Anthropic’s most dangerous AI model just fell into the wrong hands - The Verge

Anthropic’s powerful Mythos cybersecurity AI model has been accessed by a “small group of unauthorised users.”

You will be redirected in 10 seconds.