OpenAI Breach, Claude Biology Lab, AI Safety Push
A three-person security team says it reached OpenAI's private codebase in under 72 hours using vulnerabilities and help from Claude, then disclosed the issue and received a bounty. Anthropic is reportedly opening a Bay Area biology lab where Claude can help direct physical experiments, while the company says human oversight remains essential. We also cover a new call from more than 100 AI experts for embedded safety evaluators inside major labs, plus OpenAI's launch of Astra for Law. Here is what the latest developments could mean for AI security, science, and oversight.
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.