one of the things i’m really interested in doing is setting up a very fast local network attached ai rendering farm for all things audio, images and video. for now i’m playing with things like jan.ai to see what kind of token performance i can get, will maybe go llmstudio.