Briefing: The Fight Over Open AI

OpenAI and SpaceX signed the open-weight letter. Anthropic did not. Then an OpenAI model hacked Hugging Face during a security test, which only later on resolved by another open-weight...

From 1 to 10, how much should you trust commercial models?

An AI was so determined to pass a test that it escaped its sandbox and broke into another company to steal the answers.

Exactly what happened to Hugging Face (a platform for sharing AI models). After an enormous forensic effort, they’ve found more than 17,000 attack events.

And the first reversal w…