Modal Blog company blog

We reverse-engineered Flash Attention 4

modal.com published 2025-09-26

Asynchrony, fast approximate exponents, and 10x more efficient softmax.

Points

134

HN comments

48

Submissions

1

Cited by

0 commenters

AI & Machine LearningSoftware Engineering2025

Related posts from the same era

Statement on US government directive to suspend access to Fable 5 and Mythos 5

Anthropic20263,158 points2,314 commentscited by 11 details

GPT-4o

OpenAI Blog20243,138 points2,366 commentscited by 16 details

Astra can read and write weights to improve Qwen

RRikiya Takehi20262,949 points892 comments details

HackerTube: YouTube built from videos recommended by HN

karpathy20262,882 points640 comments details

Claude Fable 5

Anthropic20262,626 points2,159 commentscited by 10 details

Claude Code is steganographically marking requests

Thereallo20262,445 points750 commentscited by 3 details