Hacker News new | past | comments | ask | show | jobs | submit
manmal | on: AMD acquires Taalas to boost inference performance by etching models in silicon
That’s their API with extra steps, or am I missing something? That was always faster.