- A federal hearing is underway today before Judge Rita Lin in California where Anthropic is contesting the Pentagon's designation of the company as a "supply-chain national security risk" — the first time such a label has been applied to a US-based AI firm.
- Anthropic's head of policy and head of public sector filed sworn declarations denying that Anthropic sought veto authority over military operations.
Snapshot — March 24, 2026
38 stories
Agentic AI Meta / creati.ai / CXO Digital Pulse
AI Safety Protesters March on OpenAI, Anthropic, and xAI Offices
- Amazon Web Services and Cerebras Systems announced a collaboration to deliver the fastest AI inference available in the cloud, launching through Amazon Bedrock in the coming months.
- The solution uses "inference disaggregation" — splitting prefill on AWS Trainium and decode on Cerebras CS-3 — connected by Amazon's Elastic Fabric Adapter.
- AMD and Meta have officially expanded their multi-year AI infrastructure partnership around the deployment of up to 6 gigawatts of AMD Instinct GPUs based on a custom MI450 architecture.
- The deal, structured around AMD's Helios rack-scale design and 6th-Gen EPYC CPUs, marks one of the largest chip agreements in semiconductor history.
- Anthropic's Computer Use feature — in research preview for Claude Pro and Max on macOS — allows Claude to autonomously control a user's desktop: clicking, typing, opening apps, and completing tasks remotely.
- The "Dispatch" companion lets users send instructions from their iPhone to be executed on their Mac.
- Anthropic has agreed to terms on a $10 billion funding round co-led by Coatue and Singapore's sovereign wealth fund GIC, pushing the company's valuation to roughly $350 billion — nearly double its $183 billion valuation just six months ago.
- The raise makes Anthropic one of the twenty most valuable private enterprises globally.
Apple Announces WWDC 2026 with Explicit Promise of "AI Advancements"
Apple confirmed WWDC 2026 will run June 8–12, and its announcement press release explicitly teased "AI advancements" — a departure from Apple's typically vague developer conference previews. Analysts interpret this as a commitment to finally deliver the long-delayed Siri overhaul, including deeper contextual awareness, on-screen intelligence, and multi-step task execution, potentially powered by Apple's Google partnership. iOS 27 is expected to be the primary vehicle for these updates.
- As OpenAI — valued at $730 billion after its record $110B fundraise — eyes a public market debut, CEO Sam Altman has pivoted toward more measured infrastructure spending.
- The company is scaling back plans to build its own massive data centers and instead leaning into purchased cloud capacity, including through its $500B Stargate partnership with Oracle and SoftBank.
AWS and Cerebras Partner for Industry-Leading AI Inference Speed on Bedrock
*Coverage from newsletter subscriptions for 2026-03-24*
Cursor Discloses Composer 2 Is Built on Moonshot AI's Kimi K2.5
- Cursor revealed that its recently launched Composer 2 coding model — marketed as "frontier-level coding performance" — was fine-tuned from Kimi K2.5, an open-source model by Chinese AI startup Moonshot AI (backed by Alibaba).
- The disclosure sparked debate about model provenance transparency in the developer tools space.
Developer Tools & Platforms AWS / About Amazon
- Google DeepMind released a research paper introducing a cognitive framework for systematically measuring progress toward Artificial General Intelligence.
- The paper defines capability milestones across reasoning, planning, memory, and generalization — offering a more rigorous vocabulary for a debate long hampered by definitional ambiguity.
Google DeepMind's AlphaProof — the reinforcement learning system that achieved silver-medal performance at the International Mathematical Olympiad by bridging natural language and symbolic reasoning — was formally published in Nature this week. The paper details the RL loop enabling AlphaProof to translate natural language math problems into formal Lean proofs, a milestone in AI's capacity for rigorous mathematical reasoning with implications for scientific discovery.
- Google Labs' March 18 update transformed Stitch from a simple UI mockup tool into a comprehensive AI design platform capable of generating full, production-ready interfaces from plain text or voice.
- Powered by Gemini 3.1 Pro, the update introduced an infinite AI canvas, a context-aware Design Agent, voice interaction via Gemini Live, editable Figma export, and an MCP server for direct integration with coding tools like Cursor and Claude Code.
Hardware & Infrastructure AIToolly / The Verge
- In a Monday episode of the Lex Fridman podcast, Nvidia CEO Jensen Huang stated "I think we've achieved AGI" — a significant and deliberately provocative claim given the lack of an industry-standard definition for artificial general intelligence.
- The statement adds weight to a growing CEO consensus that AI systems have crossed a meaningful threshold of generalized capability, though benchmarks remain contested.
Industry & Policy Economic Times / TechCrunch
Is this email difficult to read? View in browser - The Wall Street Journal The Wall Street Journal - Meta Names New Leader of Company’s Efforts To Become AI Native - Andrew Bosworth, Meta’s chief technology officer, will oversee the company’s “AI For Work” initiatives as focus shifts away from the metaverse. - Alerts Center - Cookie Policy
Mark Zuckerberg is personally testing a "CEO agent" that automates executive information retrieval and streamlines communications at Meta, according to The Wall Street Journal. Internally, Meta employees are also experimenting with "My Claw" (an agent for navigating internal chats) and "Second Brain" (a chief-of-staff agent for project documents), with AI adoption now factored into performance reviews.
- Meta has hired the entire founding team of Dreamer — an agentic AI startup valued at $500 million — to join its Superintelligence Labs division.
- The team includes David Singleton (ex-Stripe CTO), Hugo Barra (ex-Meta Oculus VP), and Nicholas Jitkoff (ex-Chrome OS design lead).
- Dreamer built technology for creating custom AI agents using natural language.
Model Releases & Products Anthropic / BusinessToday
Nearly 200 activists from Pause AI and QuitGPT marched through San Francisco yesterday, stopping at the offices of Anthropic, OpenAI, and xAI to demand that CEOs publicly commit to pausing development of frontier AI models. The protest marks an escalation in organized public pressure on leading AI labs at a moment when multiple companies are simultaneously announcing more powerful and autonomous AI capabilities.
Nvidia Open-Sources Nemotron-Cascade 2: Efficient 30B MoE for Agentic Workflows
Nvidia released Nemotron-Cascade 2, an open 30-billion-parameter Mixture-of-Experts model with only 3 billion active parameters at inference, making it highly cost-efficient for deployment. The model is specifically designed for agentic AI tasks and continues Nvidia's push to pair hardware dominance with open-source software contributions, positioning it as a key option for enterprises building on the NemoClaw agentic platform announced at GTC 2026.
OpenAI Courts Private Equity with 17.5% Guaranteed Returns for Enterprise AI Ventures
OpenAI is pitching private equity firms including TPG and Bain Capital on joint ventures offering a guaranteed minimum return of 17.5%, early access to frontier models, and a $4 billion capital commitment target. The structure reflects OpenAI's strategy of converting PE capital into enterprise deployment and distribution channels ahead of a potential IPO, with early-model access as a key differentiator over competitors in the enterprise AI market.
- OpenAI discontinued Sora after six months, citing ~$1M/day compute costs against a user base that collapsed from ~1M to under 500,000.
- The shutdown killed a $1B Disney content deal — Disney learned less than an hour before the public announcement.
- Sam Altman reportedly stated the company cannot afford "side quests." Sora's fate is the clearest data point yet that high-fidelity AI video generation remains economically unviable at consumer scale with current infrastructure costs.
Research & Science Google DeepMind
Slack, Workday and LinkedIn Are Resisting Customers’ AI Agents [2026-03-24] · The Information
Sources: BusinessToday · xix.ai · CNBC · creati.ai · The Verge · AIToolly · Cybernews · Wall Street Journal · CXO Digital Pulse · teknogadyet.com · Google DeepMind · Nature · MarkTechPost · Economic Times · 9to5Mac · MacRumors · AWS News Blog · About Amazon · The Neuron · AgentDock Academy
The Information logo - Laura Bratton headshot - By Laura Bratton - Sponsor Logo - or in the case of Amazon employees, trepidation - we first reported on with Slack nine months ago - Jyoti reported last week - A message from Google Cloud - From 8 weeks to 8 hours: How Kraft Heinz and Google Cloud are rewriting the recipe for innovation - Monday, April 27 — Financing the AI Revolution
The latest news on Meta Platforms Inc. [2026-03-24] · Wall Street Journal
This digest was compiled from publicly available news sources published in the last 24 hours. All stories verified against original reporting at time of compilation.
Zuckerberg Is Building an AI Agent to Help Run Meta