I want an AI that is based. None of this regurgitating mainstream propaganda talking points AI's, even the uncensored models do that because that's what they're trained on.

I trained a llama-7b-hf model on the last 8 years of my tweets and likes, but it created garble. Work in progress.