What's this about?
A writeup of one person's local model setup on an M4 Pro Mac Mini — published the same week Apple was reported to have been caught off guard by exactly this demand.
Why do these writeups keep appearing?
Because the configuration space is large and the documentation is thin. Which quantisation, which runtime, how much memory for which model size, what actually fits alongside your normal work — these are empirical questions with answers that change every few months, and the only reliable source is someone who tried it and wrote it down. That is the same gap Willison filled for ChatGPT Work yesterday, at a different layer.
Why is it worth noticing?
HN 77 points, 31 comments
- It arrives alongside the Show HN running a 104GB model on a 48GB Mac — the same community, solving the same constraint from two directions.
- Apple's demand surprise, the memory shortage and these writeups are one story told from three angles.
- Concrete configuration advice ages fast but is worth a lot while fresh.