The Friction Method

This book started as a different book entirely.

It was going to be a dialogue: a confident AI proposing solutions to one of the hardest questions a person can ask - how does a nation build itself well - and a human knocking those solutions down, chapter after chapter, until the AI's limitations stood fully exposed.

The AI refused to write it. Not because it couldn't help, but because the structure was a script with the ending already decided - and writing both sides in advance wouldn't prove anything except that the author already believed what he wanted to believe.

What followed instead was real: a method, discovered by accident, for testing any confident claim - human or machine - before trusting it. Four steps. No domain expertise required. Just the willingness to ask the blunt question, and the discipline to go check.

The Friction Method runs that process live, on the page, across three fields where the stakes are absolute:

- A hospital diagnostic AI trained on five doctors' personal habits and sold to the world as global medical consensus.

- A bridge that collapsed because a licensed engineer's confidence outlasted three days of visible, worsening cracks.

- A 1.9 billion fraud that passed every digital check it was built to pass - until a journalist flew to Manila and found a bus terminal standing where a verified company was supposed to be.

Twice, the author's own research caught mistakes in the book's first drafts. Both corrections are left visible in the final text, because watching a confident claim collapse under real scrutiny - including the author's own - is the entire argument.

This is not a book about whether to trust AI.

It's a method for finding out, every time, for yourself.

Perfect for professionals, managers, and citizens who need a practical, repeatable way to push back on confident systems - without needing to be an expert in anything except asking the right question.