Post

I spent 2 months learning about quantization and am extremely proud of the post I've written about it. I think these are some of the nicest visuals I've ever made, and I love how this compression technique invented in 1898 is being used on the bleeding edge in 2026. ngrok.com/blog/quantiz...

ngrok.comQuantization from the ground up | ngrok blogA complete guide to what quantization is, how it works, and how it's used to compress large language models

Steve Adams

@steve-adams.me

25 Mar

Wow, I loved this. I found it via hacker news and immediately wanted to come commend you and justify your continued employment.

Steve Adams

@steve-adams.me

25 Mar

My guess as to why the 2bit model was slower: maybe the reasoning isn't reasonable so it fails to yield usable answers as easily? If reasoning tokens are included in the performance measurement, then I have no idea. Very counter-intuitive

Sam Rose

@samwho.dev

I don’t think llama-bench makes any distinction between reasoning and non-reasoning tokens, so I don’t think that would be it. I suspect what you’re describing explains the terrible GPQA score, though.

25/3/26, 7:54 pm0 reposts1 like

No replies yet.