Generative AI Newsletter
Anthropic Partners With SpaceX and Now Runs on Mus...
Plus: Chrome installed a 4GB AI model on your device without permission, the Pentagon's Scale AI deal jumped 5x, Snap-Perplexity falls apart.
Meta's AI scans your height and bones
Meta analyzes user pictures with AI, Apple is fined $250M for lying about how smart their phones are and the EU found a way for AI to help with disasters.
Wall Street Just Backed Anthropic Over OpenAI
Plus, a Trump executive order on AI review, the Palantir cofounder funding ocean AI compute and Shopify inside ChatGPT and Claude.
Harvard put AI against ER doctors and AI won
Plus, your ChatGPT logs are evidence now, the Treasury Secretary warns on bank hacks and TikTok defaults posts into AI memes.
Musk admitted he trained Grok on ChatGPT
Plus, Claude Opus 4.7 ships as a security product, OpenAI opens its strongest models to defenders, Meta's $145B capex kills a record rally.
The White House refuses to let Mythos out
Plus, Stripe gives AI agents access to wallets, Wall Street loses its nerve after the latest OpenAI news, Amazon's desktop AI application is out.
Inside OpenAI's worst pre-IPO week
Plus, Claude plugs into Adobe and Blender, the Pentagon gets Google AI with no limits and a Claude agent wiped a production database in 9 seconds.
Musk dragged Altman into court over $134B
Plus, OpenAI ends Microsoft's cloud lock, China blocks Meta's $2B Manus and AlphaGo's creator bets $1.1B on AI that skips training on human data.
Google just let Claude and GPT run inside Workspac...
Plus, Claude just went shopping for 69 humans, the State Department's accusation against China, and the Gulf state running its government through agents.
OpenAI just retook the AI crown
Plus: Anthropic's Mythos leaks to a Discord group, ChatGPT gets team-wide agents, Google's new TPUs split training from inference
Meta is ranking engineers on AI token burn
Plus: Anthropic's Mythos leaks to a Discord group, ChatGPT gets team-wide agents, Google's new TPUs split training from inference
The image model that thinks for itself
Plus: Google puts Gemini 3.1 Pro on analyst reports, Meta logs employee keystrokes, SpaceX lines up Cursor
The best coding model on earth is free
Plus: A Google co-founder un-retires, Anthropic locks in 5GW of compute, Apple's new CEO inherits an AI problem
Claude is a designer now
Plus: three OpenAI leaders out the door, Dario's 12-month forecast, Salesforce hands its whole platform to agents
This beanie can read your mind
Plus: Opus 4.7 arrives, Codex turns into a superapp and OpenAI heads to the lab.



























English (US) ·