HackerTrans
トップ新着トレンドコメント過去質問紹介求人

feelingsonice

57 カルマ登録 8 年前

投稿

Show HN: Unify memory across agents and improve context rot, written in Rust

github.com
5 ポイント·投稿者 feelingsonice·3 か月前·1 コメント

The Missing Link Between the Transformer and Models of the Brain

arxiv.org
8 ポイント·投稿者 feelingsonice·9 か月前·1 コメント

コメント

feelingsonice
·3 日前·議論
No AdaJEPA is a deployment-time mechanism, it's not about building JEPA at all
feelingsonice
·3 日前·議論
Okay I see it now, but it's briefly mentioned without describing the mechanism. The author didn't annotate it the same way they annotated JEPA. I feel like LeJEPA should take precedence in the article since it supersedes JEPA.
feelingsonice
·3 日前·議論
LeCun's new paper takes a new approach to building JEPA, which he named LeJEPA -- https://arxiv.org/abs/2511.08544 It's not that different in the underlying theory but he removes the EMA and the twin tower, makes the whole thing more straightforward.
feelingsonice
·27 日前·議論
PASETO is great but there's not enough ecosystem support
feelingsonice
·先月·議論
Is Liquid AI still using the liquid neural network architecture?
feelingsonice
·2 か月前·議論
[dead]
feelingsonice
·2 か月前·議論
Because coming to the United States on a non-immigrant visa is pretty much the only way that a person can hope to become a US citizen (or green card holder) eventually.
feelingsonice
·6 か月前·議論
Someone else posted this on HN a while back -- https://github.com/Hugo-Dz/spritefusion-pixel-snapper

Maybe you can use that^ to snap the pixels to a perfect grid
feelingsonice
·11 か月前·議論
I have the pro plan but don't seem to have access to it?
feelingsonice
·昨年·議論
Does this mean that the o1 and o3-mini models are also using 4.1 as the base now?
feelingsonice
·昨年·議論
Why won't they add benchmarks against o1?