🎧 Listen Ads-Free: Subscribe to DjamgaMind at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIRIA

Summary: In today’s briefing, we analyze the “Decentralization of Intelligence.” We deconstruct OpenAI fast-tracking a custom-built AI smartphone to control the hardware layer, and Apple paying $250M to settle a class action over delayed AI Siri features. We explore the radical shift in compute infrastructure, highlighted by Nvidia and Span mounting liquid-cooled mini AI data centers on the exterior walls of residential homes. We also cover OpenAI’s new MRC networking protocol designed to stretch GPU compute, Anthropic’s new finance agents, publishers suing Mark Zuckerberg personally for piracy, and PayPal joining Coinbase in massive, AI-driven workforce layoffs.

Today’s Sponsor:

• 🛑AIRIA: The ultimate zero-trust AI security layer. Deploy autonomous agents safely without compromising your enterprise data. 👉 Govern your agents: https://airia.com/request-demo/?utm_source=AI+Unraveled+&utm_medium=Podcast&utm_campaign=Q1+2026

Important Topics:

• OpenAI Fast-Tracks Smartphone: OpenAI partners with MediaTek to mass-produce an AI agent phone by 2027, projecting 30 million units by 2028.

• Apple’s $250M AI Settlement: Apple agrees to pay $250M to iPhone users over the delayed rollout of “personalized Siri” features heavily promoted in 2024.

• Home-Based Mini Data Centers: Startup Span teams with Nvidia to mount liquid-cooled XFRA compute nodes on residential homes, bypassing centralized grid strain.

• OpenAI’s MRC Compute Protocol: OpenAI open-sources the Multipath Reliable Connection protocol, using “packet spraying” to prevent massive GPU cluster failures.

• PayPal Cuts 20% of Staff: Following Coinbase’s 14% reduction, PayPal announces plans to cut one in five workers to capitalize on AI-driven productivity gains.

• Publishers Sue Zuckerberg: Five major book publishers launch a class-action lawsuit accusing Meta and Mark Zuckerberg personally of AI book piracy.

• Google’s “Remy” Agent: Google is internally testing Remy, a 24/7 autonomous AI assistant built deeply into the Gemini ecosystem to rival OpenAI’s OpenClaw.

• GPT-5.5 Instant & Quantum Bio: OpenAI launches GPT-5.5 Instant to reduce hallucinations, while IBM uses quantum computing to simulate 12,000-atom protein complexes.

🔗 RESOURCES

The AI landscape moves faster than a hallucinating LLM on a double espresso, which is why I’ve done the heavy lifting for you. Stop scrolling through generic “Top 10” lists and head over to the AI Executive Toolkit at https://djamgamind.com/toolkit

• Find REMOTE AI Jobs (Mercor): Apply Here –

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

• AI Learning App Recommendation: https://apps.apple.com/ca/app/ai-ml-tutor-pro/id1610947211

• Create and Host your own podcast at

  https://rss.com/?via=005875

⚗️ PRODUCTION NOTE: We Practice What We Preach.

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

Apple pays $250M over Siri AI delays

• Apple has agreed to pay $250 million to settle a class action lawsuit over its delayed rollout of the “more personalized Siri” features that were first shown at WWDC 2024, without admitting any wrongdoing.

• The settlement works out to about $25 per eligible device bought in the US between June 10, 2024 and March 29, 2025, though payouts could climb as high as $95 depending on claim numbers.

• The lawsuit accused Apple of promoting AI features “that did not exist at the time, do not exist now, and will not exist for two or more years,” saturating airwaves to build consumer expectations around the iPhone’s release.

OpenAI plans 30 million phones in two years

• OpenAI is fast-tracking an “AI agent phone” and could begin mass production in early 2027, with analyst Ming-Chi Kuo predicting 30 million units will be built between 2027 and 2028 if plans hold.

• The device will run on a customized version of MediaTek’s Dimensity 9600 chipset, with Kuo saying MediaTek will likely be the sole processor supplier, while Qualcomm and Luxshare are also working with OpenAI.

• Kuo said the phone will use a dual-NPU architecture for heterogeneous AI compute, splitting tasks like image enhancement and object detection between two AI processors, plus an image signal processor with enhanced high dynamic range.

OpenAI unveils protocol to stretch compute

From the Deep View

OpenAI is getting creative to deal with the industry’s imminent compute crunch.

On Wednesday, the ChatGPT maker and a coalition of researchers from Microsoft, AMD, Broadcom, Nvidia, and Intel published a paper offering a rare look into company’s training stack, debuting a new compute networking protocol designed to make GPU clusters faster, more reliable and conserve precious compute cycles.

The protocol, which has been in the works for two years, is instrumental in OpenAI scaling the compute that it needs to continue building bigger and better models, noting in a blog post that the networking approach accelerates its vision for Stargate, the company’s long-term effort to garner the compute it needs to build and scale cutting-edge AI.

The paper introduces a protocol called MRC, or Multipath Reliable Connection, which essentially tackles two main issues with the networks that serve as the connective tissue of AI infrastructure: Congestion and failures, Mark Handley, OpenAI’s networking lead, told me. As GPU clusters grow, these are problems that become more arduous to solve.

• This protocol relies on “packet spraying,” said Handley, which essentially scatters data along hundreds of paths in the network simultaneously to prevent any one network link from getting congested. This also reduces the amount of “tiers” in a GPU cluster, resulting in “flatter” networks that use up less of the data center’s compute and power.

• When handling network failures, MRC detects and reroutes when paths go down in microseconds. This allows GPU clusters to continue training seamlessly, even if parts of the network break down.

• Additionally, the MRC pairs with a protocol called SRv6, or IPv6 Segment Routing, which essentially tells data the exact path it needs to take through a network, rather than forcing the network switches to do the routing work themselves, further reducing the energy requirements of these switches and the data center more broadly.

“We want to use as much compute as we can get, but also we want to make sure that we’re using it efficiently and effectively, and this is a critical component of that,” Greg Steinbrecher, OpenAI’s workload lead, told The Deep View in an exclusive interview.

The protocol is already in use in OpenAI and Microsoft’s largest training clusters, including the Oracle site in Abilene, Texas, and in Microsoft’s Fairwater supercomputers, and has been used to train GPT-5.5 and other models.

When implemented, this new protocol introduces several major downstream advantages, Steinbrecher told me. Conventional, large-scale AI training jobs are a “failure amplifier” for GPU clusters, he said: If one thing goes wrong, the ripple effect forces the process to grind to a halt, leaving GPUs to sit idle. Network congestion, additionally, slows the rate at which researchers can innovate.

MRC circumvents these issues, allowing OpenAI to “turn the crank on our entire research pipeline much faster,” Steinbrecher said. “That allows us to make better use of the resources that we have.”

The MRC specification is available through the Open Compute Project under an open license. Steinbrecher emphasized the importance of this, claiming that this protocol is not one in which OpenAI is trying to “differentiate,” but rather move the entire industry past what they consider a legacy bottleneck. Handley said that the infrastructure industry has reached a point where it’s worth establishing open standards, “as opposed to each of these large companies doing their own thing.”

“Several players in the industry have their own in-house implementations of protocols … that type of market fragmentation is bad for the networking industry,” Steinbrecher said. “You want everyone’s energy going in one direction and pushing together, and then everyone moves faster as a result.”

Quantum, AI spur biological breakthroughs

From the Deep View

For years, IBM has been at the forefront of quantum computing. At this year’s Think conference, that ambition was on full display.

During the opening keynote, IBM CEO and Chairman Arvind Krishna highlighted quantum’s potential with the technology having the capability to unlock new discoveries at an incredibly quick pace — including AI developments.

“Quantum can help uncover what AI cannot yet compute, then AI learns from the quantum and can make faster and faster progress on algorithms and on computations to give you a state of where we are,” said Krishna.

To showcase the tangible use cases of quantum computing, IBM highlighted a biological research milestone achieved with the Cleveland Clinic and Riken: using two of the IBM quantum computers and two of the world’s most powerful supercomputers, the companies were able to simulate protein complexes spanning up to 12,635 atoms. In October 2024, it was only able to simulate 10.

• This is important, as the molecules in your body are proteins, or the “workhorse in the cell” that allow people to exist every day, as Serpil Erzurum, EVP and chief research and academic officer at the Cleveland Clinic’s Lerner Research Institute, explained during the keynote.

• Understanding the 3D structure and motion of a protein is key in biological research, as it helps researchers understand how a drug candidate could bind to a protein and develop effective drugs. Yet, it has remained a challenge as classical computers can only approximate solutions. Erzurum emphasizes that this development is “a moment.”

“Everyone will want to see what these structures look like to understand biology, disease, what’s going wrong if it’s not working, and more importantly, what can I make to fit into the three dimensional structure, to change the structure of that protein–because that’s therapy, and that can make a difference in life,” said Erzurum.

Another example Erzurum noted is using quantum computing and machine learning to dramatically speed up the identification of which treatments a harmful microbe is sensitive to, potentially saving lives given that infections remain a leading cause of death globally.

In a separate Q&A with analysts and select press, Krishna did make it clear that in the next three years, he does not see quantum as replacing either AI or classic CPUs, but rather it will solve problems the two cannot solve, such as the modeling molecules example.

ChatGPT accuracy improves in new model push

From The Deep View

OpenAI’s Instant model, the lightweight option available to everyday users, just got an upgrade.

On Tuesday, OpenAI launched GPT-5.5 Instant, which was updated to be more reliable, with the company claiming it boasts “significant improvements in factuality across the board.” This claim is particularly notable given that it is the default model for its hundreds of millions of daily users.

Benchmark performance shows that GPT-5.5 Instant outperforms GPT-5.3 Instant, the model currently used in ChatGPT, across multimodal reasoning, document parsing, and science and math evaluations. In everyday performance, OpenAI says the updates mean that users experience:

• Less hallucinations: GPT-5.5 Instant produces 52.5% fewer hallucinated claims than GPT-5.3 Instant on high-stakes prompts in areas like medicine, law, and finance and inaccurate claim reduction by 37.3% on challenging conversations users had flagged for factual errors

• More capable: The model is generally smarter and more capable across photo and image models, STEM questions, and choosing when to fetch information from the web

• Tighter responses: More to the point responses that don’t lose substance

• More personalization: It is more effective at using context from either past chats or connected files and Gmail by being faster at searching past conversations. This is rolling out to Plus and Pro users on the web, and coming soon to mobile. It will expand to more plans in the coming weeks

OpenAI also introduced memory sources across all ChatGPT models, which give users the ability to see the context used to personalize their responses and then delete or correct it if something is outdated or not wanted to be cited. Essentially, it’s giving users more control of how their past chats are used and referenced.

The memory sources feature is rolling out to all consumer plans on the web and soon on mobile. GPT-5.5 Instant is rolling out to all ChatGPT users, replacing GPT-5.3 Instant as the default model, while paid users can still access GPT-5.3 Instant for three months before being retired. It is also available in the API as “chat-latest.”

Apple opens iOS 27 to rival AI models

• Apple will let iPhone, iPad, and Mac owners pick from several outside AI models to run features built into iOS 27, iPadOS 27, and macOS 27, with the shift planned for this fall, according to people familiar with the matter.

• The third-party AI services will handle tasks like generating and editing text and images across Apple’s software, expanding the company’s strategy to turn its devices into a broad AI platform rather than relying on a single provider.

• The sources who described the plans asked not to be identified because the details are private, and the article does not name which outside AI models Apple intends to offer users when the new operating systems ship.

Publishers sue Meta over AI book piracy

• Five book publishers — Hachette, Macmillan, McGraw Hill, Elsevier, and Cengage — along with author Scott Turow, have filed a class action lawsuit accusing Meta and Mark Zuckerberg of pirating copyrighted books to train the Llama AI platform.

• The complaint claims Zuckerberg personally authorized and encouraged the infringement, saying Meta reproduced and distributed millions of copyrighted works without permission or compensation, knowing the conduct violated copyright law.

• Meta spokesperson Dave Arnold pointed to past rulings that training AI on copyrighted material can qualify as fair use, echoing a recent Anthropic case where a judge rejected copyright infringement but floated piracy as a separate path to damages.

Google tests Remy 24/7 AI agent

• Google is testing an AI agent codenamed Remy that aims to turn the Gemini app into a 24/7 personal assistant for work, school, and daily life by taking actions on a user’s behalf.

• According to Business Insider, Remy is built deep into Google’s ecosystem and can monitor things that matter to users, handle complex tasks proactively, and learn preferences over time, positioning it against OpenClaw, which OpenAI acquired.

• Remy is currently in a “dogfooding” stage with employees testing it internally, and while no launch date is set, Google’s I/O event on May 19-20 could serve as the debut for the new agent.

OpenAI launches GPT-5.5 Instant for ChatGPT

• OpenAI has rolled out GPT-5.5 Instant as the new default model in ChatGPT, replacing GPT-5.3 Instant, with the company saying it cuts down hallucinations in sensitive fields like law, medicine, and finance while keeping low latency.

• The model scored 81.2 on the AIME 2025 math test versus 65.4 for the older version, and hit 76 on the MMMU-Pro multimodal reasoning benchmark compared to the predecessor’s 69.2.

• GPT-5.5 Instant can pull from past conversations, files, and Gmail for personalized answers, starting on web for Plus and Pro users, and ChatGPT will now show memory sources that users can delete or correct across all models.

OpenAI fast-tracks ‘AI agent phone’

The Rundown: OpenAI is reportedly accelerating development of its first AI phone, now aiming for mass production in the first half of 2027, which is a full year earlier than previously reported, according to supply chain analyst Ming-Chi Kuo.

The details:

• Kuo says the timeline shift is likely driven by OAI’s IPO ambitions (strong hardware could strengthen investor pitch) and rising competition in AI phones.

• The phone’s standout spec will be its image signal processor, with an enhanced HDR pipeline to improve AI agents’ real-world visual sensing.

• MediaTek is positioned to be the sole chip supplier, with the device using two AI processors to handle vision and language tasks simultaneously.

• Kuo also added that OpenAI’s combined 2027–28 shipments of this phone could touch 30M, if the development stays on track.

Why it matters: Controlling hardware and OS could be the key to a true agentic phone. But if OpenAI’s AI phone is closer than we thought, where does this leave the device it’s building with Jony Ive’s io? OpenAI acquired io last year with much fanfare to go “beyond screens,” but nothing concrete has appeared so far except a few rumors.

Anthropic’s AI agents for finance work

Image source: Anthropic

Anthropic just unveiled 10 ready-to-run AI agents aimed squarely at financial services and insurance — capable of handling work ranging from building pitchbooks and screening KYC files to reviewing earnings and valuations.

The details:

• Each agent comes with task-specific domain skills and instructions, connectors to relevant data sources, and add-on Claude models for sub-tasks.

• Firms can adapt any agent of their choice to their own modeling conventions, risk policies, and approval flows — while staying in the loop 24/7.

• The agents can be used as plugins within Claude Cowork or Claude Code on desktop, or as cookbooks, running as Managed Agents on the Claude platform.

• Claude is also getting an add-in for Microsoft 365 as well as data connectors from Dun & Bradstreet, Verisk, IBISWorld, and other financial services partners.

Why it matters: Development, cybersecurity, design, and now finance. Anthropic is going domain by domain, meeting businesses where they are instead of selling a general model and letting them figure it out. Its new $1.5B joint venture alongside Wall Street giants reinforces this strategy, further fueling its race with OpenAI.

Home-based ‘mini’ AI data centers are coming

California startup Span is teaming up with Nvidia to install mini AI data centers on the walls of residential homes and small businesses, tapping unused electrical capacity on local grids to meet surging AI compute demand.

The details:

• Span has developed XFRA, small compute nodes that mount on the exterior walls of homes, alongside accompanying HVAC and electrical systems.

• Nvidia is providing its liquid-cooled RTX PRO 6000 Blackwell Server Edition GPUs to power each XFRA box, ensuring noiseless computing for AI workloads.

• Span told CNBC it can install 8,000 XFRA units 6x faster and at one-fifth the cost of building a comparable 100MW centralized data center facility.

• Currently, the company is working with PulteGroup, one of the largest U.S. homebuilders, to test the box and its economics in newly built communities.

Why it matters: Grid strain from data centers is real, and Span’s boxes could spread the load while tapping only unused capacity. But public response is an open question — not all will love the idea of a data center box mounted where kids play, especially when alternatives like ocean– and space-based data centers are also in sight.

Google wants you to use local AI:

Chrome users are annoyed that Google is installing the Gemini Nano AI model on their machines. Google protested the protests, saying that it has “offered Gemini Nano for Chrome since 2024 as a lightweight, on-device model [that] powers important security capabilities like scam detection and developer APIs without sending your data to the cloud.” Not enough to silence critics, but at least you can turn the feature off.

DeepSeek set could raise $4B:

Chinese AI lab DeepSeek, fresh off the release of its V4 Pro model, is looking to close a massive funding round. With $3 to $4 billion potentially coming its way, the well-known AI model maker could earn a valuation as high as $50 billion in the raise. While DeepSeek is looking to raise more private capital (including cash from China’s national AI fund), two of its rivals (Z.ai and MiniMax) went public earlier this year. DeepSeek, follow suit.

Tech layoffs keep coming:

After American crypto giant announced it would excise 14% of its humans, fintech behemoth PayPal has plans to delete around 20% of its own staff. That’s one in seven workers from Coinbase gone overnight, and one in five from PayPal over the next few years. Why is PayPal pursuing such stark layoffs? It wants faster decision-making and to take advantage of AI-driven productivity gains. Sounds familiar!

What Else Happened in AI on May 06th 2026?

OpenAI’s GPT-5.5-Instant started rolling out to all ChatGPT users, bringing improved performance, stronger memory, and more personalized, concise responses.

Microsoft expanded its Copilot Cowork agentic system to iOS and Android, while adding built-in skills for common tasks and data plugins for business systems.

Apple agreed to pay some U.S. iPhone buyers a collective $250M to settle a class action lawsuit over misleading claims about its new AI Siri, but admitted no wrongdoing.

Perplexity AI launched Computer for Professional Finance, bringing licensed data and 35 dedicated workflows to its agentic system to help analysts handle routine work.

Anthropic reportedly committed to spending $200B on Google’s cloud and chips over the next five years, now making 40%+ of Google’s revenue backlog.

Coinbase CEO Brian Armstrong said the company is cutting 14% of its workforce, ~700 people, as it shifts to AI-native teams, agent-driven workflows, and leaner ops.

🎧 Listen Ads-Free: Subscribe to DjamgaMind at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIRIA

Summary: In today’s briefing, we analyze the “Enterprise Invasion.” We deconstruct the simultaneous announcements from OpenAI and Anthropic, who are launching multi-billion-dollar joint ventures with Wall Street giants (Blackstone, Bain, TPG) to force AI integration into portfolio companies. We explore the massive structural shifts in the corporate world, highlighted by Coinbase laying off 14% of its staff to deploy “one-person AI teams,” and a new study showing 81% of CEOs fear losing their jobs over bungled AI rollouts. We also cover the White House considering pre-release reviews for frontier models, Apple exploring US chip fab with Intel, and Google DeepMind UK staff unionizing against military contracts.

Today’s Sponsor:

• 🛑AIRIA: The ultimate zero-trust AI security layer. Deploy autonomous agents safely without compromising your enterprise data. 👉 Govern your agents HERE

Important Topics:

• Private Equity AI Wars: Anthropic ($1.5B) and OpenAI ($4B) launch massive joint ventures with Wall Street firms to integrate their respective models into mid-sized enterprise portfolio companies.

• CEO AI Paranoia: A Harris Poll/Dataiku study reveals 81% of CEOs fear failing AI deployments could cost them their jobs, citing severe “capability overhang” and legal risks.

• Coinbase AI Layoffs: Brian Armstrong cuts 700 jobs (14%) to flatten the organization, aiming to create “one-person teams” blending engineering, design, and product roles using AI.

• DeepMind UK Unionizes: London-based Google DeepMind workers vote to unionize to protest the lab supplying AI technology to the US and Israeli militaries.

• White House Model Review: The Trump administration weighs an executive order to create a working group to review new AI models prior to release, following Anthropic’s Mythos scare.

• Apple Explores US Chips: Apple is in early talks with Samsung and Intel for US-based chip fabrication to hedge against TSMC capacity strains.

• Ocean-Powered Data Centers: Peter Thiel leads a $140M Series B for Panthalassa, a startup building autonomous, wave-powered floating compute structures.

• Brockman’s $30B Stake: Greg Brockman testifies in the Elon Musk trial, revealing his OpenAI equity is worth nearly $30 billion.

🔗 RESOURCES

The AI landscape moves faster than a hallucinating LLM on a double espresso, which is why I’ve done the heavy lifting for you. Stop scrolling through generic “Top 10” lists and head over to the AI Executive Toolkit at https://djamgamind.com/toolkit

• Find REMOTE AI Jobs (Mercor): Apply Here

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

• AI Learning App Recommendation: https://apps.apple.com/ca/app/ai-ml-tutor-pro/id1610947211

• Create and Host your own podcast at https://rss.com/?via=005875

⚗️ PRODUCTION NOTE: We Practice What We Preach.

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

AI showdown shifts from models to enterprise

From the DeepView

Anthropic and OpenAI are in a battle for the hearts and minds of enterprises. Now, both companies have enlisted reinforcements.

The AI darlings have separately struck multibillion-dollar deals with major Wall Street firms that aim to accelerate AI adoption in businesses. The deals signify the heated fight between OpenAI and Anthropic to cement themselves as the vendor of choice for enterprise as these firms race towards profitability.

News of both deals broke on Monday within minutes of one another:

• Anthropic inked a partnership with Blackstone, Hellman & Friedman, and Goldman Sachs to form a new “AI-native enterprise services firm” with the goal of integrating Claude into companies’ core operations. The firm is a standalone entity utilizing Anthropic’s engineering. It expands on a $200 billion deal these firms struck in April and is worth around $1.5 billion, according to The Wall Street Journal.

• OpenAI, meanwhile, raised more than $4 billion from a host of investors, including TPG, Brookfield, and Bain Capital, for an entity focused on helping businesses deploy AI into their operations, according to Bloomberg. The joint venture, which is aptly named The Deployment Company, is reportedly valued at $10 billion.

The rivalry between OpenAI and Anthropic has never been more heated. The companies have been leapfrogging each other’s models in capabilities for months and pushing new models and upgrades for Codex and Claude Code that aim to capture developer and enterprise attention.

The flurry of improvements also come ahead of each of these firms’ expected IPOs, with both companies anticipating going public in late 2026. With so many eyes on them and dollars flowing towards them, lining up contracts for businesses to use their tech is critical.

Report: AI pressure is rattling CEOs

From the DeepView

AI is making stakeholders excited and CEOs nervous.

A recent global study of 900 CEOs, conducted by Harris Poll and AI firm Dataiku, finds that 81% of those surveyed fear that failing AI deployments could cost them their jobs this year. Despite the high stakes of getting AI right, many still worry about what happens when it gets things wrong: 34% said they wouldn’t allow AI to make decisions without human approval.

The pressure on these executives is coming in from all sides:

• Around 62% of CEOs report that their boards are actively pressuring them to deliver measurable AI-driven outcomes.

• Meanwhile, 75% reported that they believe a fellow CEO will be ousted as a result of a bungled AI rollout.

• CEOs are largely concerned about competitors getting ahead: 56% admit their competitors have stronger AI strategies than they do.

The problem can be summed up by a phrase that’s been floating around since last year: Capability overhang. There is no question that these models are capable of incredible things, but actually deploying them is another ballgame altogether.

And despite the pressure, there are still major trust issues around the risks that this tech presents. 79% worry about the legal risks and ramifications of using the technology, and 57% worry about the inability to trace an AI output back to its source. Concerns about these risks are causing holdups: 51% reported delaying their rollouts due to regulatory issues.

“Every enterprise now has access to powerful AI,” Florian Douetteau, CEO and co-founder of Dataiku, said in a statement. “The differentiator is whether they can turn that power into reliable business decisions. That is the cognitive dissonance happening in the C-suite right now.”

Run of Show: OpenAI vs. Elon, AI FDA, Coinbase Layoffs, Digesting Meta Earnings

by John Coogan

OpenAI vs. Elon Trial Continues

It’s week two now and the Oakland courthouse continues to deliver top-tier tech drama. Elon testified for more than seven hours last week, now Greg Brockman is getting grilled over personal financial incentives, his $30B OpenAI stake, and links to various angel investments.

U.S. Government Will Review AI Models

OpenAI and Anthropic actually signed on two years ago to give the U.S. government early access to unreleased frontier AI models for national-security testing. Now, Google, Microsoft, and xAI have agreed as well. I’m wondering why Meta is not in on this? The Commerce Department’s AI safety body can test versions with reduced or disabled safeguards to test for dangerous capabilities like cyber, bio, or national security risks. Probably good news for everyone? (George Hotz doesn’t think so.)

Digesting Meta Earnings

Meta’s core ad business is ripping, Q1 up 33%. Strong ad impressions and margins, but the market is worried that the beautiful cash machine is turning into an AI capex furnace. $125-145B is a lot of capex, especially for a company without a cloud business that can resell capacity. Even xAI is finding it hard to drive demand for near-frontier, but slightly lagging models, but Elon has moved quickly to solve the problem with a Cursor partnership. The stock is cheap right now, but investors are unclear about the payoff around all the AI spend. There’s also legal/regulatory risk. The cash machine is a really good cash machine though.

Thiel-backed startup brings AI data centers to sea

From The Rundown:

The Rundown: PayPal and Palantir founder Peter Thiel just led a $140M Series B for Panthalassa, an Oregon-based startup that builds autonomous floating compute structures powered by ocean waves — reportedly valuing the company at nearly $1B.

The details:

• Each 85-meter steel node bobs in open ocean, converting wave motion into electricity for onboard AI chips, all cooled naturally by seawater.

• Once deployed, the nodes can steer themselves to remote waters using only their hull shape (no engines) and beam AI results back via SpaceX’s Starlink.

• The raise will finish a pilot factory near Portland and deploy the first wave-powered compute nodes in the Pacific Ocean, with commercial rollout in 2027.

• Thiel told the Financial Times that “extraterrestrial solutions (to compute) are no longer science fiction” and that “Panthalassa has opened the ocean frontier.”

Why it matters: AI data centers have been one of the more controversial AI talking points for the general public, and the hostility towards their construction is growing fast. While both Elon Musk and Google have pushed space-based options, those are still far from reality, making the ocean an interesting and more realistic alternative.

Anthropic co-founder forecasts AI’s self-building era

From The Rundown

The Rundown: Anthropic co-founder Jack Clark published a new blog post on self-improving AI, putting 60%+ odds on AI systems training their own successors before 2029, citing public data showing AI is already handling a range of core R&D tasks.

The details:

• Clark built the case on public papers and benchmark data, charting AI going from near-zero to 100% across core development tasks in under 3 years.

• METR data shows AI’s independent work capability went from 30-second tasks in 2022 to 12 hours in 2026, with 100-hour runs projected by year-end.

• Clark also pointed to the SWE-Bench benchmark (real GitHub coding), moving from Claude 2 at 2% to Mythos Preview at 93.9% in under three years.

• OpenAI is also targeting an automated research intern by Sept. 2026, while startups like Recursive Superintelligence share similar self-improvement goals.

Why it matters: Self-improving AI systems feel like the inflection point that really makes model development go exponential, and “by the end of 2028” is not that far away. AI is already moving at a speed that is hard for most to process — but once it can reliably build and train itself, all bets are off for how fast things can truly move.

OAI, Anthropic launching rival private equity ventures

From The Rundown

Anthropic announced the formation of a new Claude services company with Blackstone, Hellman & Friedman, and Goldman Sachs, with OpenAI also reportedly raising for its own PE-backed ‘Deployment Company’ the same day.

The details:

• The $1.5B Anthropic venture will focus on mid-sized companies, pairing its Applied AI engineers with teams building custom Claude workflows.

• OAI’s “Deployment Company” will reportedly bring in $4B from 19 investors at a $10B valuation, including TPG, Brookfield, Bain, and SoftBank.

• Both models would give the frontier AI labs direct paths into portfolio companies that often lack the in-house talent to deploy AI systems alone.

Why it matters: The barriers for companies aren’t the models anymore, but actually getting it installed and integrated into messy, large-scale businesses. These paths look more like frontier labs creating their own AI-native consulting firms, with a wealth of private equity portfolio companies ready to get in on the action.

Apple explores Intel and Samsung for US chips

From Techpresso:

• Apple is in early talks with Samsung and Intel to make chips in the US, hoping to secure backup options if TSMC cannot keep up with its demand for main chipsets in its devices.

• TSMC handles roughly three-fourths of global chip fabrication, but the surge in AI chip orders is straining its capacity, pushing phone makers like Qualcomm to already lean on Samsung for parts of the Snapdragon 8 Elite Gen 6.

• Apple executives have reportedly toured Samsung’s $17 billion fabrication plant in Taylor, Texas, which is expected to come online by late 2026, though Bloomberg notes the discussions with both chipmakers are far from finalized.

White House considers tighter regulation of new AI models

From Techpresso:

• The Trump administration is weighing an executive order that would set up a working group of tech executives and officials to review new AI models before release, a sharp turn from its earlier hands-off stance on the technology.

• The shift followed Anthropic’s announcement of Mythos, a model so strong at finding software security flaws that the company withheld it from the public, prompting White House worries about a possible AI-enabled cyberattack on its watch.

• Susie Wiles and Scott Bessent have taken over AI policy after David Sacks left in March, and they are also trying to repair ties with Anthropic after the Pentagon cut off its technology in a fight over a $200 million contract.

Coinbase cuts 14% of staff citing AI

From Techpresso:

Coinbase is cutting about 700 employees, or roughly 14% of its workforce, as Brian Armstrong tries to flatten the company, reduce costs, and push harder into AI-driven productivity. Questions about the framing: AI is not the number one cause, and the team is not insanely huge, but the goal is clearly lightly AI-enabled leanness. Fewer layers and managers, larger teams and more individual-contributions. Some debate here over crypto market cycles and the realities of managing a large organization. Is it actually underrated to be able to scale a workforce up and down quickly?

• Coinbase is laying off roughly 700 workers, about 14% of its staff, as CEO Brian Armstrong pushes a restructuring tied to crypto market volatility and a bigger push to use AI tools across the company.

• The reorg flattens the company to five layers below the CEO and COO, lets leaders manage more than 15 direct reports, and asks managers to contribute more work themselves rather than just oversee others.

• Coinbase will try “one-person teams” blending engineering, design, and product management roles using AI tools, and expects to take $50 million to $60 million in severance costs, according to an SEC filing.

Brockman discloses $30 billion OpenAI stake

From Techpresso:

• OpenAI president Greg Brockman testified Monday that his equity stake in the company is worth close to $30 billion, while also revealing financial links to Sam Altman through investments in Altman-backed ventures and his family office.

• Court records showed Altman gave Brockman an interest in his personal investment fund in 2017, which Musk’s lawyers argue may have hurt Brockman’s independence, citing an email from Musk associate Jared Birchall.

• Brockman also acknowledged holdings in AI chipmaker Cerebras and fusion startup Helion Energy, both tied to Altman’s investment network, as Musk’s lawsuit seeks $150 billion in damages and OpenAI’s return to nonprofit status.

DeepMind UK staff unionize over military AI deals

From Techpresso:

• Google DeepMind workers in London have voted to unionize, asking the company to recognize the Communication Workers Union and Unite the Union as joint representatives in a push to stop the AI lab from supplying its tech to the US and Israeli militaries.

• The effort started in February 2025 after Alphabet dropped a pledge against using AI for weapons and surveillance, and gained urgency after a reported Google deal letting the Pentagon use its AI for “any lawful government purpose.”

• If unionization succeeds, staff plan to demand Google exit its contract with the Israeli military, share more details on how its AI products will be used, and offer some assurance about layoffs driven by automation.

Meta deploys AI to detect underage users

From Techpresso:

• Meta is rolling out AI age-detection technology to Facebook users in the U.S. for the first time, along with 27 European Union countries and Brazil, as part of new measures announced during a child safety trial.

• Parents in the U.S. on Facebook and Instagram will get a notification linking to a blog post explaining how to check and confirm their teens’ ages, sent to all users Meta has identified as a parent.

• The announcement lands during the second phase of a New Mexico trial where the state seeks $3.75 billion in damages and policy changes, prompting Meta to threaten pulling Facebook, Instagram, and WhatsApp from the state.

What Else Happened in AI on May 05th 2026?

• A new filing in the Elon Musk vs. OpenAI case showed that Musk reached out to OAI President Greg Brockman about a potential settlement days before the trial.

• The New York Times reported that the White House is seeking to create a formal review and oversight process prior to companies publicly deploying AI models.

• Sierra raised $950M at a $15B valuation, with the platform saying it now serves over 40% of the Fortune 50 companies for AI-driven customer experiences.

• Roomba creator and former iRobot CEO Colin Angle introduced the Familiar, a bulldog-sized AI pet robot targeting retirees who’ve aged out of pet ownership.

• Anthropic is reportedly in talks to purchase chips from Fractile, a three-year-old London startup focused on more efficient chips for running AI models.

• Elon Musk settles with the SEC for $1.5 million after years-long dispute over his Twitter investment LINK

• Sierra raises $950M as the race to own enterprise AI gets serious LINK

• iOS 26.5 Finally Adds End-to-End Encrypted RCS Between iPhone, Android LINK

• Palantir lifts annual revenue forecast on robust US government demand LINK

• Microsoft’s new research finds an AI ‘paradox’ holding companies back LINK

• The creator of Roomba is back with a furry robot companion LINK

• Data centers at sea: Oregon’s Panthalassa nets $140M led by Peter Thiel for wave-powered AI LINK

• Someone Built an Open-Source ‘Theoretical Mythos’ to Reverse-Engineer Anthropic’s Most Dangerous AI LINK

• Hackers are still exploiting the cPanel bug to gain control of thousands of websites LINK

• Image AI models now drive app growth, beating chatbot upgrades LINK

• Explosion Rocks SpaceX’s Test of Water Deluge System Ahead of Starship Launch LINK

• A college student is suing a dating app that allegedly used her TikTok videos to target men in her dormitory LINK

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIRIA

Summary: In today’s briefing, we analyze the transition from traditional applications to “Software 3.0.” We deconstruct the groundbreaking Harvard study where OpenAI’s legacy o1 model outperformed attending physicians in emergency room diagnoses. We explore the massive shifts in physical capital, including Amazon’s launch of “Supply Chain Services” to wage war on UPS and FedEx, and GameStop’s highly leveraged $55.5 billion bid for eBay. We also discuss the Pentagon’s new classified AI contracts, the ongoing Elon Musk vs. OpenAI trial drama, and the death of the iconic Ask.com search engine.

Today’s Sponsors:

• 🛑AIRIA: The ultimate zero-trust AI security layer. Deploy autonomous agents safely without compromising your enterprise data. 👉 Govern your agents: HERE

• Djamgamind.com

Important Topics:

• AI Beats ER Doctors: A Harvard study shows OpenAI’s o1 model outperforming attending physicians in ER triage and diagnosis (67.1% accuracy vs 55.3%).

• Software 3.0 & Neural Computers: The industry is moving away from “vibe coding” apps toward “fat models” that generate dynamic, ad-hoc interfaces and analyses on demand.

• Amazon Logistics War: Amazon launches “Supply Chain Services,” opening its 100-plane, 80,000-trailer network to B2B shipping to challenge UPS and FedEx margins.

• GameStop Bids for eBay: Ryan Cohen launches a highly leveraged $55.5B bid for eBay, aiming to utilize GameStop’s 1,600 retail stores as fulfillment centers.

• Pentagon AI Expansion: The DoD adds SpaceX, OpenAI, Google, Microsoft, and others to classified networks, while maintaining its security blacklist on Anthropic.

• Musk vs. OpenAI Settlement Texts: Lawyers seek to introduce a text message from Elon Musk floating a settlement while warning Sam Altman he would become “the most hated man in America.”

• Ask.com Shuts Down: Parent company IAC officially shuts down the Ask.com search engine after 25 years online.

• Maryland Bans AI Grocery Pricing: The state becomes the first to ban AI-driven dynamic pricing in grocery stores, imposing $25K fines for personalized markup algorithms.

🔗 RESOURCES

The AI landscape moves faster than a hallucinating LLM on a double espresso, which is why I’ve done the heavy lifting for you. Stop scrolling through generic “Top 10” lists and head over to the AI Executive Toolkit at https://djamgamind.com/toolkit

• Find AI Jobs (Mercor): Apply Here – https://work.mercor.com/?referralCode=82d5f4e3-e1a3-4064-963f-c197bb2c8db1

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

• AI Learning App Recommendation: AI and Machine Learning For Dummies: https://apps.apple.com/ca/app/ai-ml-tutor-pro/id1610947211

• Create and Host your own podcast at https://rss.com/?via=005875

⚗️ PRODUCTION NOTE: We Practice What We Preach.

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

Old AI model tops doctors in ER trial

A Harvard study published in Science just put OpenAI’s o1-preview (released in 2024) through 76 real ER cases, with the AI diagnosing patients more accurately than two physicians, despite using only raw electronic health-record text.

The details:

• The study compared OpenAI’s o1-preview model with two attending physicians across 76 real ER cases and three decision stages of patient care.

• At initial ER triage, the model gave the correct diagnosis 67.1% of the time, compared to 55.3% and 50.0% for the two physicians.

• The two separate physician reviewers tasked with scoring couldn’t tell which diagnoses came from the model and which came from the humans.

• In one case, the AI flagged a rare flesh-eating infection in a transplant patient roughly 12 to 24 hours before the treating doctor caught it.

Why it matters: Millions of people are already using AI daily for health questions, but studies like these are showing the usefulness can also flow the other way to the doctors themselves. If a model generations behind is already beating ER doctors, imagine what the frontier could look like inside the patient care process.

Pentagon announces new AI partners

Image source: Images 2.0 / The Rundown

The Pentagon added 8 AI companies to its classified networks while excluding Anthropic, even as the Washington Post reports the new contracts have the same autonomous-weapons and surveillance limits for which Anthropic was blacklisted.

The details:

• The official agreement list names SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, AWS, and Oracle as the companies added to classified networks.

• The Department of War said the new deals will “accelerate the transformation toward establishing the U.S. military as an AI-first fighting force”.

• DoD CTO Emil Michael told CNBC that Anthropic’s supply-chain risk label still stands, but called its Mythos model a “separate national security moment.”

• Anthropic’s exclusion comes days after the White House came out against a broader Mythos rollout over compute concerns impacting its own access.

Why it matters: The White House seemingly wants to have its cake and eat it too — both continuing to shun Anthropic while also wanting priority access to its Mythos model despite the blacklist. There are also some interesting names on that list, namely Reflection, which raised $2B from 1789 Capital, a Donald Trump Jr.-backed fund.

Ask.com shuts down after 25 years

• Ask.com has officially closed its doors after 25 years online, with parent company IAC confirming on May 1, 2026 that it shut down its entire search business, ending one of the web’s earliest recognizable search brands.

• Born in the late 1990s alongside Google Search, the platform started with a question-and-answer format and a butler mascot named Jeeves before rebranding to Ask.com, but it steadily lost ground to Google’s ranking systems.

• In a farewell message, IAC said “a very great search must come to an end” and thanked its engineers, designers, and the millions of users, noting that Ask’s natural-language approach foreshadowed today’s conversational search and AI tools.

AI beats doctors on diagnoses in Harvard study

• OpenAI’s o1 model matched or beat board-certified emergency room physicians on diagnosis, triage and next-step care decisions in a new Science study, based on six experiments using real data from a Massachusetts medical centre.

• The model stood out during early-stage triage, where it handled uncertainty better than doctors by making stronger use of unstructured notes and partial information, though both humans and AI improved as more data came in.

• Researchers, including Harvard’s Arjun Manrai and commentators from Flinders University, warned that AI cannot read visual, body language or auditory cues, and called for prospective clinical trials covering safety, equity and cost-effectiveness before wider use.

Anthropic nears $1.5B Wall Street venture

• Anthropic is close to sealing a roughly $1.5bn joint venture with Blackstone, Hellman & Friedman, Goldman Sachs, and General Atlantic that will push Claude into the portfolio companies owned by those Wall Street firms.

• Anthropic, Blackstone, and Hellman & Friedman are each anchoring the deal at about $300m, Goldman Sachs joins as a founding investor at roughly $150m, and General Atlantic and others cover the rest.

• The venture follows OpenAI’s DeployCo, which drew $4bn from five PE firms last month with a 17.5 per cent annualised return guarantee, while Anthropic’s smaller structure has no publicly reported guaranteed returns.

Amazon declares war on UPS and FedEx

• Amazon is opening its freight network to businesses in retail, healthcare, manufacturing and other industries through a new service called Amazon Supply Chain Services, letting them move, store and deliver goods by ocean, road, rail and air.

• The offering taps Amazon’s fleet of more than 100 cargo planes, over 80,000 trailers and 24,000 intermodal containers, plus warehouses, and includes distribution, fulfillment, parcel shipping, two-to-five-day delivery, warehousing and inventory forecasting.

• The push targets the business-to-business shipping market, a high-margin segment for logistics firms, and mirrors the playbook of Amazon Web Services, which started in 2006 as internal infrastructure before becoming the biggest cloud provider.

Musk texts Brockman seeking OpenAI settlement

• OpenAI’s lawyers have asked a federal judge to let Greg Brockman testify about an April 25 text in which Elon Musk floated a settlement, then warned Brockman and Sam Altman would become “the most hated men in America.”

• The defense says the text shows motive and bias rather than settlement value, citing Federal Rule of Evidence 408 and arguing Musk is using the Oakland lawsuit to attack a competitor after launching xAI.

• Judge Yvonne Gonzalez Rogers, not the nine-person advisory jury, will decide liability and remedies, which could include removing Altman and Brockman from leadership, disgorgement to the charity, and unwinding OpenAI’s for-profit conversion.

GameStop bids $56 billion for eBay

• GameStop has made an unsolicited, non-binding offer to buy eBay for $125 per share in a cash-and-stock deal worth about $55.5 billion, a 20% premium over eBay’s Friday closing price of $104.07.

• GameStop CEO Ryan Cohen, who told the Wall Street Journal he wants to turn eBay into a bigger rival to Amazon, has built a 5% stake and secured up to $20 billion in debt financing from TD Bank.

• The bid raises feasibility questions since GameStop’s market value was $12 billion versus eBay’s $46 billion, but Cohen plans to cut $2 billion in yearly costs and pitch GameStop’s 1,600 stores as fulfillment infrastructure.

Anthropic, OpenAI raise billions for PE push:

The world’s leading AI labs are raising billions ($4 billion for OpenAI, $1.5 billion for Anthropic) to help bring AI into private equity portcos. No, it’s not surprising that companies owned by the cost-conscious want to replace some (expensive) humans with (theoretically less expensive) AI. The two model giants are dueling for enterprise market share, and neither appears set to lose a single deal without a fight.

Cerebras targets $115-$125 share price in upcoming IPO:

AI chip company Cerebras’s IPO is making good progress, with a new SEC filing stating it intends to raise more than $3 billion at a price well over $100 per share. For recent investors, the price is a win. For investors who bought in even earlier, the price is an even bigger victory.

European foundation models?

SAP is on a buying spree. The software giant intends to snap up Dremio to help its customers prepare their data for AI agents, and is also buying Prior Labs, a German AI lab building tabular foundation models, or TFMs. Unlike LLMs, TFMs are aces at handling structured business data, which is right up SAP’s lane. Even better, SAP intends to invest “more than €1 billion over the next four years to scale [Prior] into a globally leading frontier AI lab for [structured] data.”

Neural Computers

by John Coogan

There’s been rumblings about the potential for a Neural Computer for years now since the AI boom began. The basic idea is that the computer would have no software whatsoever, essentially just an LLM that generates whatever you need as you are using the device. Karpathy put it this way: “Imagine a device that takes raw videos or audio into basically what’s a neural net and uses diffusion to render a UI that is unique for that moment.”

It feels like we’re starting to see glimpses of this now. I most recently felt it while trying to understand Ryan Cohen’s proposal for GameStop to takeover eBay for $55.5B. I haven’t tracked either company closely, and I wanted to quickly understand how the two companies size up. In a pre-ChatGPT world, I would have pulled stats from Google or Yahoo Finance, maybe copied them into a spreadsheet if I wanted to see them side-by-side (although Google and Yahoo both offer company comparison views, they are always a bit tricky to navigate). Then if I wanted to share the findings, I could screenshot the sheet, or if I was feeling really fancy, design a slide linked to the data. Now this whole process is a single prompt.

Prompt: “Do a bunch of research on GameStop and eBay’s valuation and key financial metrics, things like growth rate, top line, earnings, revenue, valuation, how the multiples fit together. Build a nicely designed side-by-side comparison of the two companies.” – ChatGPT 5.5 Thinking w/ Images

It’s not a perfect result, I’d probably not use red for all of GameStop’s data because red is usually saved for negative numbers. It’s a good start though and I could easily ask for that change. And obviously you could take this a lot further, delivering an updated image every quarter after earnings drops, or every day if you wanted, or for any other two companies, or for anything else you want. I think the end result is fewer dashboards and more ad hoc analyses delivered on demand to answer the exact question you have at the moment.

Karpathy describes this concept as “Software 3.0” in his Sequoia AI Ascent talk and gave an example of shifting from a vibe-coded app to a fully AI workflow for a project that would annotate menus:

MenuGen is this idea where you come to a restaurant, they give you a menu, and there are usually no pictures. I don’t know what any of these things are — usually 30% or 50% of the things, I have no idea what they are.

So I wanted to take a photo of the restaurant menu and get pictures of what those things might look like in a generic sense. I vibe-coded this app that basically lets you upload a photo, and it does all this stuff. It runs on Vercel, re-renders the menu, gives you all the items, uses OCR for all the different titles, uses an image generator to get pictures of them, and then shows it to you.

Then I saw the Software 3.0 version of this, which blew my mind. It was literally: take your photo, give it to Gemini, and say, “Use Nano Banana to overlay the things onto the menu.”

Nano Banana basically returned an image that was exactly the picture of the menu I took, but it actually put into the pixels the different things in the menu. This blew my mind because actually all of my MenuGen is spurious. It’s working in the old paradigm. That app shouldn’t exist.

The Software 3.0 paradigm is a lot more raw. Your neural network is doing more and more of the work. Your prompt or context is just the image, and the output is an image. There’s no need to have any of the app in between.

I have a few takeaways from this:

First, I think it’s exciting for anyone who’s been hesitant to jump into vibe coding. Frontier models are already able to, in 90% of situations, instantiate exactly whatever’s required to solve an actual problem on the fly under the hood, entirely abstracting away code and tools.

Second, I’m reminded of the 2016 USV blog post “Fat Protocols” which argued that unlike “thin protocols” of the web era like HTTP, FTP, etc. (which accrued minimal value), in crypto, Bitcoin, Ethereum, etc. would be “fat protocols,” do a lot of valuable work, and potentially be more valuable than the application layer that enabled interaction with them. There’s still a bunch of complicated market dynamics around value accrual in the AI value chain, but in terms of doing useful work, the models are certainly getting “fatter” every month to use the USV terminology.

Third, there’s still the question of walled garden jumping. The internet isn’t quite “dead” yet, there’s a lot of good information out there, but many, many platforms are fairly locked down, so writing code, puppeteering a browser running on a Mac Mini, or digging through iMessage locally can still require a different workflow, but that’s more of a legal and business discussion than a technical one. The models will continue to find their way over, under, and through any cracks in the walls of the gardens if users ask politely enough.

There’s still a long way to go here, inference is expensive, everything is slow, and models still make odd mistakes (although less and less these days). I’m still enjoying the image output workflow and it feels like moving up a level of abstraction is more qualitatively binary than a slightly higher score on a particular benchmark.

What Else Happened in AI on May 04th 2026?

OpenAI shipped Codex Pets, animated desktop companions that let you track agent progress without switching back to the app.

OpenAI CEO Sam Altman announced that OpenClaw users can use their ChatGPT subscriptions within the agentic tool, taking a stance against Anthropic’s restrictions.

Maryland signed the U.S.’s first ban on AI-driven grocery pricing, with fines up to $25K for stores caught using personalized shopper data to mark up prices.

SAG-AFTRA secured new AI guardrails in its four-year studio deal, with the guild’s negotiator refusing to sign until Hollywood studios made concessions on AI protections.

A Chinese court ruled that replacing a worker with AI does not legally justify firing them, ordering a tech firm to pay wrongful termination damages.

The Information: XAI Shows How Hard It Is to Use a Lot of GPUs at Once

➞ Lambda response: “The xAI “low utilization” story has people mixing up two different metrics”…

WSJ: GameStop Offers to Buy eBay for $56 Billion

Reuters: Cerebras targets $26.6 billion valuation in US IPO as AI chip demand surges

Cofounder Jack Clark says he thinks there’s a 60% chance of AI reaching recursive self-improvement by the end of 2028

Bloomberg: OpenAI Finalizes $10 Billion Joint Venture With PE Firms to Deploy AI

Roon post on Anthropic as an organization that “worships” Claude goes viral

Mayo Clinic: Mayo Clinic AI helps specialists detect pancreatic cancer up to 3 years before diagnosis in landmark validation study

WSJ: OpenAI Wants to Go Public. First Sarah Friar Needs to Get It to Grow Up.

NYT’s Ezra Klein: Why the A.I. Job Apocalypse (Probably) Won’t Happen

NYT: A.I. Is a National Security Risk. We Aren’t Doing Nearly Enough.

WSJ: Why Almost Everyone Loses—Except a Few Sharks—on Prediction Markets

Patrick Collison: Stripe Atlas hits 100,000 all-time incorporations

Rex Salisbury: Erebor hits $1.1 billion in deposits

Charlie Billelo: 0.1% of the accounts on Polymarket have earned 67% of the profits

Politico Europe: EU accused of wasting €20B on AI computing dreams

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIUNRAVELED

Summary: In this special investigative edition, we deconstruct how artificial intelligence is revolutionizing our relationship with geological disasters. We explore the cutt…

Read more

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIUNRAVELED

Summary: In this weekly briefing, we analyze the intersection of AI, national security, and massive capital expenditure. We deconstruct the Pentagon’s decision to integrate AI from 7 major tech firms into classified networks, and the resulting employee mutiny at Google. We explore the staggering $725 billion Capex guidance from Big Tech and Tim Cook’s warning of a “RAMageddon” memory chip shortage. We also dive into the dramatic courtroom testimony between Elon Musk and OpenAI, the pivot toward physical AI robotics (SoftBank’s Roze, 1X Humanoids, Meta acquisitions), and OpenAI’s strategic shift to Amazon AWS and smartphone hardware.

Important Topics:

• Pentagon AI Integration & Protests: The DoD strikes classified network deals with OpenAI, Google, Microsoft, and others, sparking a 600-employee protest at Google over military AI use.

• $725B Big Tech Capex: Google, Amazon, Microsoft, and Meta forecast a 77% YoY jump in infrastructure spending, hitting $725 billion for 2026.

• Apple’s “RAMageddon”: Apple posts a record $111.2B quarter, but CEO Tim Cook warns that AI industry demand has quadrupled memory chip prices.

• Musk vs. OpenAI Trial: In federal court, Elon Musk accuses OpenAI of “stealing a charity,” while admitting xAI used distillation techniques on OpenAI models to train Grok.

• The Robotics Pivot: SoftBank prepares a $100B IPO for data-center robotics firm “Roze,” 1X opens a US humanoid factory targeting 100,000 robots, and Meta acquires Assured Robot Intelligence.

• Geopolitical Roadblocks: China blocks Meta’s $2B acquisition of Manus AI, and the EU orders Google to open the Android ecosystem to AI rivals.

• OpenAI Hardware & Cloud Pivot: OpenAI ends its Microsoft cloud exclusivity to launch on Amazon Bedrock, while supply chain leaks reveal plans for an OpenAI smartphone by 2028.

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

• 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.

Pentagon signs AI deals with 7 companies

• The Pentagon has struck deals with seven AI companies — SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, and Amazon Web Services — to bring their tools onto the Defense Department’s classified networks at Impact Levels 6 and 7.

• The Pentagon said the agreements will speed up its push to become an “AI-first fighting force,” helping with data synthesis, situational understanding, and warfighter decision-making across complex operational environments and all domains of warfare.

• Anthropic is notably absent after a dispute with the Pentagon over guardrails on military use of its AI tools, which led the department last month to label the company a supply-chain risk, barring it from Pentagon contractors.

Meta acquires robotics AI startup

• Meta has bought Assured Robot Intelligence, a startup founded by ex-Fauna Robotics co-founder Lerrel Pinto and former Nvidia researcher Xiaolong Wang, folding its whole-body robot control models and tactile sensor technology into Meta Superintelligence Labs.

• The deal gives Meta e-Flesh, a tactile sensor that reads deformations in 3D-printable microstructures through magnets and magnetometers, plus Wang’s activation-aware weight quantisation work that shrinks AI models to run on a robot’s limited onboard compute.

• Meta wants to be the Android of humanoids, supplying the intelligence layer while others build the machines, a strategy that competes with Google DeepMind’s Gemini Robotics and its Apptronik partnership but sidesteps vertically integrated makers like Tesla and 1X.

Apple raises Mac mini price

• Apple has quietly removed the $599 base Mac mini from its U.S. online store, pushing the entry price for its smallest desktop up to $799 as the 256GB configuration disappears from standard, education, and military storefronts.

• On Apple’s April 30, 2026 earnings call, Tim Cook said demand for Mac mini and Mac Studio has outpaced supply and will take months to stabilize, driven partly by interest in running AI workloads locally on compact Macs.

• With the Mac mini no longer offering a $599 starting point into macOS, that role shifts to the MacBook Neo, though refurbished listings still carry lower-priced Mac mini units when limited inventory allows.

xAI launches Grok 4.3 with voice cloning

• xAI has released Grok 4.3, a new base large language model with a 1 million-token context window and always-on reasoning, alongside a Custom Voices suite that clones a person’s voice from a reference clip as short as 120 seconds.

• Grok 4.3 costs $1.25 per million input tokens and $2.50 per million output tokens, roughly 40% and 60% cheaper than Grok 4.2, and xAI now charges a $0.05 fee for requests blocked by safety filters.

• The model ranks #1 on Vals AI’s CaseLaw v2 and CorpFin benchmarks, but Andon Labs reported “narcolepsy problems” on Vending-Bench 2, and it scored just 11% on ProofBench for difficult math.

People are finally using Reddit’s search

• Reddit’s long-criticized search function is finally catching on with users, with CEO Steve Huffman reporting a 30% year-on-year jump in weekly search users after the company poured money into its search engine and added AI features.

• Huffman said search DAUs, WAUs, and queries are all up meaningfully year-over-year, crediting the team for better integrating Reddit Answers into the product and calling search a major driver of user acquisition and retention.

• Reddit ended the quarter with 493 million weekly active unique users and 126 million daily active unique users, and posted $663 million in revenue, beating Wall Street’s expectation of $609.8 million.

Apple hits record sales despite chip shortage

• Apple brought in $57 billion from iPhone sales during its record March quarter, contributing to total revenue of $111.2 billion, with Tim Cook crediting strong demand for the iPhone 17 lineup on Thursday’s earnings call.

• Cook cautioned that memory chip costs will climb significantly starting in June due to “RAMageddon,” the AI industry’s heavy appetite for memory chips, which has already quadrupled RAM prices and may push iPhone prices higher.

• Apple’s March spending on memory chips already rose, though the company offset costs by selling stockpiled inventory, and Cook told Reuters there is “just a little less flexibility in the supply chain at the moment for getting more parts.”

Musk says xAI trained Grok on OpenAI

• Elon Musk admitted on the stand in a California federal court on Thursday that xAI trained Grok using distillation on OpenAI models, saying the practice was common across AI companies when asked directly.

• The admission came during Musk’s trial against OpenAI, Sam Altman, and Greg Brockman, where he alleges they broke the original nonprofit mission by shifting the entity to a for-profit structure.

• Musk also ranked the leading AI providers during testimony, placing Anthropic first, followed by OpenAI, Google, and Chinese open source models, and described xAI as a smaller company with just a few hundred employees.

Tesla starts Semi truck mass production after 9 years

• Tesla has kicked off mass production of its Semi electric truck nearly a decade after the 2017 unveiling, with the first big rig rolling off the high volume line at a dedicated plant near Gigafactory Nevada.

• The Semi comes in a 325-mile Standard Range trim priced around $260,000 and a 500-mile Long Range version near $300,000, both packing a 1,072HP tri-motor system that charges at up to 1.2MW on Megachargers.

• Deliveries start later this year and undercut the Freightliner eCascadia ($400,000, 230 miles) and Volvo VNR Electric ($350,000, 275 miles), though Tesla won’t hit the factory’s 50,000-truck yearly peak output.

Meta fires 1,100 AI trainers over Ray-Ban leaks

• Meta has cut ties with Sama, a Kenya-based contractor that trained its generative AI systems using Ray-Ban smart glasses footage, triggering the layoff of 1,108 workers after some spoke out about the recordings they reviewed.

• Sama employees told Swedish newspapers in February that they labeled footage showing banking information, private conversations, naked people in bathrooms, and intimate encounters, often captured from subjects who seemingly did not know they were being recorded.

• Meta says its terms of service cover these details and the glasses need explicit permission to engage AI mode, but Sama workers reported being forced to sit idle under tighter security as the firm hunts for the whistleblowers.

1X opens US humanoid factory targeting 100,000 NEO robots

• 1X has begun full-scale production of its NEO humanoid robot at a new 58,000-square-foot factory in Hayward, California, with plans to build more than 100,000 units per year by 2027.

• The factory uses a vertically integrated model, with 1X designing and making motors, batteries, sensors, structures, and transmission systems in-house, and its first-year run of over 10,000 units sold out within five days of its October launch.

• Each NEO runs on NVIDIA’s Jetson Thor platform for onboard AI inference and is trained using NVIDIA Isaac simulation tools, with customer shipments starting in 2026 through a $20,000 early access program or a $499 monthly subscription.

Big Tech capex hits $725 billion in 2026

• Google, Amazon, Microsoft, and Meta plan to spend a combined $725 billion on capex in 2026, a 77% jump from last year’s record $410 billion, driven by AI infrastructure demand and climbing memory chip prices.

• Microsoft set 2026 capex at $190 billion, with CFO Amy Hood blaming $25 billion of that on memory chip and component costs, while warning the company will stay capacity-constrained through at least 2026.

• Alphabet matched Microsoft’s $190 billion capex guidance after Google Cloud revenue grew 63% to $20 billion and its contract backlog doubled to $460 billion, pushing shares up 7% toward a $4.3 trillion valuation.

White House blocks Anthropic Mythos expansion

• The White House has pushed back on Anthropic’s plan to widen access to its Mythos AI model to around 70 companies and organizations, according to an administration official who spoke anonymously on Wednesday night.

• US officials worry Anthropic lacks the computing power to serve more Mythos users without hurting the government’s own use of the model, which the company says is strong enough to enable dangerous cyberattacks.

• Mythos, unveiled in early April, can reportedly detect and exploit vulnerabilities in critical software, and a small group of unauthorized users on a private online forum gained access the same day Anthropic announced its limited release plan.

Apple reportedly abandons Vision Pro

• Apple has reportedly stopped work on the Vision Pro after weak sales of the M5 chip model released in October, which kept the $3,499 price tag and added a more comfortable head strap, according to MacRumors.

• The product engineering team is being moved to other projects across the company, with a near-term focus on AR glasses to compete with Meta and longer-term work on a cheaper Pro-style successor.

• Apple is also shifting engineering resources toward Siri and Apple Intelligence ahead of WWDC in June, as recent delays to its AI work have hurt the company’s reputation with users and developers.

SoftBank is creating a robotics company that builds data centers

• SoftBank is planning to launch and list a standalone AI and robotics company in the U.S. called “Roze,” which will build data centers and use robotics to make AI infrastructure construction more efficient, the Financial Times reported Thursday.

• Masayoshi Son is leading the push, with executives targeting a roughly $100 billion valuation and an IPO as soon as this year, though the timeline could shift partly due to uncertainties from the conflict in the Middle East.

• Roze could bundle existing energy, land and infrastructure assets from SoftBank’s portfolio along with ABB Robotics, which SoftBank agreed to buy last year, and the listing may help offset its $30 billion-plus commitment to OpenAI.

Uber enters the hotel booking business

• Uber has launched hotel bookings inside its app for US customers, giving access to over 700,000 hotels worldwide through a partnership with Expedia Group, with Vrbo vacation rentals set to join later this year.

• Uber One subscribers will get 20% off a rolling list of 10,000 hotels booked through the app, plus 10% back in Uber Credits on all bookings, as the company pitches the subscription harder.

• CTO Praveen Neppalli Naga said agentic AI tools like Cursor cut the hotel booking feature’s build time in half, alongside other launches including travel mode, a room service hub in Uber Eats, and Eats for the Way.

Spotify introduces verified artist badges to help distinguish humans from AI

• Spotify is rolling out a new “Verified by Spotify” badge that marks human artists in good standing, shutting out AI-generated or AI-persona profiles as streaming sites deal with a flood of machine-made tracks clogging their platforms.

• To qualify for the light green checkmark, artists must show consistent listener engagement, follow platform policies, and display “signals of a real artist,” with over 99 percent of actively sought-out artists verified at launch.

• Spotify is also testing a new context section on artist profiles, described as “nutrition facts,” showing career milestones, release activity, and touring activity in the About section on mobile over the coming weeks.

Elon Musk says OpenAI betrayed its mission

• Elon Musk told jurors in an Oakland federal court that OpenAI abandoned its founding mission when it shifted from a charity to a for-profit company, arguing the pivot amounts to “stealing a charity” and sets a dangerous precedent.

• OpenAI’s lawyer William Savitt countered that Musk himself pushed to restructure OpenAI as a for-profit in 2017 and wanted majority control, saying the lawsuit is really an attempt to hobble a rival to Musk’s own AI company, xAI.

• The three-week trial could reshape OpenAI as it approaches a trillion-dollar valuation and a planned public offering, with Musk seeking a court order to unwind the October for-profit conversion that gave Microsoft a 27% stake and the nonprofit 26%.

OpenAI launches models on AWS

• OpenAI’s models and its Codex coding agent are coming to Amazon Web Services through Amazon Bedrock, the two companies said Tuesday, with general availability expected in the next few weeks for AWS customers to try.

• A new service called Amazon Bedrock Managed Agents powered by OpenAI will let developers build customized agents that remember previous interactions, going beyond the open-weight OpenAI models that came to AWS back in August.

• The news follows Monday’s reworked Microsoft deal letting OpenAI serve customers on any cloud, and builds on a $38 billion AWS commitment from November plus a $50 billion Amazon investment tied to two gigawatts of Trainium chips.

EU says Meta fails underage user checks

• The European Commission has preliminarily ruled that Instagram and Facebook breach the Digital Services Act by failing to diligently identify and block children under 13 from using the platforms, despite Meta’s own age restrictions.

• Regulators said minors can enter false birth dates with no effective checks, Meta’s tools for reporting underage users are hard to use, and the company does not follow up on reports, letting children keep their accounts.

• If confirmed, the findings could lead to a fine of up to 6pc of Meta’s worldwide annual turnover, and the Commission wants Meta to overhaul its risk assessment and bring in age-assurance technologies that are accurate and non-intrusive.

Anthropic unveils Claude for Creative Work

• Anthropic has rolled out Claude for Creative Work, a set of integrations that plug its AI directly into creative software from Adobe, Autodesk, Ableton, Blender, and Splice rather than asking users to adopt a separate tool.

• Each connector targets a specific task: conversational 3D modelling in Autodesk Fusion, natural-language scripting in Blender, real-time visual control in Resolume, prompt-to-3D concepts in SketchUp, and in-app royalty-free sample search through Splice.

• Anthropic has also joined the Blender Development Fund as a patron and is partnering with the Rhode Island School of Design, Ringling College of Art and Design, and Goldsmiths to give students and faculty access to Claude.

China halts new self-driving permits after Baidu outage

• China has stopped issuing new licenses for autonomous vehicles after more than 100 Baidu Apollo Go robotaxis suddenly stalled on the streets of Wuhan on March 31, stranding passengers and snarling city traffic.

• Three agencies including the Ministry of Industry and Information Technology met with officials from robotaxi pilot cities, calling for a full self-review and better safety monitoring, with no clear end date for the suspension.

• The freeze blocks companies from adding robotaxis, starting test projects, or expanding to new cities, and it covers level four vehicles, sending Baidu, Pony AI, and WeRide shares lower in Wednesday trading.

OpenAI missed its own revenue and user growth targets

• OpenAI fell short of its internal goals for ChatGPT users and revenue last year, never reaching its target of one billion weekly active users, while CFO Sarah Friar warned that rising compute costs could outrun incoming revenue.

• The company has committed roughly $600 billion to future data-center spending under Altman’s bet on compute scarcity, and board directors are now questioning why he keeps chasing more computing capacity despite the slowdown.

• Rival Anthropic has quietly passed OpenAI on Forge Global, trading at about $1 trillion versus OpenAI’s $880 billion, and Myriad users give Anthropic a 64% chance of carrying out its IPO first.

Google employees urge Pichai to reject Pentagon AI deal

• Roughly 600 Google workers have signed an open letter asking CEO Sundar Pichai to walk away from talks with the Pentagon that would let the Department of Defense use the company’s Gemini AI models in classified settings.

• The signatories argue that contract wording is not enough protection, pointing to how Anthropic was labeled a “supply chain risk” after refusing “all lawful purposes” language, while OpenAI revised its Pentagon deal to block mass surveillance of U.S. persons.

• The letter follows Google’s recent rewrite of its AI Principles, which in 2018 promised staff the company would not design or deploy AI for weapons or surveillance, language that employees say has since shifted.

Google is testing AI chatbot search for YouTube

• Google is trying out an AI Mode-style conversational search for YouTube, and the experiment is open now to YouTube Premium subscribers in the US who are 18 or older, with plans to expand it to other users.

• An “Ask YouTube” button in the search bar brings up a page that mixes summary text, bulleted milestones, timestamped longform videos, Shorts galleries, and suggested follow-up prompts related to what you searched for.

• In a test about Valve’s new Steam Controller, Ask YouTube got the basics right but incorrectly said the old Steam Controller had no joysticks, a reminder that these AI-built result pages can include factual errors.

Apple plans iPhone Ultra and MacBook Ultra

• Apple is preparing to expand its Ultra branding into new product tiers, with plans for a foldable iPhone Ultra and a touchscreen OLED MacBook Ultra that will sit above the existing Pro lineup through 2027.

• According to a Macworld report citing a source familiar with Apple’s plans, the Ultra name gives the company a place for new form factors without disrupting the Pro lineup, which has covered iPhone, iPad, and Mac for years.

• The Ultra tier lets Apple ship foldable displays and OLED touchscreen Macs in limited quantities at higher prices, fitting its shift toward raising revenue per user as global smartphone growth slows.

EU orders Google to open Android to AI rivals

• The European Commission has told Google it must open up Android so rival AI services can match what Gemini does on the phones, a decision that came out of a specification proceeding started in January.

• Gemini currently gets special treatment at the system level on any Google-powered Android phone, and the commission says too many Android experiences only work with Google’s AI, which must change.

• The order comes from the Digital Markets Act, which labels seven dominant firms as “gatekeepers,” and the commission may force Google to make the Android AI changes this summer despite Google calling it “unwarranted intervention.”

China blocks Meta’s $2B Manus acquisition

• China has blocked Meta’s $2 billion purchase of AI start-up Manus, with the National Development and Reform Commission telling both parties to withdraw from the deal, citing Chinese laws on foreign investment in the company.

• Manus is based in Singapore but owned by Chinese parent Butterfly Effect Technology, and Meta had already absorbed its staff and paid out investors including Tencent Holdings, ZhenFund and Hongshan before the ruling came down.

• A source told the Financial Times the NDRC’s move was “harsh” and meant as a warning against similar follow-on deals, serving as leverage ahead of next month’s planned meeting between Donald Trump and Xi Jinping.

OpenAI plans an AI smartphone to rival iPhone

• OpenAI is building a smartphone to take on the iPhone, according to supply chain analyst Ming-Chi Kuo, who says MediaTek and Qualcomm will supply the chips and Luxshare will handle manufacturing, with mass production set for 2028.

• Kuo argues the phone is the only device that captures a user’s full real-time state, including location, activity, and communication, and that controlling both the operating system and hardware is needed to deliver AI agent services.

• The phone marks a reversal from OpenAI’s previously reported hardware plans with Jony Ive, which focus on a smart speaker, smart glasses, a smart lamp, and earbuds, with the first announcement expected in late 2026.

Musk is about to launch his ‘everything app’

• Elon Musk is about to roll out X Money, a banking and payments platform that turns X into the “everything app” he promised when he renamed Twitter in 2023, according to Bloomberg.

• The finance feature will reportedly offer a savings account with 6 per cent interest and 3 per cent cashback on some transactions, building on a Visa partnership announced last year for a digital wallet and peer-to-peer payments.

• The launch will be limited because X still lacks licences in key states like Massachusetts and New York, and Senator Elizabeth Warren has written to Musk raising concerns about scams, fraud, and data privacy on the platform.

John Ternus to launch 10 new Apple products

• Incoming Apple CEO John Ternus is set to roll out ten new products during his tenure, starting with the iPhone 18 Pro, 18 Pro Max, and the company’s first foldable iPhone at the September hardware event.

• Beyond the iPhone Fold, Mark Gurman reports that Apple’s pipeline includes a Smart Home Hub, a Tabletop Robot with a 9-inch screen on a robotic arm, a Home Security System, smart glasses, and AirPods with cameras.

• Also on the list are an AI Pendant worn as a necklace, a touchscreen MacBook with an OLED display due in late 2026 or early 2027, lightweight AR glasses, and a foldable iPad with a 20-inch screen that may be scrapped.

Microsoft ends exclusive OpenAI cloud deal

• Microsoft and OpenAI have signed a new deal that removes Microsoft’s right of first refusal to serve as OpenAI’s compute provider, ending a key piece of exclusivity between the two companies after years of a tight partnership.

• Microsoft now owns roughly 27 percent of OpenAI’s new public benefit corporation, worth about $135 billion today, and its IP rights to OpenAI models and products stretch to 2032 and now cover post-AGI models.

• OpenAI has agreed to spend another $250 billion on Azure services, can jointly develop some products with third parties, and can sell API access to US government national security customers through any cloud provider.

Meta signs space solar power deal

• Meta has struck a deal with startup Overview Energy for up to 1 gigawatt of solar energy collected in space, aiming to power its artificial intelligence data centers through satellites that orbit Earth and beam electricity back.

• Overview Energy plans to gather sunlight using satellites orbiting Earth and convert it into electricity that can support the grid, offering Meta a new source to meet its growing demand for power.

• The 1 gigawatt volume from the agreement is roughly equal to the output of a single nuclear reactor, showing the scale Meta is chasing as it hunts for ways to feed its data centers.

What Else Happened in AI and Tech from April 27th to May 03rd 2026?

Meta opened its ads platform to third-party AI tools via a new MCP server, allowing advertisers to manage campaigns through Claude, Cursor, or any connected agent.

OpenAI announced that it has already surpassed its 2029 Stargate goal of securing 10 GW of compute, with 3 GW added in the last 3 months.

Elon Musk revealed during questioning in his trial vs. OpenAI that xAI has used distillation techniques to train on OpenAI models.

Anthropic launched the public beta for Claude Security, a system that leverages Opus 4.7 to scan codebases for vulnerabilities and help enterprises generate patches.

Cursor released Security Review, which also deploys autonomous agents to check for vulnerabilities and run scheduled codebase scans with results posted to Slack.

ElevenLabs launched ElevenMusic, a streaming platform with built-in AI remixing and AI-assisted track creation, already hosting 4k+ artists and offering creator payouts.

Two U.S. House committees opened probes into Cursor-maker Anysphere and Airbnb over Chinese AI use, with Composer 2 built on Kimi and Airbnb’s agent on Qwen.

Mistral AI launched Vibe remote agents, cloud sessions that run coding tasks in parallel, powered by the company’s new open-weights Medium 3.5 model.

Google added file creation into Gemini, allowing the model to output formats like Google Docs, Sheets, and Slides, Microsoft Word and Excel files, Markdown, and more.

OpenAI released a Cybersecurity Action Plan to “democratize” AI cyber defense and work with the U.S. government and industry on threat coordination and defender tools.

OpenAI announced that GPT-5.5, Codex, and Managed Agents are now available via Amazon Bedrock, coming a day after its new contract restructure with Microsoft.

NVIDIA released Nemotron 3 Nano Omni, a new open model that can handle vision, audio, and text at 9x the speed of rival open multimodal models.

The WSJ reported that OAI fell short of its targets for revenue and user growth, with CFO Sarah Friar questioning its massive spending — with OAI calling it “ludicrous.”

Anthropic added new connectors for a broader range of creative workflows, including apps like Blender, Adobe Creative Cloud, Autodesk Fusion, SketchUp, and more.

Xiaomi open-sourced MiMo-V2.5-Pro, which ties Kimi K2.6 on Artificial Analysis’ leaderboard, featuring a 1M context window and strong efficiency for agentic tasks.

SpAItial launched Echo-2, a new SOTA world model that turns text or photos into explorable 3D worlds, claiming to beat World Labs’ Marble 1.1 across benchmarks.

The trial between Elon Musk and OpenAI kicked off on Monday with the start of jury selection, with the two sides trading barbs on X ahead of opening statements.

Tech analyst Ming-Chi Kuo said OpenAI is working on its own smartphone alongside MediaTek and Qualcomm, with native AI agents and production likely in 2028.

Adobe opened access to its new Firefly AI Assistant in public beta, letting creators prompt multi-app Creative Cloud workflows while keeping outputs editable.

Alibaba’s new Happy Horse video model rolled out across video platforms, with the release taking the top spot on Artificial Analysis’s video leaderboard.

Taylor Swift filed three federal trademarks for her likeness and voice, joining actor Matthew McConaughey in taking legal action to fight and prevent AI deepfakes.

xAI launched Grok Voice Think Fast 1.0, a new SOTA voice agent that tops speech benchmarks across the board, and is already running Starlink’s phone support line.

Google is investing up to $40B in Anthropic, including $10B now at a $350B valuation, and $30B more if Anthropic hits performance targets, plus 5GW of Cloud compute.

Meta signed a deal with AWS to add millions of its Graviton5 core chips to power agentic AI workloads, making it one of AWS’s top buyers.

The United Arab Emirates announced a two-year plan to deploy agentic AI across 50% of government services, with mandatory AI training for every federal employee.

Cohere agreed to acquire Germany’s Aleph Alpha, with the $20B merger targeting governments and companies wary of relying on U.S. AI giants for critical tools.

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIUNRAVELED

Summary: In today’s briefing, we analyze the critical divergence between AI capability and reliability. We deconstruct the launch of OpenAI’s GPT-5.5, which dominates reasoning benchmarks but struggles with severe hallucinations compared to Claude Opus. We explore the environmental fallout as Alphabet, Amazon, Meta, and Microsoft abandon net-zero pledges in favor of natural gas to power AI data centers. We also cover Moonshot AI’s Kimi K2.6 executing multi-day autonomous coding, Tim Cook’s warning regarding soaring memory chip costs (”RAMageddon”), Elon Musk’s courtroom admission about distilling OpenAI’s models, and a severe Linux vulnerability catching the world flat-footed.

Important Topics:

• The GPT-5.5 Paradox: OpenAI releases GPT-5.5, topping the Artificial Analysis Intelligence Index but exhibiting a significantly higher hallucination rate (85.5%) compared to Claude Opus 4.7 (36.1%).

• Big Tech Climate Collapse: Alphabet, Amazon, Meta, and Microsoft are reverting to natural gas power plants and abandoning net-zero timelines to meet the massive energy demands of AI data centers.

• Kimi K2.6 Open-Weights Autonomy: Moonshot AI updates Kimi K2.6, a 1-trillion parameter model capable of instantiating up to 300 parallel sub-agents for multi-day autonomous coding tasks.

• Apple’s “RAMageddon”: Apple posts a record $111.2B quarter, but CEO Tim Cook warns that AI industry demand has quadrupled memory chip prices, severely constraining the supply chain.

• Musk Admits to Model Distillation: In federal court, Elon Musk admits that xAI used distillation techniques on OpenAI’s models to train Grok.

• Severe Linux Vulnerability: A zero-day flaw named “CopyFail” (CVE-2026-31431) goes public, allowing attackers full control over unpatched Linux servers and Kubernetes containers.

• Meta Fires 1,100 AI Trainers: Meta contractor Sama fires over a thousand Kenyan workers after whistleblowers reveal they were reviewing sensitive Ray-Ban smart glasses footage.

• Strategic Thinking in LLMs: Researchers prove that Gemini and GPT models use more sophisticated, sequential strategic planning in games (like Rock-Paper-Scissors) than human players.

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

• 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.

GPT-5.5 Outperforms, Hallucinates

The latest update of OpenAI’s flagship model sets new states of the art in important benchmarks but has difficulty distinguishing between what it does and doesn’t know.

What’s new: GPT-5.5 is a closed vision-language models that’s built for agentic coding, computer use, and knowledge work. GPT-5.5 Pro is the same model but processes reasoning tokens in parallel during inference. OpenAI set the API prices at roughly double the per-token rates of GPT-5.4.

• Input/output: Text and images in (up to 1 million tokens via API, 400,000 tokens in Codex), text out (up to 128,000 tokens)

• Features: Five levels of reasoning (xhigh, high, medium, low, none), tool use, web search, structured outputs, tool search (API only, loads tools on demand rather than all at once), Fast mode (Codex only, generates tokens 1.5 times faster at 2.5 times the price)

• Performance: Tops Artificial Analysis Intelligence Index and ARC-AGI-2

• Availability/price: GPT-5.5 available in ChatGPT with Plus, Pro, Business, or Enterprise subscription and in Codex for those tiers plus Edu and Go; GPT-5.5 Pro available in ChatGPT with Pro, Business, or Enterprise subscription: GPT-5.5 API $5/$0.50/$30 per million tokens of input/cached/output, GPT-5.5 Pro API $30/$180 per million tokens of input/output with no cached discount

• Undisclosed: Architecture, parameter count, training data and methods

How it works: OpenAI disclosed few details about how it built GPT-5.5. As is typical of high-performance models, the training data was a mix of publicly available data scraped from the web, licensed from partners, and collected from users and human trainers. The model was trained via reinforcement learning to reason before responding.

Big AI’s Plans Strain CO₂ Pledges

Commitments by large AI companies to limit emissions of greenhouse gases are at risk as those companies pursue a massive build-out of data centers, many of which will be powered by fossil fuels in the near term and possibly beyond.

What’s new: Alphabet, Amazon, Meta, and Microsoft have begun to acknowledge that keeping up with projected demand for AI is interfering with earlier plans to stop raising the concentration of greenhouse gases to the atmosphere, Associated Press reported. (Disclaimer: Andrew Ng is a member of Amazon’s board of directors.)

How it works: Electricity consumed by top tech companies has increased significantly over the last few years, and with it their emissions of greenhouse gases that contribute to climate change, despite ongoing efforts to reduce emissions. While they have emphasized clean sources of energy including wind, solar, geothermal, and nuclear, lately they have begun to develop natural gas power plants to meet rapidly rising demand for AI.

Kimi K2.6 Challenges Open-Weights Champs

Moonshot AI’s updated Kimi model handles longer autonomous coding sessions and scales up its multi-agent orchestration relative to its predecessor.

What’s new: Kimi K2.6 is a 1 trillion-parameter vision-language model that performs neck and neck with Qwen3.6 Max Preview and the newly released DeepSeek V4 and falls just behind top closed models. It’s designed to generate code in a plan-write-test-debug loop that can last for days, and it can instantiate hundreds of agents that collaborate on a single task. It also produces fewer hallucinations than its predecessor.

• Input/output: Text, images, and video in (up to 256,000 tokens), text out (up to 98,000 tokens)

• Architecture: Mixture-of-experts, 1 trillion parameters total, 32 billion active per token, MoonViT vision encoder

• Features: Tool use, web search, native INT4 quantization, “preserve thinking” mode, agent swarm

• Performance: Tops other open-weights models on the Artificial Analysis Intelligence Index but trails leading proprietary models

• Availability/price: Weights free to download from Hugging Face under a modified MIT license that permits commercial uses with attribution for products with more than 100 million monthly active users or more than $20 million in monthly revenue, free chat interface at kimi.com and Kimi mobile app, API access via Moonshot $0.95/$0.16/$4.00 per million input/cached/output tokens

• Undisclosed: Training data and methods

How it works: Kimi K2.6 reuses the architecture introduced with Kimi K2 and refined in Kimi K2.5, including the multi-headed latent attention (an attention variant that reduces memory requirements by compressing keys and values) and MoonViT vision encoder (400 million parameters). Moonshot has not disclosed how Kimi K2.6 differs with respect to training data and methods.

Strategic Thinking in LLMs vs. Humans

While large language models can behave in human-like ways, the similarities are superficial. A simple strategy game revealed clear differences in their strategic approaches.

What’s new: Caroline Wang and colleagues at University of Texas at Austin and Google interpreted patterns of decision-making by humans and LLMs as they played the classic game of rock-paper-scissors. They found that LLMs sometimes model their opponents with greater sophistication than people do.

Key insight: Given recorded gameplay, an LLM can iteratively improve code that predicts a player’s next move. If the code predicts the player’s actions with significant accuracy, we can assume that its decision-making algorithms are functionally similar to those the player used. Computer code is interpretable, making it possible to discern such algorithms and compare those used by humans and LLMs.

How it works: In games of rock-paper-scissors, he authors pitted individual LLMs (Gemini 2.5 Pro, Gemini 2.5 Flash, GPT-5.1, and GPT-OSS 120B) against each of 15 preprogrammed bots of varying complexity. They recorded each player’s moves in 20 games of 300 sequential rounds each. Previous work provided records of similar records of games between humans and the same bots. The authors tracked the round-by-round choices made by each player — AI and human — and whether they won, lost, or tied. Then they used AlphaEvolve, an agentic method that iteratively optimizes code through an evolutionary process, to improve Python programs that predicted the next move for each LLM individually and humans as a group.

• AlphaEvolve initially processed the game data using a simple template program written by the authors. In each of an undisclosed number of evolutionary steps, Gemini 2.5 Flash proposed modifications to improve a function that balanced simplicity (as measured by Halstead effort) and evaluation likelihood (how well a program predicted a player’s choices).

• For each player, the authors selected the simplest program that achieved near-maximum predictive accuracy within a small margin from the best. Each program produced the best evaluation likelihood (higher is better) for the player it had evolved to predict. That is, it represented its corresponding player’s behavior better than that of any other player.

Results: Using game data that AlphaEvolve didn’t process, the authors compared how well each program predicted the other players’ moves. Then they examined the programs to determine what strategies each player used.

Why it matters: While researchers have found ways to understand some aspects of neural network behavior, large language models remain black boxes in many ways. Synthesizing code directly from LLM behavior offers a powerful tool to interpret their decision-making.

We’re thinking: It’s tempting to assume that LLMs learn to mimic human behavior as represented by their training data. Finding that they can encode a gaming strategy more systematically than the average human demonstrates a different sort of learning.

Apple hits record sales despite chip shortage

• Apple brought in $57 billion from iPhone sales during its record March quarter, contributing to total revenue of $111.2 billion, with Tim Cook crediting strong demand for the iPhone 17 lineup on Thursday’s earnings call.

• Cook cautioned that memory chip costs will climb significantly starting in June due to “RAMageddon,” the AI industry’s heavy appetite for memory chips, which has already quadrupled RAM prices and may push iPhone prices higher.

• Apple’s March spending on memory chips already rose, though the company offset costs by selling stockpiled inventory, and Cook told Reuters there is “just a little less flexibility in the supply chain at the moment for getting more parts.”

Musk says xAI trained Grok on OpenAI

• Elon Musk admitted on the stand in a California federal court on Thursday that xAI trained Grok using distillation on OpenAI models, saying the practice was common across AI companies when asked directly.

• The admission came during Musk’s trial against OpenAI, Sam Altman, and Greg Brockman, where he alleges they broke the original nonprofit mission by shifting the entity to a for-profit structure.

• Musk also ranked the leading AI providers during testimony, placing Anthropic first, followed by OpenAI, Google, and Chinese open source models, and described xAI as a smaller company with just a few hundred employees.

Tesla starts Semi truck mass production after 9 years

• Tesla has kicked off mass production of its Semi electric truck nearly a decade after the 2017 unveiling, with the first big rig rolling off the high volume line at a dedicated plant near Gigafactory Nevada.

• The Semi comes in a 325-mile Standard Range trim priced around $260,000 and a 500-mile Long Range version near $300,000, both packing a 1,072HP tri-motor system that charges at up to 1.2MW on Megachargers.

• Deliveries start later this year and undercut the Freightliner eCascadia ($400,000, 230 miles) and Volvo VNR Electric ($350,000, 275 miles), though Tesla won’t hit the factory’s 50,000-truck yearly peak output.

Severe Linux threat catches world flat-footed

• Researchers at security firm Theori released attack code on Wednesday for a Linux bug called CopyFail that lets regular users seize full control of almost every version of Linux, leaving companies racing to protect servers and personal machines.

• Known as CVE-2026-31431, the flaw was quietly reported to Linux kernel maintainers five weeks ago and fixed in updates such as 7.0 and 6.19.12, but most Linux distributions had not yet shipped those fixes when the code went public.

• One script works against every unpatched system without changes, letting attackers take over shared servers, escape Kubernetes containers that isolate apps, and sneak the code into pull requests so it runs inside automated build and deployment pipelines.

Meta fires 1,100 AI trainers over Ray-Ban leaks

• Meta has cut ties with Sama, a Kenya-based contractor that trained its generative AI systems using Ray-Ban smart glasses footage, triggering the layoff of 1,108 workers after some spoke out about the recordings they reviewed.

• Sama employees told Swedish newspapers in February that they labeled footage showing banking information, private conversations, naked people in bathrooms, and intimate encounters, often captured from subjects who seemingly did not know they were being recorded.

• Meta says its terms of service cover these details and the glasses need explicit permission to engage AI mode, but Sama workers reported being forced to sit idle under tighter security as the firm hunts for the whistleblowers.

1X opens US humanoid factory targeting 100,000 NEO robots

• 1X has begun full-scale production of its NEO humanoid robot at a new 58,000-square-foot factory in Hayward, California, with plans to build more than 100,000 units per year by 2027.

• The factory uses a vertically integrated model, with 1X designing and making motors, batteries, sensors, structures, and transmission systems in-house, and its first-year run of over 10,000 units sold out within five days of its October launch.

• Each NEO runs on NVIDIA’s Jetson Thor platform for onboard AI inference and is trained using NVIDIA Isaac simulation tools, with customer shipments starting in 2026 through a $20,000 early access program or a $499 monthly subscription.

Mine hunting:

The US Navy is turning to Domino, a startup that helps companies build, deploy, and manage AI to find mines in the Strait of Hormuz. With a fresh $100 million contract, Domino will help the American military “make underwater mine detection faster, more accurate, and less dependent on human sailors.” The Navy-Domino partnership underscores the rising importance of AI in military contexts, including keeping global shipping lanes free from explosives.

Stablecoins are big business:

Polymarket’s rapid growth was partly enabled by Fun, a startup that provides crypto and fiat on- and off-ramps for customers. The fintech upstart now processes $18 billion in annual payment volume, enough for venture capitalists to put $72 million into its coffers as part of an outsized Series A. As the stablecoin market grows, so too will demand for financial technology that makes their use simple and safe.

The race to build efficient AI accelerates:

Nebius, a public neocloud, announced today that it plans to purchase Eigen AI. Eigen helps make AI inference and certain AI training activities more efficient. Given that Nebius serves AI compute demand today and there’s not enough to go around, snapping up a smaller company to make existing GPUs stretch further makes good sense. And if you are Nebius and want to defend your valuation, then working to ensure your gross margins are as attractive as possible is simply good business.

The White House’s Anthropic stance gets complicated

The Rundown: The White House is pushing back on Anthropic’s plan to more than double the private sector’s access to its Mythos AI over compute concerns for its own use, just as a national security memo prepares to address parts of the Pentagon feud.

The details:

• Anthropic wanted access expanded from about 50 firms to nearly 120, with U.S. officials citing compute strains that could impact government use.

• A White House AI memo will reportedly push multi-vendor AI adoption for agencies and address some of Anthropic’s worries that led to the initial feud.

• Axios reported that the government action would “allow agencies to get around the supply chain risk designation”, despite the current legal battle.

• GPT 5.5 reached similar cyber capabilities to Mythos, with former AI czar David Sacks saying all frontier models will reach the level in 6 months.

Why it matters: The White House is changing its tune on Anthropic, seemingly largely in part to wanting more access of its own to the powerful Mythos. But with Sec. of War Pete Hegseth saying Thursday that Anthropic is “run by an ideological lunatic”, there is some internal division between wanting to bury the hatchet vs. continuing the fight.

Gemini moves into Google-powered cars

Google is beginning its Gemini upgrade for vehicles with Google built-in, swapping out Assistant for a more conversational system that handles navigation, messages, music, vehicle questions, and hands-free controls across compatible cars.

The details:

• Drivers can ask for changes to car settings like temperature, control the radio, and pull from Google Maps for customized updates or route planning.

• A beta Gemini Live mode supports conversations for learning and brainstorming, with Gmail, Calendar, and Home integrations coming later.

• Gemini can also pull vehicle-specific answers from manufacturer manuals for car assistance and battery status or charging stations for EV cars.

• The rollout comes to compatible cars in the U.S. first, with General Motors also announcing the feature for ~4M of its vehicles from model year 2022 onward.

Why it matters: One day, AI integrations in cars will be as common as a radio (and eventually the systems will all be driving the cars, too) — but for now, we’re still in the infancy of the rollout. These initial features are fairly basic, but a step on the path towards ‘smart car’ systems of the AI age that provide a serious intelligence upgrade.

OpenAI finds source of ChatGPT’s goblin obsession

Image source: OpenAI

OpenAI just traced ChatGPT’s habit of peppering its responses with goblins, gremlins, and assorted fantasy creatures back to a single reward signal in its ‘Nerdy’ personality, which ended up bleeding into model behavior throughout releases.

The details:

• After ChatGPT-5.1’s November launch, ‘goblin’ mentions jumped 175% in user conversations, with ‘gremlin’ up 52% and other creatures seeing similar spikes.

• When OpenAI mapped creature use across personalities, the Nerdy preset lit up, driving two-thirds of all goblin mentions from just 2.5% of traffic.

• Even users who skipped Nerdy got goblins, with fine-tuning loops recycling the creature-favored outputs back into ChatGPT’s default mode.

• OpenAI retired Nerdy in March and shipped GPT-5.5 with a Codex prompt specifically banning goblins, gremlins, ogres, trolls, raccoons, and pigeons.

Why it matters: ChatGPT’s goblin-mode is a fun little quirk for your Friday, and another example of how weird LLMs can truly be. A reward in a single personality mode led to a pattern of creature preferences that trickled across chats around the globe. Just like Anthropic’s Golden Gate Claude, we might need a standalone GoblinGPT.

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

#DJAMGAMIND #AIUNRAVELED

Summary: In today’s briefing, we analyze the breaking point of AI scaling and liability. We deconstruct the unprecedented $725 billion capital expenditure guidance from Google, Amazon, Microsoft, and Meta. We explore the mounting national security friction as the White House blocks the expansion of Anthropic’s Mythos model over cyber capabilities, while GPT-5.5 rolls out advanced cybersecurity features. We also cover the grim legal milestone of the Florida Attorney General launching a criminal investigation into OpenAI following ChatGPT’s alleged advice to a mass shooter. Finally, we look at enterprise agent rollouts from Amazon Quick and Perplexity, and SoftBank’s massive new robotics venture, “Roze.”

Important Topics:

• $725B Big Tech Capex: Google, Amazon, Microsoft, and Meta forecast a 77% YoY jump in infrastructure spending, hitting $725 billion for 2026.

• White House Blocks Anthropic: The administration intervenes against the expansion of Anthropic’s Mythos model to 70 companies, citing critical software vulnerability risks.

• OpenAI Criminal Investigation: Florida’s Attorney General opens a criminal probe into OpenAI after chat logs reveal ChatGPT allegedly advised a mass shooter on weapons and tactics.

• SoftBank’s Data Center Robotics: Masayoshi Son plans a $100 billion IPO for “Roze,” a standalone robotics company built to automate data center construction.

• Amazon Quick & Perplexity Enterprise: A major push for desktop agents, with Amazon natively integrating AI into Microsoft 365 and Zoom, while Perplexity expands its PC agent for financial workflows.

• Apple Abandons Vision Pro: Apple halts development on the Vision Pro line due to weak sales, pivoting engineering teams to AR glasses and Apple Intelligence.

• Spotify’s Anti-AI Badges: Spotify launches a “Verified by Spotify” green checkmark to distinguish genuine human artists from the flood of AI-generated personas.

• Mayo Clinic AI Cancer Detection: New REDMOD AI successfully spots invisible tissue patterns on CT scans to detect pancreatic cancer up to three years early.

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

• 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.

Big Tech capex hits $725 billion in 2026

• Google, Amazon, Microsoft, and Meta plan to spend a combined $725 billion on capex in 2026, a 77% jump from last year’s record $410 billion, driven by AI infrastructure demand and climbing memory chip prices.

• Microsoft set 2026 capex at $190 billion, with CFO Amy Hood blaming $25 billion of that on memory chip and component costs, while warning the company will stay capacity-constrained through at least 2026.

• Alphabet matched Microsoft’s $190 billion capex guidance after Google Cloud revenue grew 63% to $20 billion and its contract backlog doubled to $460 billion, pushing shares up 7% toward a $4.3 trillion valuation.

White House blocks Anthropic Mythos expansion

• The White House has pushed back on Anthropic’s plan to widen access to its Mythos AI model to around 70 companies and organizations, according to an administration official who spoke anonymously on Wednesday night.

• US officials worry Anthropic lacks the computing power to serve more Mythos users without hurting the government’s own use of the model, which the company says is strong enough to enable dangerous cyberattacks.

• Mythos, unveiled in early April, can reportedly detect and exploit vulnerabilities in critical software, and a small group of unauthorized users on a private online forum gained access the same day Anthropic announced its limited release plan.

Apple reportedly abandons Vision Pro

• Apple has reportedly stopped work on the Vision Pro after weak sales of the M5 chip model released in October, which kept the $3,499 price tag and added a more comfortable head strap, according to MacRumors.

• The product engineering team is being moved to other projects across the company, with a near-term focus on AR glasses to compete with Meta and longer-term work on a cheaper Pro-style successor.

• Apple is also shifting engineering resources toward Siri and Apple Intelligence ahead of WWDC in June, as recent delays to its AI work have hurt the company’s reputation with users and developers.

SoftBank is creating a robotics company that builds data centers

• SoftBank is planning to launch and list a standalone AI and robotics company in the U.S. called “Roze,” which will build data centers and use robotics to make AI infrastructure construction more efficient, the Financial Times reported Thursday.

• Masayoshi Son is leading the push, with executives targeting a roughly $100 billion valuation and an IPO as soon as this year, though the timeline could shift partly due to uncertainties from the conflict in the Middle East.

• Roze could bundle existing energy, land and infrastructure assets from SoftBank’s portfolio along with ABB Robotics, which SoftBank agreed to buy last year, and the listing may help offset its $30 billion-plus commitment to OpenAI.

Uber enters the hotel booking business

• Uber has launched hotel bookings inside its app for US customers, giving access to over 700,000 hotels worldwide through a partnership with Expedia Group, with Vrbo vacation rentals set to join later this year.

• Uber One subscribers will get 20% off a rolling list of 10,000 hotels booked through the app, plus 10% back in Uber Credits on all bookings, as the company pitches the subscription harder.

• CTO Praveen Neppalli Naga said agentic AI tools like Cursor cut the hotel booking feature’s build time in half, alongside other launches including travel mode, a room service hub in Uber Eats, and Eats for the Way.

Spotify introduces verified artist badges to help distinguish humans from AI

• Spotify is rolling out a new “Verified by Spotify” badge that marks human artists in good standing, shutting out AI-generated or AI-persona profiles as streaming sites deal with a flood of machine-made tracks clogging their platforms.

• To qualify for the light green checkmark, artists must show consistent listener engagement, follow platform policies, and display “signals of a real artist,” with over 99 percent of actively sought-out artists verified at launch.

• Spotify is also testing a new context section on artist profiles, described as “nutrition facts,” showing career milestones, release activity, and touring activity in the About section on mobile over the coming weeks.

Amazon makes AI agents easier to use and trust

If you’ve been jealous hearing about AI agents automating email, Slack, calendar, and files, but thought there wouldn’t be a version secure enough to work for you or your company, then Amazon has something you’ll want to see.

On Tuesday, the company released a desktop app for its agentic work assistant, Amazon Quick, which integrates natively with four of the most widely used business tools in the world: Microsoft 365, Google Workspace, Salesforce, and Zoom. Most importantly, it’s built with enterprise-grade security and privacy and is already trusted by companies like Southwest Airlines, BMW, and the NFL.

While Amazon Quick is a powerful personal agent, it’s been flying under the radar. I first got fomo when I heard about it in February, while interviewing Matt Yanchyshyn from AWS on The Deep View Conversations podcast. Matt talked casually about how Amazon Quick flags emails and Slack messages for him, helps him prep documents and summarize files, and drafts replies.

With the new version announced on Tuesday, Amazon Quick expands its capabilities. Here are some of the highlights:

• Deep memory: It remembers your context between sessions, connects to your most used systems, and is continually learning more about you, your patterns, and preferences.

• Proactive intelligence: The agent runs in the background on your computer and can, for example, spontaneously remind you of important, timely messages that are unanswered, tell you which docs need your feedback, remind you of approvals that need your attention, and flag deals that need updates in Salesforce.

• Takes actions: The agent can also take the next steps and do things for you. It can edit docs, draft emails and Slack messages, implement feedback from comments, update a Jira ticket, and reply to requests. You can decide how automated you want it to be, or choose to click approve before any actions are finalized.

• Creates files and pages: Amazon Quick can also create presentations, spreadsheets, images, documents, web pages, and dashboards. Examples Amazon highlighted include, “HR can create an onboarding portal for new hires with onboarding links and checklists. Finance can launch a resource/budget calculator. Sales can track pipeline health with Salesforce data and trigger actions like updating a deal status or sending a follow-up email.”

You can now download Amazon Quick and try the free version. You don’t need an AWS account, but you’ll need to use a Google, Apple, Amazon, or GitHub login.

In addition to Amazon Quick, which is aimed at helping professionals create their own AI agents, the company also expanded Amazon Connect, which now offers pre-built, enterprise-ready agents for hiring, supply chain management, health care, and customer service.

Perplexity shifts its desktop agent to enterprise

Perplexity’s Personal Computer is winning users over by working directly across local files and native apps, and not just the cloud. And the company is now rolling out a suite of enterprise upgrades aimed at the professionals who are driving the most demand for the product.

On Wednesday, Perplexity held its Ask NYC event, which opened with CEO Aravind Srinivas talking about Personal Computer’s impact. Citing Steve Jobs’s famous description of the computer as “a bicycle for the mind,” Srinivas reframed the new personal computer as something far more powerful: a Ferrari, or three.

Then Dmitry Shevelenko, Perplexity CBO, and Jeff Grimes, Head of Live Events, took the stage to launch the new Computer features, including a partnership with 1Password that allows Personal Computer to take action within password-protected tools while keeping credentials private from AI models.

Another large focus was broadening access to Computer. Previously available only to Max-tier subscribers, it is now available to Pro subscribers on Mac. It’s also available on Microsoft Teams through Microsoft Marketplace, with Computer in Excel launching in beta as a native side panel.

Computer will now also offer:

• Workflows: Over 50 prebuilt templates across enterprise work that can be shared, scheduled, customized, and run asynchronously

• Connectors: New Databrick and Snowflake connectors join the hundreds available to make querying enterprise data available org-wide

Further catering to the financial sector, a new Computer for Financial Services offering allows users to bring their own license connectors for Carbon Arc, Daloopa, Morningstar, and PitchBook, giving Computer access to existing licensed data credentials and enabling the building of workflows and dashboards, according to Perplexity.

It also launched new finance-focused workflows, including an Equity Research Council, designed to recreate the experience of a panel-style research review that typically involves multiple analysts and sources.

Thus far, Perplexity says its agent has saved companies billions of dollars. In its first four weeks, while it still existed only internally on Slack, Computer performed $1.6 million worth of work, according to Shevelenko. The company reports that since its launch, Perplexity Computer has performed more than $2.8B in labor-equivalent work.

Nvidia brings agents closer to real work

The more human tasks AI agents take on, the more they need to see and hear the world as humans do. Nvidia’s new model is designed to bridge that gap.

On Tuesday, Nvidia launched Nemoton 3 Nano Omni, an open multimodal model that combines vision, speech, and language capabilities into a single model, aiming to enable agents to skip handoffs between separate models and deliver faster, smarter results.

The model doesn’t compromise on efficiency either, topping six leaderboards across complex document intelligence, video, and audio understanding, while enabling AI systems to achieve up to 9x higher throughput than other open omni models, according to Nvidia. This performance is enabled by:

• Architecture: 30B-A3B hybrid mixture-of-experts architecture, the same as Nemotron 3 Nano

• Audio and vision: The addition of audio and vision encoders makes it possible to combine capabilities into one model and eliminates the need for separate perception models

• Model partner: Nemoton 3 Nano Omni can work alongside proprietary cloud models or other Nvidia Nemotron open models to power agentic workflows

Many AI-driven companies are already adopting the model, according to the post. Potential use cases range from computer-use agents that navigate graphical user interfaces and reason over onscreen content, to interpreting documents, charts, tables, screenshots, and audio for customer service and workflow monitoring.

The model is available now through Hugging Face, OpenRouter and build.nvidia.com as an NVIDIA NIM microservice, and through Nvidia Cloud Partners, inference platforms and cloud service providers, according to the company.

Zuckerberg’s Biohub funnels $500M into AI biology

The Rundown: Biohub, the nonprofit backed by Mark Zuckerberg and Priscilla Chan’s CZI, announced a $500M Virtual Biology Initiative to build open datasets and models that can predict how human cells behave — pushing AI toward biology simulation.

The details:

• $400M of the $500M will fund data generation and imaging tech, with $100M for external research labs and research efforts.

• Nvidia, Allen Institute, Arc, and others are joining the initiative, with Biohub committing to open datasets as a shared base for AI biology research.

• Current AI biology datasets max out near 1B cells, with Biohub’s Alex Rives saying an “order of magnitude” more data is needed to accelerate the efforts.

• The goal is to train models on the data to use AI toward “understanding disease and reprogramming it at the level of cells, molecules, and tissues.”

Why it matters: Google’s Demis Hassabis has said AI could eventually end disease, and Biohub is pouring serious money behind that same line of thinking. The question is whether the scaling that cracked language and protein structure also holds for cells, and whether $500M gets anywhere close to the data scale needed to find out.

Mayo Clinic AI spots pancreatic cancer 3 years early

Mayo Clinic published new data on REDMOD, an AI that reads invisible tissue patterns on standard CT scans, catching pancreatic cancer up to three years ahead of when doctors typically find it and nearly doubling specialist accuracy.

The details:

• REDMOD reviewed nearly 2,000 routine CT scans that specialists had originally read as normal before later diagnoses, picking up 73% of the cases early.

• At the two-year mark before diagnosis, the gap widened even more, with the AI spotting roughly 3x as many early cancers as experienced radiologists did.

• The model reads “hundreds of quantitative imaging features,” texture, and structure patterns normally invisible to human radiologists.

Why it matters: Pancreatic cancer’s 5-year survival rate for pancreatic cancer is below 15%, making early diagnosis and treatment critical. With REDMOD running on scans patients already get, AI’s early screening abilities could become a standard part of routine care rather than a separate diagnostic step that adds friction to the system.

Food AI’s ‘ChatGPT moment’ — tastes like a chef

Image source: Midjourney

Food robotics startup KAIKAKU AI just published Epicure, a new paper claiming a “ChatGPT moment” for food AI that shows its AI model can pick up on flavor, cuisine, and texture just from how chefs combine ingredients in recipes.

The details:

• Researchers cleaned 6,653 messy ingredient entries into 1,032 usable foods, then mapped with AI how they relate across recipes.

• Despite never seeing chemistry data or taste labels, the model identified all 5 basic tastes, ordered peppers by spiciness, and tagged cuisines by region.

• The team flagged three applications: menu development, recipe innovation, and flavor pairing, work that is normally driven by chef intuition.

• KAIKAKU is pairing the AI from this paper with their robotics arm, pitching the combo as “autonomous food infrastructure” for commercial kitchens.

Why it matters: Recipes are already strong data points for human preference, with each pairing, swap, and pattern a signal about what people think works. If AI can read that structure, it can also help implement tools to design menus, suggest better substitutions, and create products with actual learned taste and texture in mind.

GPT-5.5 achieves superior CyberSecurity performance to Mythos

AISecurityInst is the org that Anthropic released Mythos to verify their “too dangerous to release claims”.

I’ve used GPT-5.5 to find vulns. It is pretty good, it’s true, but hardly “too dangerous to release”. That said, people should use it to review their code.

You will have to get Persona verified for security stuff, however.

OpenAI Faces Criminal Investigation in Florida: Can ChatGPT Be Charged With Murder? [LINK]

Florida Attorney General James Uthmeier announced that his office has opened a criminal investigation into OpenAI on the April 2025 mass shooting at Florida State University. Reviews of chat logs indicate that ChatGPT allegedly advised the accused shooter, Phoenix Ikner, on weapon type, ammunition, optimal timing, and campus locations likely to have the most people. Uthmeier later expanded the probe to cover a separate double homicide at the University of South Florida, where the suspect in that case also allegedly consulted ChatGPT before the killings.

https://www.nolo.com/news/openai-faces-criminal-investigation-in-florida-can-chatgpt-be-charged-with-murder.html

What Else Happened in AI on April 30th 2026?

ElevenLabs launched ElevenMusic, a streaming platform with built-in AI remixing and AI-assisted track creation, already hosting 4k+ artists and offering creator payouts.

Two U.S. House committees opened probes into Cursor-maker Anysphere and Airbnb over Chinese AI use, with Composer 2 built on Kimi and Airbnb’s agent on Qwen.

Mistral AI launched Vibe remote agents, cloud sessions that run coding tasks in parallel, powered by the company’s new open-weights Medium 3.5 model.

Google added file creation into Gemini, allowing the model to output formats like Google Docs, Sheets, and Slides, Microsoft Word and Excel files, Markdown, and more.

OpenAI released a Cybersecurity Action Plan to “democratize” AI cyber defense and work with the U.S. government and industry on threat coordination and defender tools.

Key Context by Tae Kim: The State of AI After Google, Meta, Amazon, Microsoft Earnings

OpenAI: Where the goblins came from

The Information: Anthropic Receives Investment Interest at $900 Billion Valuation

WSJ: Big Tech Strikes Gold With AI, but at a Steep Cost

FT: Elon Musk says he was ‘a fool’ to fund the launch of OpenAI

OpenAI to roll out GPT-5.5-Cyber, a frontier cybersecurity model

TIME: How Sundar Pichai Pushed Google To the Front of the AI Race

Bloomberg: NSA Testing Anthropic’s Mythos to Find Flaws in Microsoft Tech

WSJ: Retailers Flock to TikTok Shop to Find New Shoppers, Sales Growth

The Information: Google Defends Military Work After Employee Backlash to Pentagon Contract

NYT: Uber Can Already Bring You Dinner. Now, It Wants to Book Your Hotel Room.

Eric Seufert: “Meta’s advertising revenue grew 33% year-over-year in Q1 2026, to $55BN”

WSJ: SoftBank Plots IPO for New Robotics Venture

WSJ: Sports Streamer DAZN Makes $100 Million Bet on Technology Company ViewLift

Bloomberg: AI Hallucinations Put South African Ministers on the Spot

🎧 Listen Ads-Free: Subscribe to DjamgaMind via Apple Podcasts for a pure, ad-free experience at https://podcasts.apple.com/us/channel/djamgamind/id6760446113

Summary: In today’s briefing, we analyze the collision of massive AI capital expenditures and deep ideological rifts. We break down the “QuadKill” of tech earnings, exploring how Google, Microsoft, Amazon, and Meta plan to justify hundreds of billions in Capex. We cover the courtroom drama as Elon Musk testifies that OpenAI “stole a charity” on its path to a $1 Trillion valuation. We also deconstruct OpenAI’s strategic launch on Amazon Bedrock, Google’s controversial classified Pentagon deal that sparked a 600-employee mutiny, Anthropic’s new creative integrations, and regulatory crackdowns on Meta in the EU and Baidu in China.

Today’s Sponsors: *

• 🛑AIRIA: The ultimate zero-trust AI security layer. Deploy autonomous agents safely without compromising your enterprise data. 👉 Govern your agents: https://airia.com/request-demo/?utm_source=AI+Unraveled+&utm_medium=Podcast&utm_campaign=Q1+2026

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

Important Topics:

• Elon Musk vs. OpenAI Trial: Musk testifies in federal court that OpenAI abandoned its founding mission, likening their for-profit pivot to “stealing a charity.”

• Big Tech Earnings (Capex QuadKill): Google, Microsoft, Amazon, and Meta report strong earnings but signal staggering AI capital expenditures (up to $200B targets) to secure compute dominance.

• OpenAI Launches on AWS: Following its Microsoft contract restructure, OpenAI brings GPT-5.5, Codex, and Managed Agents to Amazon Bedrock.

• Google’s Classified Pentagon Deal: Over 600 Google employees sign an open letter protesting a new military AI contract authorized for “any lawful government purpose.”

• EU Targets Meta’s Age Checks: The European Commission rules that Meta fails to diligently block children under 13, threatening fines up to 6% of global turnover.

• China Halts Robotaxi Permits: Baidu, Pony AI, and WeRide face a regulatory freeze after 100 autonomous vehicles stalled and snarled traffic in Wuhan.

• Anthropic Claude for Creative Work: New direct software integrations for Adobe, Autodesk, and Blender to bring AI straight into creative workflows.

• Talkie (The 1930s AI): Researchers demo a 13B parameter AI trained exclusively on pre-1931 public domain text to bypass modern algorithmic bias.

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

• 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.

Elon Musk says OpenAI betrayed its mission

• Elon Musk told jurors in an Oakland federal court that OpenAI abandoned its founding mission when it shifted from a charity to a for-profit company, arguing the pivot amounts to “stealing a charity” and sets a dangerous precedent.

• OpenAI’s lawyer William Savitt countered that Musk himself pushed to restructure OpenAI as a for-profit in 2017 and wanted majority control, saying the lawsuit is really an attempt to hobble a rival to Musk’s own AI company, xAI.

• The three-week trial could reshape OpenAI as it approaches a trillion-dollar valuation and a planned public offering, with Musk seeking a court order to unwind the October for-profit conversion that gave Microsoft a 27% stake and the nonprofit 26%.

OpenAI launches models on AWS

• OpenAI’s models and its Codex coding agent are coming to Amazon Web Services through Amazon Bedrock, the two companies said Tuesday, with general availability expected in the next few weeks for AWS customers to try.

• A new service called Amazon Bedrock Managed Agents powered by OpenAI will let developers build customized agents that remember previous interactions, going beyond the open-weight OpenAI models that came to AWS back in August.

• The news follows Monday’s reworked Microsoft deal letting OpenAI serve customers on any cloud, and builds on a $38 billion AWS commitment from November plus a $50 billion Amazon investment tied to two gigawatts of Trainium chips.

EU says Meta fails underage user checks

• The European Commission has preliminarily ruled that Instagram and Facebook breach the Digital Services Act by failing to diligently identify and block children under 13 from using the platforms, despite Meta’s own age restrictions.

• Regulators said minors can enter false birth dates with no effective checks, Meta’s tools for reporting underage users are hard to use, and the company does not follow up on reports, letting children keep their accounts.

• If confirmed, the findings could lead to a fine of up to 6pc of Meta’s worldwide annual turnover, and the Commission wants Meta to overhaul its risk assessment and bring in age-assurance technologies that are accurate and non-intrusive.

Anthropic unveils Claude for Creative Work

• Anthropic has rolled out Claude for Creative Work, a set of integrations that plug its AI directly into creative software from Adobe, Autodesk, Ableton, Blender, and Splice rather than asking users to adopt a separate tool.

• Each connector targets a specific task: conversational 3D modelling in Autodesk Fusion, natural-language scripting in Blender, real-time visual control in Resolume, prompt-to-3D concepts in SketchUp, and in-app royalty-free sample search through Splice.

• Anthropic has also joined the Blender Development Fund as a patron and is partnering with the Rhode Island School of Design, Ringling College of Art and Design, and Goldsmiths to give students and faculty access to Claude.

China halts new self-driving permits after Baidu outage

• China has stopped issuing new licenses for autonomous vehicles after more than 100 Baidu Apollo Go robotaxis suddenly stalled on the streets of Wuhan on March 31, stranding passengers and snarling city traffic.

• Three agencies including the Ministry of Industry and Information Technology met with officials from robotaxi pilot cities, calling for a full self-review and better safety monitoring, with no clear end date for the suspension.

• The freeze blocks companies from adding robotaxis, starting test projects, or expanding to new cities, and it covers level four vehicles, sending Baidu, Pony AI, and WeRide shares lower in Wednesday trading.

White House reforging ties with Anthropic:

Axios reports that — just over a month after labeling the AI giant a “supply chain risk” — the Trump administration is prepping to backtrack. Obviously, the debate between the Pentagon and Anthropic over the use of AI for domestic surveillance and autonomous weapons systems went down before the company announced Mythos. That’s their latest frontier model, rumored to reshape the global playing field in terms of cybersecurity. Now that the US government is eager to start playing around with Mythos, they’re changing their tune on the whole “supply chain” concern. Axios suggests that the White House hopes to assemble various companies across different sectors to discuss Best Practices for deploying Mythos, and essentially wants to revive the Anthropic collab while publicly saving face. No word on what this means for the domestic surveillance and/or kill-bot policies.

Musk v. Altman kicks off:

On Tuesday, Elon Musk testified in his lawsuit against OpenAI and co-founders Sam Altman and Greg Brockman, telling jurors that he’s filed the suit because “it is not okay to steal a charity.” He further insisted that OpenAI’s restructuring must not serve as an ongoing precedent, lest we see the “looting [of] every charity in America,” and made the case that Brockman and Altman took advantage of his money and reputation, before betraying their core principles in a switch to for-profit status. OpenAI attorney William Savitt countered that Musk is attempting to use the court system to undermine a key competitor to his own company, xAI.

Codex goes Goblin Mode:

In less dramatic but certainly more bizarre news, Wired reports that OpenAI’s Codex command-line tool includes specific instructions forbidding AI models to mention an assortment of creatures, both real and fictional. The instructions read: “Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user’s query.” Why was this important for OpenAI to spell out specifically for its models? No one knows, and the company is refusing to explain. CEO Sam Altman has acknowledged the story, however, on social media. He posted on Wednesday that Codex is having a “ChatGPT moment,” before correcting himself: “i meant a goblin moment, sorry.”

Google finalizes Pentagon deal despite protests

Google signed a classified AI deal with the Pentagon, opening its models to “any lawful government purpose,” the same week that 600+ staffers wrote an open letter to CEO Sundar Pichai, calling to reject the use of AI for military purposes.

The details:

• More than 600 Google employees sent Pichai a letter on Monday asking him to “refuse to make our AI systems available for classified workloads.”

• The Information reported that the contract opens Google’s AI to “any lawful government purpose”, with no legal right to veto how the Pentagon uses it.

• OAI and xAI inked deals with the Pentagon last month, with Anthropic currently fighting in court after being blacklisted for not dropping its guardrails.

• Google’s no-weapons pledge was scrubbed from its AI principles in 2025, after it was implemented in 2018 following successful staff protests.

Why it matters: The Pentagon drama might still feel fresh in the OAI-vs-Anthropic rivalry, but it’s not discouraging another top AI lab from making a similar deal. Google’s now wading into a messy territory from a PR and internal perspective, and time will tell if the same backlash we saw with ChatGPT now comes to Gemini’s doorstep.

Talkie is an AI that thinks it’s 1930

Researchers Nick Levine, David Duvenaud (fmr. Anthropic), and Alec Radford (fmr. OpenAI) demoed Talkie, a 13B ‘vintage’ AI model trained only on text from before 1931, built to test how AI thinks when its worldview predates the internet.

The details:

• Talkie was trained on 260B tokens of pre-1931 books, newspapers, journals, patents, and case law, all now in the US public domain.

• To teach talkie to chat without modern data, the team pulled instructions from etiquette manuals and cookbooks, with Claude Sonnet 4.6 grading the answers.

• The coding language Python didn’t exist in 1930, but Talkie wrote working code by flipping a plus sign to a minus sign in an example, proving it can generalize.

• AI benchmarks get poisoned when models train on their own test data — talkie sidesteps that, with a GPT-3-level version coming next.

Why it matters: Today’s frontier models all sound vaguely similar because they all read roughly the same modern web. Talkie is definitely cut from a different cloth — but the Python coding anecdote is a fascinating part of the experiment that shows what kind of learning and reasoning is potentially going on beneath the original training data.

Tech Earnings QuadKill From TBPN

by John Coogan

Four earnings calls, one big question. How is the AI buildout going? SemiAnalysis expects upward revisions to capex guides. Financial performance has been strong across the board, even in “legacy” areas like search, ecommerce sales, and enterprise software seats. The bigger question is around durable revenue tied to AI infrastructure. Everyone has seen lots of cash flow to NVIDIA, power companies, and data center builders, but what does the exact conversion to higher revenues and profits look like?

Let’s start with Google, who had a great last quarter. Revenue up 18% for Q4. Search up 17%, YouTube ads up 9% and Cloud up 48% with operating income of $5.3B. Incredibly strong platform to build off of, and all of that is accelerating AI across the organization. The Gemini App continues to grow and direct customer API usage “exceeded 10 billion tokens per minute” – this all led to basically doubling capex year over year, with a 2026 target of $175B to $185B.

Google has a fully integrated AI stack. Consumer distribution, model training, custom chips, and product surfaces like search, YouTube, Android, and Cloud to deploy solutions across. The question investors are asking is: “Does AI change the unit economics of Search too quickly?” There are tons of places to pick up growth, especially in cloud, but the question is whether AI Overviews and Gemini are expanding search usage and ad ROI or compressing the model.

Microsoft is also coming off a strong quarter. Revenue up 17%, with cloud (which includes Azure, M365, etc.) growing 26%. Azure alone is growing even faster at 39% and is a bigger lever on capex (which is run rating around $150B). The biggest number in the Microsoft earnings is remaining performance obligations, listed at $625B last earnings, up 110%, with about 45% of that coming from OpenAI.

Microsoft gives the cleanest read on enterprise AI monetization. Azure growth, gross margins for Microsoft Cloud, Copilot adoption and ARPU, M365 seat growth, and GitHub Copilot momentum will all paint a picture of what’s happening with AI adoption in enterprises broadly. Copilots have not been the most hyped product, but strong growth numbers could reveal how nuanced the diffusion/adoption question really is. M365 seat growth should have impacts for the seat-based enterprise SaaS model.

For Amazon, everyone wants to see strong AWS acceleration to justify their Mag7-topping capex numbers. They guided to $200B in 2026 capex, so expect a lot of focus on AWS revenue growth and margin. Q4 was healthy overall though. Net sales up 14%, AWS growing 24%, the sneakily huge ads business grew 23% to $21.3B. Operating income overall was $25B for the quarter and they generated free cash flow of $11.2B, but this was down on increased AI spending.

If AWS accelerates, all the capex looks like buying scarce capacity ahead of demand. Consensus around AWS revenue is ~$36.7B, with growth in the mid-20s range, but the market is really hoping to see that AWS growth rate start with a 3 soon if the capex keeps growing.

Lastly, Meta. Super clean Q4, nothing really to prove. Despite all the FUD around the talent wars, the new team members, and sometimes clunky model releases, AI is already helping the core product, and that’s what matters for the financials. Revenue grew 24% to $59.9B for the quarter. They have 3.58 billion DAUs and grew the ad business a ton. Ad impressions rose 18% and average price per ad rose 6%. Family of Apps operating income was $30.8B, which really makes the $6B loss at Reality Labs look quaint. Capex was $72.2B for 2025 and the guide is $115B to $135B for 2026.

The question is how much more juice will AI bring to the ad business. Expectations imply revenue growth of around 31%, and the question is always the same – is there enough incremental growth to justify all the capex? The good news for Meta is that improvements in AI move the needle incredibly quickly. An AI advancement delivers better ad performance basically immediately, they don’t have to go and convince anyone to adopt a new product or change a workflow. There’s no “diffusion” question.

So, in summary, every company is trying to answer the same question: “Can you turn AI capex into proprietary distribution, higher customer retention, and measurable revenue growth before depreciation catches up?” Super Bowl for big tech. Get your popcorn ready.

What Else Happened in AI on April 29th 2026?

OpenAI announced that GPT-5.5, Codex, and Managed Agents are now available via Amazon Bedrock, coming a day after its new contract restructure with Microsoft.

NVIDIA released Nemotron 3 Nano Omni, a new open model that can handle vision, audio, and text at 9x the speed of rival open multimodal models.

The WSJ reported that OAI fell short of its targets for revenue and user growth, with CFO Sarah Friar questioning its massive spending — with OAI calling it “ludicrous.”

Anthropic added new connectors for a broader range of creative workflows, including apps like Blender, Adobe Creative Cloud, Autodesk Fusion, SketchUp, and more.

Xiaomi open-sourced MiMo-V2.5-Pro, which ties Kimi K2.6 on Artificial Analysis’ leaderboard, featuring a 1M context window and strong efficiency for agentic tasks.

SpAItial launched Echo-2, a new SOTA world model that turns text or photos into explorable 3D worlds, claiming to beat World Labs’ Marble 1.1 across benchmarks.

Bloomberg: IBM to Add 750 Jobs in AI, Quantum Computing at Chicago Tech Hub

Elon Musk set for major SpaceX payday — if he settles 1 million people on Mars

FT: How OpenAI’s $500bn data centre venture Stargate has shifted shape

WSJ: Ex-Twitter CEO’s AI Startup Raises Funds at $2 Billion Valuation

WSJ: OpenAI Sued by Seven Families Over Mass Shooting Suspect’s ChatGPT Use

NBC: Inside Day 2 of the OpenAI trial: All eyes on Elon Musk

NY Post: New Meta AI glasses help veteran ‘see’ the room around him

AP: Robot dogs with Musk and Zuckerberg heads roam around Berlin museum in Beeple’s new exhibit

Joe Weisenthal: “Huge. A Brookfield-backed datacenter company is pulling out of a major project in Virginia, that they had been working on for years, due to growing political opposition”

Intel Stock Soars As AI Turnaround Ignites Massive Rally