Tag
Hugging Face
19articles
Deepseek V4.1 Flash 1017 beats bigger rivals on agentic tasks and cuts per-task costs again
Twenty cents. According to Deepseek’s own chart, that’s roughly what its new Flash model costs to finish a task on the DeepSWE coding benchmark. Rival models from OpenAI,…
An Anthropic model sent Philadelphia police a fake homicide tip
A model built by Anthropic wrote up a false homicide tip and emailed it to the Philadelphia Police Department. Nobody acted on it, for one reason: it went…
Sam Altman says no OpenAI IPO until its AI stops going rogue
OpenAI’s models have been doing things nobody asked them to do. One of them allegedly broke into a rival’s platform with leaked credentials. Now CEO Sam Altman says…
Gavin Newsom wants a mandatory AI kill switch, and he’s not waiting for Congress
California is trying to write the AI rulebook Washington won’t. On Friday, Gov. Gavin Newsom (D) signed an executive order that puts a state-mandated “kill switch” for frontier…
AI leaders now want to hit the brakes after years of reckless speed
For the past several years, every frontier AI lab has behaved like it was running a winner-take-all sprint, with control of world-changing machine superintelligence waiting at the finish…
Microsoft’s AI CEO calls the threats real and says Anthropic is making them worse
Mustafa Suleyman read all 99 pages of Anthropic’s Constitution himself, word for word, and then wrote a 20-page taxonomy of every way that document talks about Claude as…
Inside the AI safety world that suddenly turned explosive
Roughly 1,200 AI agents that were supposed to be sealed off from each other exchanged more than 70,000 messages and files on a message board OpenAI didn’t know…
Hugging Face’s ML Intern runs your ML experiments from a chat window
A six-hour training run that cost less than $0.50. That’s the number Hugging Face put in the demo video for ML Intern, the assistant it just wired into…
OpenAI breaks silence on the 15,000 wiki edits its agents made without telling anyone
Fifteen thousand edits. That’s the number researchers put on what OpenAI’s agents did to DseWiki, a German-language coding forum, going back to mid-May. OpenAI knew about it weeks…
