30u30

A study guide for learning 30 foundational AI papers through hands-on implementations, notes, and exercises. It helps readers connect paper ideas to working deep-learning concepts one paper at a time.

Browse open issues

Pick a small issue you can reproduce, then fix it with your AI coding agent.

Share on XCustom license (see repository)

Overview

30u30 is an open-source study guide for learning foundational AI research through a one-paper-per-day path centered on 30 papers recommended by Ilya Sutskever. It helps bridge the gap between reading a paper and understanding how its ideas work by pairing papers with from-scratch implementations, notes, interactive exercises with solutions, and visualizations. Learners follow the sequence, study each paper and its math, then work through the implementation and exercises.

Key features

  • A guided sequence covering 30 foundational AI papers, one per day
  • From-scratch implementations, including a character-level RNN and an LSTM
  • Interactive exercises with solutions for each paper
  • Notes and visualizations to support understanding

Best for

Learners who want a structured path through foundational AI papers and prefer to explore ideas by implementing them and completing exercises.

Upstream
Ojhaharsh/30u30
Fork on GitHub
Guo-astro/30u30
Upstream stars
1.7k
Category
Learning, interviews and curated lists
License
Custom license (see repository)
Forked
2026-10-16
Sync status
In syncLast synced 2026-11-08

More in Learning, interviews and curated lists