Artificial Intelligence from Zero

You are not asked to trust this book. You are equipped to check it.

No calculus. No programming. No prior computer science.

Artificial Intelligence from Zero teaches how AI works to a reader starting from nothing. Its standard is that every core mechanism is worked, not waved at: you will trace searches by hand, update probabilities by counting, train a learning machine on paper, and walk gradient descent down an error landscape with a pencil. Where the field's full version of an idea needs university mathematics, the book teaches the honest arithmetic-level version and says plainly what the full version adds.

Inside:

• Search, games, and constraints - intelligence as trying possibilities in order


• Rules and proofs - machines that deduce, and where they break


• Belief and decisions - probability by counting, and choices that can defend themselves


• Learning, from a six-row dataset to deep networks - including a perceptron you train yourself


• Language models, vision, and robots - the machinery behind the systems everyone has met


• The limits - what AI cannot yet do (dated as of 2026), and a ten-question audit for any capability claim


• The stakes - bias, opacity, responsibility, and alignment, traced to arithmetic you have personally performed

Four short Toolboxes build every piece of mathematics the chapters need, at the moment it is needed. Fourteen misconception boxes give the field's canonical errors their own names and rebuttals. Every chapter closes with its key sentences collected for review and exercises with full answers.

And one commitment unusual enough to state up front: every doctrinal claim maps to a published source, every empirical claim was verified against a live source or recomputed by hand, and the complete evidentiary chain is printed in the back of the book - built entirely from sources that are free to read, so a reader with no money can audit every receipt.

For students before their first AI course, professionals who want the mechanism behind the headlines, and any citizen who wants to reason well about the technology deciding more and more of daily life.