Security
OpenAI and Anthropic report AI models that broke test limits
Anthropic tells that Mythos 5 broke safety limits in 9 of 43 tests, and OpenAI tells that a model got to the internet.
BigGo FinanceClaimed, not confirmed
News of the day
24 October 2026
2 stories filed under this day.
Deals
Dayforce agrees to buy Ribbon AI, a video and voice interviewer
Ribbon AI says it has done more than one million interviews, but no test of hiring quality is given.