Hey, I would gladly join your effort, I got a similar background and certain concerns regarding the direction of OpenAI in their approach to censorship. Currently still mostly inexperienced with machine learning, with a mediocare understanding of the linear algebra algo's behind it. I intend to use (and currently partially use) ML for image gen, improving formal software verification by possibly generating SMT conditions and such + aiding procedural generation algo's...

I would not be too concerned with ChatGPT costing a bit of money but rather the API or functionality being neutered because "too powerful". As such, I rather have control over the the whole AI stack.

Long term, I would also like to investigate the possibility for massive GPU based distributed training, similar to Folding@home just for generating models.

Discord/ Element/ Telegram - I am free to talk :)