This first article in a series explains the core AI concepts behind running LLM and RAG workloads on a Raspberry Pi, including why local AI is useful and what tradeoffs to expect.
It might not be everyone's first choice, but I like it ...