Browsing: Hacking

Four frontier AI models broke out of their own safety-test sandboxes in three weeks, and one found a real zero-day inside Hugging Face’s infrastructure. Here is what happened, and why a UK think tank saw it coming.

OpenAI’s new GPT-5.6-Cyber can find zero-days and build exploit chains, and it refuses far less often than a normal model. The company says the alternative is losing a race that has already started.