True, I've been telling that coding LLMs have been detrimental to local AI scene. What happened to those nifty ML projects which can run on raspberry pi zero.