Why I Revised My Book

I wrote the first edition of Punching Above Your Weight in a week, in June 2025. I’ve told that story enough times that some of you could tell it back to me: the draft, the ICU, the month of five a.m. mornings to get it published. Speed was the point. My dad had just come back from fifteen minutes without a heartbeat, and I had stopped believing in “later.”

Sixteen months on, I went back and rewrote it. The revised edition is out now. This is why.

Sixteen months is a long time now

When the first edition went to press, the models on the marquee were GPT-4o, o3, Claude Opus 4, and Gemini 2.5. Those names are trivia already. Anyone who memorized the first edition’s tool list has my sympathy.

The capability moved just as fast. A research nonprofit called METR measures how long a task an AI system can finish on its own, counted in the time the same job takes a skilled person. The best model on sale when I published could handle about two hours of work and succeed half the time. By this spring, the measurement had reached about sixteen hours, and by METR’s count the figure has been doubling roughly every four months since 2023.

Here is the part that should interest a skeptic. Gallup has asked American workers about AI since 2023. That year, about one in five used it at work at least a few times a year. By May 2026 it was half of us. Daily use went from four percent to fifteen. The machines got dramatically better in sixteen months, and the average workday did not.

I don’t read that as proof the tools are useless. I read it as the distance between owning a tool and knowing how to use one. Gym memberships have the same problem.

So the new edition teaches roles and habits, and it treats product names like weather. That change alone meant rewriting the Introduction from the first line.

I couldn’t trace my own quotes

This is the part I’m least proud of and most glad I did.

The first edition had quotations I’d carried around for years. Lines attributed to fighters, Stoics, and founders, the kind that show up on posters and in keynotes. When I went back and checked each one against a real source, several could not be traced to anything reliable. Some were paraphrases that had hardened into quotes. Some were simply invented, decades ago, by somebody who needed a good line.

I removed them or replaced them with words I could verify. A few others are now marked “attributed,” where a saying is widely credited to someone but the original source is uncertain.

I wrote a book that tells you to check what the machine hands you. A book like that has to meet its own standard. Mine didn’t yet, and no amount of good intentions changes that. Knowing about the tool and checking the tool are different habits, and only the second one protects you. I include myself in that.

The research is in the text now

The first edition made claims the way most business books do: confidently, with the evidence somewhere behind the curtain. The revised edition puts the research in the text and the sources at the back. There’s a page at the end called “The Evidence on One Page,” because if I’m going to ask you to change how you work, you deserve to see what I’m standing on.

Rules, because enthusiasm gets people in trouble

I’ve spent the past year building AI products inside a large company and talking to people at library branches and civic groups about how they use these tools. The pattern I kept seeing was the same in both rooms. People get excited, and then they get in trouble, because nobody gave them rules.

So the revised edition has rules for using AI at work. What you can paste into a chat window and what you can’t. When to check and how. What to do when the tool is confidently wrong, which it still is, more often than the marketing suggests. None of it is complicated. All of it is the kind of thing people wish someone had told them before the meeting where it went badly.

What didn’t change

The boxing. The twelve rounds. The song at the end of every chapter. The preface about the morning after, which I didn’t touch, because I can’t improve on how that felt. The five moves still spell FIGHT, and the stance is the same: fight smarter than your size suggests, and keep what’s yours.

The language got plainer. The standard of proof got higher. The book is still for the smallest person in the fight.

Train, fight, revise

In the first essay I wrote about this book, I ended with a line I meant as advice: “Train, fight, revise, and then send it out into the world.”

It turned out to be a schedule.

If you read the first edition, thank you. You read it while it still felt urgent, which is what I asked of you. If you’re coming to it now, you’re getting the version I can stand behind line by line. Either way, the fight hasn’t changed. Only the tools have, and they’ll change again before I finish this sentence.

That’s fine. The book was never about the tools.