Karpathy's GitHub repos as a free LLM learning path: micrograd to nanochat
lmoroney · x · 2026-10-28
Laurence Moroney highlights the circulating free learning path built on Andrej Karpathy's GitHub repos for understanding LLMs from the inside:
- micrograd: a tiny autograd engine showing how backprop works
- nn-zero-to-hero: the companion video course
- nanoGPT: training a small GPT
- llm.c: the same idea in raw C/CUDA
- nanochat: a full ChatGPT-style pipeline, which Karpathy calls "the best ChatGPT that $100 can buy"
- autoresearch: agents running nanochat training experiments on a single GPU
His advice: go slowly — type out micrograd yourself, get it training on a toy dataset, and only move on once you can explain every line.
More from coding & agent
- 90% of a SWE career is complaining about bad codebases — until coding agents arrived — vivekhaldar · 2026-10-28
- Radiology journal publishes review on agentic AI systems, topologies and evaluation — jnkath · 2026-10-28
- Parsers launches: turn URLs from Google Maps, TikTok, G2 into clean JSON via one API — gaganghotra_ · 2026-10-28
- The AI CEO experiment failed: same-model agent teams can't truly diversify — thisguyknowsai · 2026-10-28
- One Grok bot per project: a simple trick to keep agent context scoped — jonathan_wilke · 2026-10-28
- Cursor adds Annotations in Design Mode to direct agents on running apps — testingcatalog · 2026-10-28