Follow @avatar42 on Micro.blog.
TurboQuant technique reduces AI RAM requirements 92% and increases AI speed 800%