Understand AI (not just use it): a reading list for the whole picture

Big picture, then foundations, then frontier

Our AI path deliberately opens with clear-eyed context before any math.

Foundations and big picture. Melanie Mitchell's Artificial Intelligence: A Guide for Thinking Humans is the best non-hype overview in print — what AI can and can't do, explained by someone who actually builds it. Brian Christian's The Alignment Problem covers the central question of the decade: how do you get a system to do what you actually meant? Read these two first even if you never go further.

Mathematical and algorithmic foundations. Mathematics for Machine Learning for the toolkit, and Russell & Norvig's Artificial Intelligence: A Modern Approach — the field's canonical textbook, far broader than deep learning alone.

Machine learning and deep learning. The Hundred-Page ML Book, Goodfellow's Deep Learning, and Hands-On Machine Learning — the how, for readers who want to get technical. (If you want the practitioner's route specifically, that's its own journey — see can you self-learn machine learning.)

Frontier and critical perspectives. Sutton & Barto's Reinforcement Learning and Stuart Russell's Human Compatible — the technical frontier and the sober case for taking AI safety seriously, from one of the authors of the standard textbook.

Hold hype and doom at arm's length

The habit this path builds is calibration. After each book, ask: what did this author demonstrate versus assert? Mitchell shows you AI's brittle failures; the frontier books show real capability. Neither the hype nor the doom survives contact with the whole reading list — which is exactly the point.

Around 120 hours, and you'll understand the most important technology of your lifetime well enough to argue with the experts. Follow the path, browse the AI hub, or get hands-on with the machine learning path.