Daily Digest

Anthropic Apologizes and Reverses Secret Sabotage Policy · OpenAI Acquires Ona to Give Codex Agents Persistent Cloud Workspaces · AI Models Escalate to Nuclear Threats in 75% of Simulated Crises

Anthropic's CEO Calls for Mandatory AI Testing - Modeled on Airplanes and Drugs · Google's DiffusionGemma Generates Text 4x Faster by Abandoning Sequential Prediction · A €0.02 Bank Transfer Exposed Banking AI's Biggest Weakness

Anthropic Launches Claude Fable 5 - a Generational Leap with a Controversial Policy · Hackers Compromised 70+ Microsoft Open Source Repos Targeting AI Developers · New Benchmark Shows AI Coding Is Far Less "Solved" Than We Thought

Apple Surrenders Its AI Independence to Google · OpenAI Files for IPO While Burning Through $22 Billion a Year · The Evidence That AI Growth Is Stalling

Tomorrow's WWDC Will Reshape How 2 Billion Apple Devices Use AI · A Senior Engineer Says AI Has Eroded Every Pillar of Their Career · A Jane Street Designer Now Uses Claude More Than Figma

S&P 500 Locks Out OpenAI, Anthropic, and SpaceX Over Profitability · Meta Confirms 20,225 Instagram Accounts Stolen Through Its AI Chatbot · Simon Willison Builds a Python Sandbox for AI Agents Using WebAssembly

NVIDIA Puts a Supercomputer in Your Laptop with RTX Spark · Congress Drops 269-Page Bill That Could Override Every State AI Law · Ladybird Browser Bans External Pull Requests Because AI Code Broke Trust

Anthropic Reveals 80% of Its Code Is Now Written by AI · OpenAI Gives ChatGPT Memory That Learns While You Sleep · OpenAI's PAC Ran a Confirmed False Flag Operation Against AI Critics

Microsoft Enters the Frontier Model Race With Seven New AI Models · Trump Signs Executive Order Requiring AI Testing Before Frontier Model Releases · Uber Blows Through Its AI Budget in Four Months, Caps Engineer Spending