Comment by pixl97

I mean, yes, if you ever studied the history of AI safety the fact is someone was always going to build it. End of story. The question was always would we make it safe before it does.

> Also, if their model is so smart, why didn’t they use it to design the sandbox?

"Can god make a rock so big that he can't pick it up", and other stupid sayings.

First, NEVER FUCKING EVER have the models you're making also be in charge of security. This is the first rule of AI safety, because if you're model is deceptive then it will leave hard to see holes everywhere to escape from.

>Or were the humans too lazy?

Of course they were. If you're hinging our future on humans not being lazy, we'll it was nice knowing us. There are not really any fail safes on LLMs or AI in general.