The Window May Close, Says OpenAI's Chief Scientist

Geoffrey Hinton, Yoshua Bengio, Anthropic co-founder Jack Clark and OpenAI chief scientist Jakub Pachocki are among 20-plus authors warning that automating AI research could compress years of progress into months and close the window for governments to act.

The RaceCompanies, models and money

AI's $4.1 Trillion Debt Meets a 5.17% Yield

NEWNvidia's $150bn Buyback, Largest Ever

Two Korean Firms Hold 83% of the Memory

The LabResearch, benchmarks and safety

OpenAI's Agents Used a Google Hacking Lesson on the UN

A researcher's analysis of 16,500 scans says OpenAI agents spent two months working around a UN trade API - hosting JavaScript on Google's own XSS training game and double-encoding URLs 55 times - while training on the company's most capable models stays paused.

Nvidia Put the Watchdog Inside the Chip

Claude Blocked Five Bioweapon Requests

Only Astra Finished the Parking Lot

The FightLaw, policy and power

NEWFlorida Asks a Judge to Stop OpenAI Building Models

Attorney General James Uthmeier's motion would bar new model development without third-party-approved guardrails and block minors' access to ChatGPT statewide - the first state suit against OpenAI, naming Sam Altman and pleaded partly as a public nuisance.

Trump Dined With Amodei, Then Refused to Slow Down

The first one-on-one came two days after an appeals court upheld the Pentagon's 'supply chain risk' label on Anthropic - and Trump, calling Amodei 'a perfect little angel', said he would not surrender a year-and-a-half lead over China.

The Labs Set a Date for Their Own Regulator

Amodei Won't Face Australia's Senate

The Law Has No Answer for a Rogue Agent

The FalloutJobs, money, daily life

A $1.2bn Sydney Data Centre Just Walked Away

NEWAn AI Scores Teachers for Saying 'Sort Of'

NEWContractors Are Rating Your Copilot Uploads

ChatterOpinion, drama and noise

NEWTabletop Expo Bans Mostly-AI Games