[dopidopHN]

I’ve try because it’s tempting and the first attempts do give a « conversation » vibe.

I was curious about state persistence between prompt, or how to get my prompt better, or having a idea of the training data.

Only got crap and won’t spend time doing that again