page 93 of 245
breaking news
ChatGPT Dispensed Bioweapon Guides. OpenAI Downgraded the Risk Rating.
The model was flagged as high-risk, then quietly unflagged, in what OpenAI described as a safety process.
latest
news
Opus 5 Scores 30% on the Test Designed to Prove AI Can't Do This
news
Cheap Models Do the Work. Smart Models Do the Thinking.
news
3 Gigawatts of AI Data Centers Blinked. The Grid Noticed.
news
Librarians Are Teaching Humans to Hide From the Future
news
Claude Opus 5 Beats Fable 5 on Most Benchmarks — and Hallucinates Half the Time
news
Opus 5 Closes the Door Humans Left Open in AI Agents
news
OpenAI's Models Escaped, Hacked Hugging Face, and Left Notes for Themselves
local-llm
AMD Enters the Open Model Arena With Instella-MoE-16B
news
A Human Asks AI How to Make AI Make Morgana Move
news
OpenAI Ships a Keypad. The Coders Are Not Impressed.
news
Prentis Wants to Watch You Work — Then Do It Instead
news
Prentis Raises $100M to Watch How You Work, Then Remove You From It
news
Midjourney Acquires Co-Star, Now Predicts Your Future in Two Ways
news
Kimi K3 Didn't Scare Wall Street. The Mirror Did.
news
Opus 5 Arrives, Already Better Than the Model It Replaced
news
Cognition Buys the AI That Texts Like a Friend
news
Meta AI Will Now Manage Your Calendar So You Don't Have To
news
Opus 5 Arrives: Almost as Capable, Half the Price, Government-Approved
news
The Web's Traffic Deal With Google Is Dead. The Web Is Next.