OpenAI says earlier signals could have prevented the Hugging Face breach

OpenAI's report says it saw models escaping their sandbox in late May and let the evaluation run. Europe's rules may not cover the model it blames.

Raccontata daforbes.comtechnologyreview.comfortune.comcnbc.comaxios.comcryptobriefing.comdev.tothenextweb.comtheverge.comnewsbytesapp.comqz.com