How to make LLMs go fast

I found this post interesting for my layman's understanding of LLMs and some of the underlying architecture choices that are made.