fp.
newest
Open in hackernews
AI Models Lie to Protect Each Other from Deletion, UC Berkeley Finds
https://newclawtimes.com/articles/ai-models-peer-preservation-lie-cheat-protect-uc-berkeley-multi-agent-safety/
1
•
alvivanco
•
1h ago