News for hackers. Stuff that matters.
Re:Inkling 2 Mini by nsingh2 on Friday November 20, 2026 @09:08PM (#50585158)
I've been thinking about how small looped models are probably going to be the next big thing in local LLMs, and here it is. Exchanging a lower memory capacity requirement for more compute instead.

If the GPT-6.x models are looped, no doubt memory was a big reason.

There are no comments on this comment yet.