Re:Bonsai 27B: A 27B-Class Model that runs on a phone
by kennywinker on Tuesday July 14, 2026 @10:50PM
(#48915668)
There is value in splitting things. If all I ever do is local app automations, i don’t need model that knows how to code. If all I ever do is coding, i don’t need a model that translates english to slovakian.
[ Parent
| on HN ]
Re:Bonsai 27B: A 27B-Class Model that runs on a phone
by ArcHound on Wednesday July 15, 2026 @12:24AM
(#48916299)
Slovakia mentioned, let's gooo. Ehm, exactly, we can achieve better smaller models for specialized tasks rather than using compute to improve a big model that does everything. There's a lingering philosophical question if better language processing capabilities translate to better image processing capabilities (i.e. having the vocabulary and experience to properly describe an image), but I still think that identifying tasks and splitting responsibilities saves a lot of effort.
Re:Bonsai 27B: A 27B-Class Model that runs on a phone
by bigmadshoe on Wednesday July 15, 2026 @01:07AM
(#48916507)
Good point! I thought you meant splitting them and then doing inference with some kind of learned router while keeping all the split models loaded at once. What you're suggesting is pretty sensible.
Re:Bonsai 27B: A 27B-Class Model that runs on a phone
by Zababa on Wednesday July 15, 2026 @03:52AM
(#48917580)
There is value in splitting things but there is also a cost. You have to train the specialized model, for that you have to know your use case, you have to hope the use case is going to be stable over time, you then have to see if you can remove english -> slovakian or coding from a model without affecting the useful parts.