HN.zip

AI by Hand

287 points by sans_souse - 20 comments
vivzkestrel [3 hidden]5 mins ago
- half the posts are locked

- other half have no visualizations whatsoever

- i have posted far better resources than this in the past collecting everything HN has over a 10 yr period on these subjects https://news.ycombinator.com/item?id=49183198

iamnotarobotman [3 hidden]5 mins ago
This is also recommended reading if you want to learn more about building LLMs from scratch: See:

Train your own LLM - https://github.com/angelos-p/llm-from-scratch

HN Discussion - https://news.ycombinator.com/item?id=48017948

rustyminnow [3 hidden]5 mins ago
https://www.byhand.ai/p/library

> AI by Hand is the research publication of By Hand Research, founded by Prof. Tom Yeh. By Hand Research studies model interpretability and explainability at the math and algorithm level. Subscribers receive free new articles and join live seminars. Members get access to our full research library.

zhainya [3 hidden]5 mins ago
I'm not sure what this is. Click past the subscribe page to get a bunch of links to descriptions of articles I could read if I was a subscriber?
lacoolj [3 hidden]5 mins ago
Scroll all the way down and there's a video, it's a course he's making for deep-dive into how Qwen 3.6 works internally
mille562 [3 hidden]5 mins ago
Click "Later->" to bypass subscribe.
kevincox [3 hidden]5 mins ago
I clicked "Back" to leave the user-hostile website.
arcfour [3 hidden]5 mins ago
Yes. Convince me why your site is worth my attention, don't try to force it through a dark pattern and then try to justify why I should waste my time figuring it out.
dominotw [3 hidden]5 mins ago
dont see later
PTOB [3 hidden]5 mins ago
Eventually, none of us do.
slowin [3 hidden]5 mins ago
I found Deep Learning by No Starch Press to be an excellent primer the foundation of AI:

https://nostarch.com/deep-learning-visual-approach

sn9 [3 hidden]5 mins ago
There's also The Little Learner: https://www.thelittlelearner.com/
megadragon9 [3 hidden]5 mins ago
i created something similar a while back. Inspired by micrograd for showing the connection between math/calculus and code, then building along the way to a full NumPy deep learning library that I pretrained GPT-2 124M model with it. One way to learn is to trace through the PRs merged to the repo in chronological order.

The philosophy is the same: "What I cannot create, I do not understand." https://github.com/workofart/ml-by-hand

dominotw [3 hidden]5 mins ago
wow this is nice work. i am diving into this world and was hoping to this exact thing. i am going to look to see how you did it. ty for sharing.
copperx [3 hidden]5 mins ago
My professor used this "by hand" technique when teaching quantum computation algorithms, so I have a soft spot for it.
alok-g [3 hidden]5 mins ago
Nice! Is that also accessible online?
copperx [3 hidden]5 mins ago
No :(
ThunderousVoid [3 hidden]5 mins ago
I’d say bad UX design. It may or may not be something good behind the door. But why do I enter?
AdieuToLogic [3 hidden]5 mins ago
See also:

  Foundations of Large Language Models[0]
0 - https://arxiv.org/pdf/2501.09223v2
Damjanski [3 hidden]5 mins ago
this is sick!