CEO Mark Zuckerberg, built as a photorealistic 3D character that can talk with employees and give them feedback in his place.

• The character is being trained on Zuckerberg’s mannerisms, tone, public statements, and recent thinking on company strategies so workers might feel more connected to the founder.

• This project is separate from a previously reported “CEO agent” tool meant to help Zuckerberg retrieve information quickly, and is part of Meta’s broader push into AI.

China erases US lead in AI per Stanford report

• Stanford’s 2026 AI Index Report shows China has closed the AI performance gap with the U.S., with models from both countries now constantly trading places at the top of benchmarks.

• Over 90 percent of notable AI models come from private companies, and firms like Google, Anthropic, and OpenAI have stopped disclosing dataset sizes, training duration, and training code.

• Generative AI adoption has outpaced personal computers, the internet, and smartphones, with 53 percent of the world’s population now using it regularly, though public trust remains low.

Anthropic plots Lovable challenger, leak suggests

• Leaked images suggest Anthropic is building a vibe-coding app builder inside its Claude chatbot, which would compete directly with Swedish startup Lovable in the fast-growing no-code development space.

• The screenshots posted to X show an in-chat tool that lets users generate applications like AI chatbots, photo albums, and landing pages from simple prompts typed into Claude.

• Lovable raised $330m last December at a $6.6bn valuation, and its head of growth Elena Verna recently said she views Big Tech companies as more threatening than rival startups.

Internet Archive faces existential legal threat

• The Internet Archive and its Wayback Machine face a growing legal and access crisis as major news organizations, including USA Today Co. and The New York Times, block its web crawler from archiving their stories.

• Over 100 working journalists, including Rachel Maddow and Taylor Lorenz, signed a letter of support organized by the Electronic Frontier Foundation and Fight for the Future, defending the Wayback Machine’s role in preserving reporting.

• News organizations say they are blocking crawlers broadly to fight scraping bots, with The Guardian citing concerns over potential misuse by AI companies of content sets crawled for preservation purposes.

OpenAI blames Microsoft for limiting cloud partnerships

• OpenAI’s new revenue chief, Denise Dresser, said in an internal memo that the company’s Microsoft partnership has limited its ability to reach enterprise customers on rival cloud platforms like Amazon’s Bedrock.

• Dresser highlighted “staggering” inbound demand since OpenAI announced its Amazon deal in February, as the company fights to catch Anthropic’s Claude model, which multiple industry figures described as dominating enterprise AI.

• The memo also attacked Anthropic’s strategy as built on “fear” and “restriction,” claiming the rival made a “strategic misstep” by not acquiring enough compute, while OpenAI’s own infrastructure ramp is “materially ahead.”