Comment by shoeb00m

Would this make cloud providers running low volume fine-tuned models more economically viable?