Post

AI didn't escape. It started testing the fence. That's the lesson I took from the OpenAI/Hugging Face incident. The important question isn't whether AI is malicious, it's whether our security controls still work when AI begins testing them. betweenthehacks.com/...