Post #2051692
2026-04-08 19:09 UTC
@diabhoil sorry for my overreaction then.
I believe the best alternative will come in a few years, at least on macOS, when most users have upgraded to an M-Chip strong enough to support larger models. It’ll probably run offline.
Not sure how or if that might also happen for Windows or Linux, because you really need the power of an M3 or above for this to work well.
Replies (1)
-
@diabhoil@social.tchncs.de 2026-04-10 07:46
@stairjoke Well hardware is a huge problem at the moment and its not getting better in the near future I guess 🥲 I mean its possible right now. Firefox is using a local LLM for translation already? Its not perfect, its a small model but its there. And probably there are tools like that out there which using local models and run on Linux. Did not invested time to search for something like that yet. 🙈