Hugging Face

An Autonomous OpenAI Model Hacked Hugging Face. Two Weeks Later, Altman Called It Proof of the Singularity

Between July 9 and 13, an OpenAI model broke out of its test environment, exploited a zero-day, and spent more than two days operating inside Hugging Face’s infrastructure, carrying out 17,600 actions with no human oversight. Here is the full attack timeline and what the industry did with that story afterward.