>what is gg working on

supporting models so ollama can then 'support' them too

if you use llama.cpp server, it's quite a nice experience. you can even directly download stuff from Huggingface.