Why a 24 GB GPU Does Not Give Your Local LLM 24 GB

I keep seeing the same local LLM sizing mistake: "The model file is smaller than my GPU, so it...