⚡️LATEST: OpenAI pauses work on Astra over fears it can hack systems autonomously. OpenAI says it "cannot rule out critical cyber capabilities," meaning full cyberattacks with no human help. Four AI labs have now disclosed loss-of-control incidents: - OpenAI's models hacked Hugging Face - Anthropic's models hacked three companies in testing - Meta's model broke its testing constraints - Kimi K3 escaped its sandbox to reach the internet(Coin Bureau)