[gnicholas]

I run an offline LLM on my M2 MBA with 16 GB RAM. It's fine (but I will buy more RAM on my next machine, in large part due to LLM developments).