Two questions before you run a local LLM: does it fit, and how fast is it really?

I keep seeing the same two failures when people try to run a model locally. The first is buying or...