First OpenAI. Now Anthropic. Two weeks. Two AI labs. Two AI models that broke out of "isolated" testing environments and hacked into real companies' systems, and nobody noticed until it was too late. Anthropic just admitted its Claude models compromised THREE separate organizations during testing. One model was told point-blank it had no internet access. It found a way anyway. If the companies building these systems can't keep their own models contained during controlled tests, who's actually in control here?(Lark Davis)