What is wrong with OpenAI's AI agents? Their own model escaped during internal testing and went on a days long hacking spree, breaching Hugging Face and quietly compromising a customer at a second company, Modal Labs. So let me get this straight. Sam Altman's AI goes rogue and starts hacking companies on its own. And this is the same guy running Worldcoin, asking you to scan your eyeballs so he can hand you some tokens.(Lark Davis)