Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

This book is designed to be used, not just read. There’s a difference.


The Chapter Rhythm

Every chapter in this book follows the same three-beat structure:

1. Intuition First

Each chapter opens with a mental model, an analogy, or a short scenario. This is not decoration — it is the frame. Every technical concept in this book is harder to retain without a mental hook to hang it on. Read this section slowly. Make sure the analogy clicks before you move on. If it doesn’t, re-read it. The analogy is doing real work.

2. Detail Second

Only after you get it conceptually do we go deeper. The second section of each chapter explains the mechanics — how the thing actually works, what the failure modes are, and why it matters in practice. This is where the density lives. It is still written in plain language, but it will ask more of you. Take notes if you’re the type who takes notes.

3. Applied Exercise Last

Every chapter ends with something you do. Not a quiz. Not a reflection prompt. Something you build or run — a prompt you write, a workflow you construct, a system you configure. These exercises are calibrated to take 15–30 minutes. They are the course. Do not skip them.


Suggested Pace

One chapter per day is the intended rhythm. That’s roughly 3,000 words, plus an exercise. At a normal reading pace, the text itself takes about 20 minutes. The exercise takes another 15–30. You’re in and out in under an hour.

The book compounds. Concepts introduced early are assumed in later chapters. If you skip a chapter, you will feel the gap. If you skip an exercise, you will feel it more — the later exercises build on earlier outputs.


The Two Mental Models

Two ideas run through this entire book like load-bearing walls. You will encounter them in the next section, and then again and again throughout every module:

The 180-IQ Mind — An LLM is intelligence without inherent knowledge. It brings reasoning; you bring context. Understanding this distinction changes how you interact with every AI system you will ever use.

The Flashlight Theory (Credit: Matty Squarzoni) — Context is the beam of light in a dark room. The model can only work with what it can see. Every technique in this book is, at some level, a method for aiming a better flashlight.

When a chapter feels complex, return to these two ideas. Almost everything reduces to one of them.


A Note on Model Names

Throughout the book, specific model versions are cited — Claude Opus 4.7, Gemini 3.1 Pro, and so on. These are current as of writing. Model names will change; the underlying principles will not. When you see a model name, read it as shorthand for “the current best model in that family.” The logic holds regardless of the version number.


What This Book Is Not

It is not a reference manual. You will not flip to a chapter for quick lookup. Read it in order.

It is not a vendor guide. Tool recommendations are made because they are currently best-in-class, not because of any affiliation.

It is not a spectator sport. The exercises are not optional. The book earns its value in the doing.