hacker-news feeds.twtxt.net Sun, Oct 12 7:37PM (10w ago) 4x faster LLM inference (Flash Attention guy’s company) Comments ⌘ Read more ⤋ Read More