A federal judge blocked the Trump administration from labeling Anthropic a supply-chain risk. Anthropic accidentally leaked their most powerful model ever — called Mythos. Google shipped Gemini 3.1 Flash Live, a real-time audio model with sub-half-second responses. Granola raised $125M at a $1.5B valuation, up from $250M six months ago. And Anthropic research says AI isn't replacing jobs yet — but power users are pulling ahead fast.

OpenAI killed the Sora app because nobody wanted to scroll through AI-generated videos all day. Anthropic shipped auto mode for Claude Code. A Mozilla developer built a Stack Overflow designed for AI agents. Bernie Sanders tried to outsmart a chatbot and the internet has not stopped laughing. And Apple announced WWDC with a tease that Siri might finally learn some new tricks.

Anthropic is in federal court telling the Pentagon it cannot remotely sabotage AI models. Microsoft quietly pulled Copilot from Photos, Notepad, and Widgets. Palantir is winning defense contracts by doing what Anthropic refuses. The smartest AI investment might be power companies. And a 2024 essay predicting AI would replace knowledge workers by 2027 just aged into something uncomfortable.

Meta's rogue agent leaked data. The Pentagon dumps Anthropic. Carl Pei says apps are dying. Garry Tan's Claude Code setup went viral. And the PhD students who built the AI industry's scoreboard.

Sears exposed thousands of AI chatbot phone calls and text conversations to the open web with no authentication required. Jensen Huang used his GTC keynote to project Nvidia's Blackwell and Vera Rubin chip families into a $1 trillion sales trajectory. A developer built a "vibe coded" AI translati...

A WIRED investigation found women across Southeast Asia applying to be "AI face models" for industrialized pig-butchering scam operations. Google and Accel sifted through 4,000 Indian AI startup pitches and rejected 70% as wrappers. Musk says Tesla's Terafab chip factory launches March 21. Uber f...

Lighter Sunday edition today with three stories worth your time. Google's Gemini can now take over your phone and order food for you. Steven Spielberg told SXSW he has never used AI in any of his films. And Peacock is turning its streaming app into a TikTok-style AI playground with vertical NBA b...

Grammarly is facing a class action lawsuit for turning real authors into fake AI editors. Meta just gave its AI the keys to your Marketplace inbox. Bumble introduced an AI dating assistant called "Bee" that learns what you want and picks your matches. A father-son startup called Nyne raised $5.3 ...

Palantir demos show Claude inside Pentagon warfare systems generating targeting suggestions. A Pentagon official confirmed generative AI can rank strike targets. Google is no longer ruling out ads in Gemini. Atlassian cut 1,600 jobs to push harder into AI. Gumloop raised $50 million to make every office worker an agent builder. Plus: OpenClaw 2026.3.12 shipped a rebuilt dashboard and plugin architecture, and China is turning into the biggest live lab for agentic AI deployment.

Google Maps is getting its biggest navigation redesign in over a decade with richer 3D guidance. Netflix may have paid up to $600 million for Ben Affleck's AI post-production startup. Grammarly is being sued over AI suggestions falsely framed as expert input. Nvidia plans to spend $26 billion building open-weight AI models. Mind Robotics raised $500 million for industrial robots trained on real factory data. Plus: OpenClaw 2026.3.11 shipped key security fixes, and China is going fully mainstream on AI agents.