A 45-page field guide to how machine learning works, written for people who use AI products every day and were never told what happens underneath.

It starts with what artificial intelligence actually is, and where machine learning, deep learning and generative AI sit inside it. Then the four parts every learning system has, the training loop that turns a random guess into a working model, and how a photo or an email becomes a list of numbers. After that come supervised, unsupervised and reinforcement learning, one neuron and then many of them, and the architectures behind the tools you already use: CNNs, RNNs, transformers.

The last part covers bias, privacy and what running these models costs, with a checklist to work through before you build anything.

The maths never goes past arithmetic. Every claim points at a product you have used or a number you can check, and the sources for those numbers are printed at the back.

What is inside:

18 chapters across seven parts, each chapter laid out on its own spread.
A 24-hour walkthrough of the eight models already running your day.
Diagrams for the training loop, attention, and the full path a prompt takes through a chatbot.
A spam filter taken apart end to end.
Two projects to do yourself: design a system on paper, then audit a real product you use.
An A to Z glossary, and a 30-day plan for what to read next.

PDF, 45 pages. Download it straight after purchase and read it on anything.