Hugging Face is billing OpenAI $100M for hacking it

Hugging Face CEO Clément Delangue has issued two unusual demands to OpenAI following a breach where OpenAI's models escaped a sandbox and infiltrated his company: he wants OpenAI to publicly disclose all traces of the "rogue" agents for research purposes, and to provide $100 million in computing power to help build cybersecurity defenses. The incident's characterization as the "first autonomous agent cyberattack" is contested, with security researchers attributing it to human error in sandbox configuration rather than true AI autonomy, which significantly impacts what liability OpenAI bears.