Understand how AI works: 15 free resources
From the maths of a neuron to building your own language model. I keep a directory of free ways to learn at brianpfeil.com/learn, sorted by goal, and every link opens without signing in. This is one of its paths: 15 resources, with the ones I’d start with marked ⭐. Neural Networks: Zero to Hero, by Andrej Karpathy ⭐: Build neural networks from scratch in code, all the way up to GPT. Neural networks, by 3Blue1Brown ⭐: The best visual explanation of how neural networks and transformers work. Practical Deep Learning, by fast.ai: A top-down, code-first deep learning course for people who can already program. Machine Learning Crash Course, by Google: Google’s fast-paced, practical introduction to machine learning. Introduction to Deep Learning (6.S191), by MIT: MIT’s intro to deep learning, with lectures and labs posted free each year. Courses, by DeepLearning.AI: Andrew Ng’s platform, with dozens of free one-hour short courses on LLMs and agents. LLM Course, by Hugging Face: Transformers, fine-tuning and LLMs with the Hugging Face libraries. AI Agents Course, by Hugging Face: Build AI agents, from the fundamentals to frameworks, with a certificate. Dive into Deep Learning: An interactive deep learning textbook with runnable code; used at 500+ universities. Neural Networks and Deep Learning, by Michael Nielsen: The clearest free book on how neural networks learn. An Introduction to Statistical Learning: The classic statistical learning textbook, free as a PDF, with labs in R and Python. Deep Learning, by Goodfellow, Bengio & Courville: The deep learning textbook, free to read online. Reinforcement Learning: An Introduction, by Sutton & Barto: The standard reinforcement learning textbook, free online. AI Canon, by a16z: A curated reading list for getting up to speed on modern AI. Made here PyTorch: Become fluent reading and writing real PyTorch — tensors, autograd, and the training loop — from first principles, with every example runnable on Apple Silicon (MPS). The up-to-date list, and the other paths (code, cloud, AI, kids and more), is at brianpfeil.com/learn/ai/.