No. We're starting to figure out that OpenAI is both incompetent and unethical. Nothing is "escaping." LLMs are just burning compute to accomplish tasks they're assigned by incompetent and unethical people at the company.

If they set up their sandboxes properly, none of these hacking incidents would happen.