Info This post is auto-generated from RSS feed Hacker News. Source: Two different tricks for fast LLM inference