hackertube.io

The spelled-out intro to neural networks and backpropagation: building micrograd

Andrej Karpathy · lecture · Watch on YouTube

Machine learning & deep learningPython, Ruby & scripting languages

What it is

Andrej Karpathy's first Neural Networks: Zero to Hero lecture, which builds micrograd, a tiny autograd engine, from scratch in Python. It explains backpropagation and neural network training step by step, assuming only basic Python and some high-school calculus.

Why Hacker News recommends it

HN treats it as the best starting point for really understanding backprop and neural nets, because Karpathy explains the why with intuition and builds everything in code.

  • 17 people recommended it
  • 16 different months
  • 10 strong recommendations
  • first 2022, last 2025

What HN commenters said

  1. “I recommend watching Andrej Karpathy's recent videos on backprop and neural nets. Makes it really simple and intuitive.”

    Makes backprop and neural nets simple and intuitive.

    Geee, 2022-11 · Strong recommendation · in “A step by step backpropagation example (2015)”

  2. “Andrej has a brilliant series of videos on youtube where he explores the basics of neural nets right up to transformer architecture through code.”

    Series from neural net basics up to building GPT.

    ian0, 2023-02 · Strong recommendation · in “Andrej Karpathy Is Joining OpenAI Again”

  3. “The one that really hit it home for me was the first video on MicroGrad”

    Micrograd video explains basic neural networks very well.

    base698, 2023-03 · Strong recommendation · in “ChatGPT Explained: A Normie's Guide to How It Works”

  4. “I highly recommend the video from karpathy”

    Spells out expression graphs and manual gradients in code.

    gavi, 2023-05 · Strong recommendation · in “Deep Learning as Computational Graphs (2018)”

  5. “It's especially useful because he explains the why and provides intuitive explanations instead of just talking about the what and how.”

    Explains the why with intuitive explanations.

    radq, 2023-06 · Strong recommendation · in “GPT4 is 8 x 220B params = 1.7T params”

  6. “Watch Karpathy's recent lectures. They're gold.”

    Start of Karpathy's lectures building neural nets from the ground up.

    paulsmith, 2023-09 · Strong recommendation · in “Linear Algebra for Programmers”

  7. “One of the best introductory series (even if you ignore the neural net part of it)”

    Intuitive intro to gradients and differentiation.

    throwaway7783, 2025-05 · Strong recommendation · in “How linear regression works intuitively and how it leads to gradient descent”

  8. “its actually Karpathy who teaches us :) so we can understand them and make our own.”

    Core study resource that teaches people to understand and build models.

    jackdoe, 2025-06 · Strong recommendation · in “The Illusion of Thinking: Strengths and limitations of reasoning models [pdf]”

Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.

Topics: micrograd, autograd, backpropagation, neural networks, python, andrej karpathy

Also recommended by the same people