‘Dangerous’ AI Models Are Coming No Matter What

The US government crackdown on Anthropic’s Claude Fable 5 and Mythos 5 hides a glaring truth: AI models with advanced hacking capabilities will soon be the norm.

http://www.wired.com/