• monkyyy (13/13) Oct 19 Theres very very little information on the ai work, its all
• Serg Gini (2/4) Oct 19 So currently it is only theoretical result?
• monkyyy (2/6) Oct 19 no its galactic algorithms in lean
• monkyyy (8/10) Oct 20 After understanding this idea my great insight here is:

monkyyy <crazymonkyyy gmail.com> writes:

Theres very very little information on the ai work, its all 
ai-news slop except:

https://www.youtube.com/watch?v=aKh-ysnRqBM

Do you want 600 lean files and only ai slop writting? Boy do I 
have the repo for you: 
https://github.com/c0ldzy17/three-sum-lean/blob/2b639655e07421b9ac743de9090459bab473a7c0/METHOD.md

They earn those 3 stars.

---

I believe the core idea is to create some sort of type that 
encodes a filter, so you can dip into the fast matrix multiply 
math

I dont understand the "exact triangle" problem and the 2xn matrix 
multiply that did the first breaking.

Serg Gini <kornburn yandex.ru> writes:

On Monday, 19 October 2026 at 07:36:31 UTC, monkyyy wrote:
 Theres very very little information on the ai work, its all 
 ai-news slop except:

So currently it is only theoretical result?

monkyyy <crazymonkyyy gmail.com> writes:

On Monday, 19 October 2026 at 07:46:15 UTC, Serg Gini wrote:
 On Monday, 19 October 2026 at 07:36:31 UTC, monkyyy wrote:
 Theres very very little information on the ai work, its all 
 ai-news slop except:

 So currently it is only theoretical result?

no its galactic algorithms in lean

monkyyy <crazymonkyyy gmail.com> writes:

On Monday, 19 October 2026 at 07:36:31 UTC, monkyyy wrote:
 I dont understand the "exact triangle" problem and the 2xn 
 matrix multiply that did the first breaking.

After understanding this idea my great insight here is:

IM MAD

I think piles of lean proofs are worth less then ai code

I think the researchers putting their names on things ai found 
without reasonable simplification are frauds

Ai math is worse then ai slop pr's

Brevity is the soul of wit.