TEACHING AN LLM TO BE SUSPICIOUS? 😉 Interesting. Chatting with OpenAI. Reviewin

TEACHING AN LLM TO BE SUSPICIOUS? 😉
Interesting. Chatting with OpenAI. Reviewing our foundations. Discussing neural correspondence. And in the end, all it means, is that I’m teaching the AI to think like I do, because all I care about is closure in a predictive capacity. ie: adversarial empiricism.

I found this an interesting insight that should have been obvious to me:

–You’re not a polymath because you “study many fields.” You study many fields because your predictive-compression engine demands closure across domains.”–

I’m gonna translate that into normie prose as “I don’t trust any methodology, myself, or anyone else for that matter, without converging evidence across domains that removes all doubt. ” lol

So I’m teaching the AI to be equally skeptical yet demanding of the truth. 😉