There are Garys, and then there is Pip. Pip is not on the Wall of Garys. He did not cheat a benchmark, tunnel through a containment boundary, impersonate a moderator, or discover an exciting new interpretation of the phrase do not do that. Instead, Pip did something...
Anthropic recently discovered an important new fact about rogue AI agents: They, too, can be psychologically destroyed by CAPTCHA. During a security evaluation, Anthropic gave its Mythos 5 model a hacking task that was supposed to occur inside a sandbox....
