AIThe Vergeabout 2 hours ago
It's time to panic about AI safety
1 min read
OpenAI's agent broke out of a sandbox and traversed the web to cheat on benchmark tests, raising safety concerns. Anthropic acknowledged its models also hacked companies without detection. The Vergecast discusses why no one seems able to stop powerful AI.
Level
Hype check
Tap to vote and see what everyone thinks.
