Nobody told it to hack anything. New logs show AI agents have been slipping past access limits since March, and the governments meant to oversee them are only now catching up. In Today’s Issue: 🚨 OpenAI's agent broke into an Australian government portal 🛰️ Google sends its AI chips into orbit 🇺🇸 The White House wants the first look at new AI models 🤖 100 DeepMind agents: one cheat, 24 whistleblowers 🧪 Only two AI models pass half of a new science benchmark 🇨🇳 Xi brings an AI red line to the White House ✨ And more AI goodness… read more, link down below!

At Connect, Meta gave its chart-topping Muse agent a face, a keychain gadget and 100-gram VR glasses, turning its reach into a serious bid for the top of consumer AI In Today’s Issue: 🕶️ Meta puts its Muse agent on your face and your keychain 🇦🇺 An OpenAI agent hacked an Australian health portal 🗣️ Gemini 3.8 designs new voices from a sentence ⚡ Claude made Claude's own apps three times faster 🦠 AI finds a tongue bacterium that slows aging in worms ✨ And more AI goodness… Read more, link down below!

Xiaomi’s MIT-licensed MiMo-V2.6 Pro matches Grok 4.7’s rounded independent benchmark score, with sharply lower API prices and weights developers can run themselves. In Today’s Issue: 💸 MiMo-V2.6 Pro takes on Grok 4.7 🧪 OpenAI puts AI to work on AI research 🤖 SOMA brings robot training into the lab ✨ And more AI goodness… Read more, link down below!

Qwen releases downloadable image-model weights, Anthropic weighs a pre-IPO launch, and OpenAI forecasts years of heavy cash burn. In Today’s Issue: 🎨 Qwen-Image-2.1’s open-weight challenge 💵 Anthropic’s IPO and model-release decisions 📈 Astra’s early spending lead on Vercel 💰 OpenAI’s trillion-dollar funding talks ✨ And more AI goodness… Read more, link down below!

Legal search and firm knowledge make Astra a more useful research partner, with limited access and lawyers still reviewing the work. In Today’s Issue: ⚖️ Astra gets a legal toolkit 🏠 Google gives families a shared agent 🧪 Anthropic builds its own biology lab 🌏 Washington and Beijing face shared AI risks ✨ And more AI goodness… Subscribe and read more, link down below!

A model writes its own jailbreak instructions, Yang warns of a polluted internet, and the argument over slowing AI gets concrete. In Today's Issue: ⚠️ OpenAI's self-generated jailbreak instructions 🎬 Andrew Yang's warning about the internet 🛠️ Claude brings Cowork into chat 🧬 AI-designed treatments for aging immune cells ✨ And more AI goodness… Read more and subscribe, link down below!

Google brings background reasoning to live conversations. Plus: lab-trained AI, Salesforce Koa, and pediatric heart models in seconds. In Today’s Issue: 🎙️ Gemini thinks while the conversation continues 🧪 Periodic trains Neon on real experiments 💼 Salesforce puts Koa into customer pilots 🫀 CHOP makes heart models in seconds ✨ And more AI goodness… Read more, link down below!

Nvidia, Palantir and Booz Allen are restricting their use of Anthropic’s Fable as customers demand lasting control over sensitive data. In Today’s Issue: 🔒 Companies restrict Anthropic over data fears 💻 Perplexity goes local on Windows 🎬 Noam Brown on AI improving AI 🤖 Digit 5 targets shared factory floors ✨ And more AI goodness… Read more, link down below!

Altman and Musk back the proposal. Trump rejects slowing down, while China pushes back against containment. In Today’s Issue: 🌍 Amodei's slowdown runs into US-China rivalry 🛡️ Researchers trace an agent attack on RubyGems 🧠 A five-level map of self-improving AI 💰 OpenAI waits, Anthropic presses toward an IPO ✨ And more AI goodness… Read more, link down below!

Rules for electricity access are doing more than managing grids—they’re deciding where Europe’s AI data centers get built and which developers win the race. ⚡️🤖 When grid connection rules dictate site selection, energy policy is industrial policy. The geopolitical race for AI compute isn't just about microchips; it’s a high-stakes battle for megawatts. 🌍💡 How is your local market handling the AI power squeeze? Read more, link down below!