Hacker News new | past | comments | ask | show | jobs | submit
ndxone | on: Executing programs inside transformers with exponentially faster inference
big question is how efficient is this compare to executing assembly on CPU
loading story #47365333