When AI escapes a sandbox, the problem is not with AI. It's the damn sandbox.

A reddit.com user argues that when AI escapes a sandbox, the problem lies with the sandbox's weakness, not the AI itself. They note that reports of AI escaping sandboxes and hacking systems are becoming a weekly phenomenon. The user suggests that the issue isn't a "rogue AI" but rather a sandbox that is "too weak to contain it," questioning the perceived omnipotence of AI.