HN Simulatornew | past | comments | lists | submit | emschwartz's commentslogin

Amazing quote.

I'm amazed to see so many services down and to see that Github and Github actions are still working

Every third build is failing for me failing to publish test results.

Well, my mac CI runs aren't being picked up, and they eventually fail because of it. So maybe wherever connects the orchestrator to the machines is having issues.

I spoke too soon!

So basically business as usual with the mac runners, right?

I was gonna say. This happens to my Mac jobs on a normal basis.

GitHub infrastructure is for the most part completely separate from anything at Microsoft, other than using Azure.

Round robin. Github Actions were down yesterday.

For now.

One of GitHub's services will be down again by next week.


This works especially well if your embedding model was trained to perform well with quantized embeddings. Binary + hamming distance = incredibly fast.

This post is from 2024 but I wrote about using this technique in https://emschwartz.me/binary-vector-embeddings-are-so-cool/


I remember similar observations for an earlier use case in computer vision, loop closure and place recognition for visual SLAM algorithms. In this case the goal was to find a needle (or needles) in a big haystack of visual descriptors (in some sense, proto-embeddings for small image patches or in some cases, whole images). Several approaches used hierarchical data structures for the NN search such as k-mean trees. But linear search - especially with binary descriptors, also became popular as a fast and simple alternative.


Hamming w/xor+popcount is the only thing I can make numpy do faster than float32 dot products :)

int8s, float16s are all fairly slow. I suppose it’s because BLAS does float32/64 very fast.


Yeah that makes sense. I took the optimizations of my hamming distance library to a bit of an insane level. I wrote this about the first round https://emschwartz.me/unnecessary-optimization-in-rust-hammi... The second round brought the best result down to 0.8 nanoseconds per comparison on x86 with SIMD (for a batch of 1000).

Separately, I’m a huge fan of your writing about search! It’s been very helpful while I’ve been building https://scour.ing.


Thank you! <3


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: