News for hackers. Stuff that matters.

Icko_

Icko_ on Hacker News

Re:How Distillation Makes AI Models Smaller and Cheaper on Thursday July 24, 2025 @04:39AM (#44668442)
distillation and teacher-student models are definitely way older than 2024.
Re:Breaking Quadratic Barriers: A Non-Attention LLM for Ultra-Long Con... on Monday June 16, 2025 @05:20PM (#44293502)
Not even that. With KV-caching, it's linear with the size of the context; and if someone figured out a way to have e.g. NlogN complexity, I imagine with KV-caching it may go down to logN complexity. (If the new algorithm permits that.)