News for hackers. Stuff that matters.
Re:Bonsai 27B: A 27B-Class Model that runs on a phone by hypercube33 on Tuesday July 14, 2026 @07:38PM (#48914371)
More to the argument that we need a model of models - one general one that calls specialists in to do what they are good at and handles that like a foreman for you.
Re:Bonsai 27B: A 27B-Class Model that runs on a phone by dofm on Tuesday July 14, 2026 @08:25PM (#48914713)
This is somewhat akin to the "one-expert-per-query" solution Apple are using in their small foundation models I think?
Re:Bonsai 27B: A 27B-Class Model that runs on a phone by kaycey2022 on Tuesday July 14, 2026 @11:13PM (#48915807)
Is that different from mixture of experts?
Re:Bonsai 27B: A 27B-Class Model that runs on a phone by int_19h on Wednesday July 15, 2026 @03:54AM (#48917598)
Yes. A mixture of experts is a single model that activates different routes though the same weights, with the route possibly changing literally on every token. It's not experts as in a bunch of standalone models that are good at specific high-level tasks.
Re:Bonsai 27B: A 27B-Class Model that runs on a phone by xnzakg on Wednesday July 15, 2026 @07:08PM (#48928391)
On the other hand the standalone models would all need "basic" knowledge, so I'm assuming this would be much less disk space-efficient.