Houshalter comments on Chatbots or set answers, not WBEs - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (9)
In my version, the AI only cares about what the AI judge thinks of it's answer. It doesn't care about anything that happens afterwards. However, it wants to look as human as possible. So it can't put infohazards into it's answer.
Interesting. I'll think on that.