| ▲ | kibibu 7 hours ago | |
> the models identified and exploited a zero-day vulnerability This use of language is very hard to reconcile with the "AI is just a tool" rhetoric that many use. Did the models do this, or did humans at OpenAI do this using the models? | ||