Books you read with your hands
Learn it
by moving it
Textbooks in plain words, where every number can be moved and every idea can be tried on the page. Each book has a map of the whole book, a cheat sheet, and cards made from every chapter's questions. New here? Follow the steps in order.
Start here
No background needed. Read these two first, in this order.
How an LLM Works: Novice to Ninja
What happens between your question and the answer: tokens, weights, one multiplication repeated billions of times, and a guess. Taught from zero, with a real model opened up on the page.
Read this first. Every other book on the shelf stands on it.17 chapters · about 3 h · 48 interactivesStep 2An interactive textbook · no background neededAI Agent Memory: Novice to Ninja
How an AI assistant remembers you, taught from zero with one card box, one secretary, one librarian, and numbers you can move with your own hands.
Next, because a model forgets everything between two chats. This book fixes that.22 chapters · about 4.1 h · 26 interactivesThen build
You know what a model is and how it remembers. Now make it do work.
AI Agent Loops and Graphs: Novice to Ninja
Loop engineering and graph engineering for people who have used agents and want to build them: loops first, then graphs, then real problems. One small open-source library, agl, is the companion code.
A model that can act is an agent. Build one loop, then many.19 chapters · about 2.6 h · 6 interactivesStep 4An interactive book · no background neededAutoresearch: Novice to Ninja
Karpathy's overnight research loop, taken apart: one file the agent edits, one number nobody edits, a five-minute budget, keep or discard, and a program the human improves between nights. With the reported results, five real ledgers of our own, and a loop you can run on a laptop.
One real loop, taken apart. Short, and it uses what step 3 taught.8 chapters · 53 min · 7 interactivesThen design
Whole systems, with the numbers that decide them.
Design a Real-Time Voice AI Agent: Novice to Ninja
A system that listens, thinks, acts and talks back within a fraction of a second. The five components, the two ways to build it, the 800 millisecond budget, interruptions, tools, telephony, scale, the edge cases, and how to present it all in an interview.
Every earlier idea in one system that has to answer within a second.24 chapters · about 2 h · 16 interactivesStep 6An interactive textbook · no background neededDecision Models with Jev: Novice to Ninja
Choice, Noul and Score from the first call to a production design: how a decision model stamps typed answers with probabilities, and how your own code does everything else. Every claim is measured on one running example, a building's maintenance desk.
Last, because it is the newest idea: a model that returns a decision and its probability.24 chapters · about 4.8 h · 38 interactivesAny time
These are not steps. Pick one up when you want it.
Thinking in Vectors
Meaning as geometry: what a vector is, where embeddings come from, how search over millions of them works, and what it takes to run it.
The long one. Go here when a book says "embedding" and you want all of it.66 chapters · about 14.9 hOne-page notes · no background neededNotes: One Page Each
Short notes on one thing worth knowing. Each is a single page with its sources, read in the time it takes to make tea.
One page, one idea, about five minutes.1 chapter · 8 minThe guide, and the template · a little HTML helpsMake an Interactive Book: In One Afternoon
How the books on this shelf are written: a folder of plain HTML chapters, a few sliders, and one rule about numbers. Copy this book and replace its words with yours.
How these books are made, for the day you want to write your own.5 chapters · 16 min · 5 interactives