In the Middle Ages, the peasant brought his lord to the table by walking off the land. Today you hold the same leverage — because AI models get dumber without human data.
Every age has a master, and the master is whoever controls that age’s scarcest resource. When the resource changes, the master changes with it — but the structure almost never does: a lord at the top, an intermediary class that administers, and the people who actually produce.
From Feudal Lords to Algorithms traces that structure across four transformations — the agricultural revolution, the commercial revolution, the industrial revolution, and now artificial intelligence — and asks a single question: what leverage has the ordinary person ever had, and what is it now?
The answer offered here is specific and uncomfortable. The scarce resource of the AI age is not computing power. It is fresh, real, human-generated data — the one input these systems cannot manufacture for themselves without degrading. Which means the leverage is already in your hands. You simply haven’t been told.
Read a chapter before you buy
Chapter 10, The Five Conditions of Leverage, is published here in full and free to read. It is the framework the rest of the book runs on: why the same plague dissolved serfdom in Western Europe and hardened it in the East, and what that says about the age of AI.
Editions
Kindle
69 pages · published 24 August 2026 · English
Buy on Amazon
Paperback
68 pages · ISBN 9798194245512 · English
Buy on Amazon
About the author
Ant Gökçek was trained as an econometrician and spent twenty-five years watching power move from one scarce resource to another — usually from the inside. International banking in Istanbul, administration and finance for an infrastructure programme in Libya, and cross-border real estate across Turkey, Spain and Lithuania. He now builds AI tools for small and mid-sized companies through Ernex, UAB in Vilnius.
He researched and wrote this book in close collaboration with AI, and says so plainly in the preface — a book about the value of human data could hardly hide how it was made.