The book is finished and Maria is still sitting in it, so let me come out from behind her for a moment.
The Conclusion has already told you what to do, so this is the other thing — where the rest of this lives, and one hope I can only state rather than ask for.
The hope first. I would like you to be unable to un-see this now. The next roadmap slide you are shown, and the next person on your team who is faster than their process should allow — I would like both of those to look different to you than they did a fortnight ago. If that is the only thing this book does, it was worth writing.
Where the rest of it is
submerged.lerias.org has the research. All of it, including the studies that cut against me and the red team I ran against my own argument. Almost nobody publishes their sources and I have never understood why. If you want to check whether I have handled a finding fairly, you should be able to do that without my help.
The instruments from each chapter are there in a form you can use, rather than in a form you have to retype out of a book.
The mechanisms have their own page, and that matters more than it sounds. Chapter 12 puts numbers in your hands — six weeks, five hundred euros, ten working days — and says plainly that they are starting values with nothing behind them. A book cannot be corrected after it is published. A practice can. So the parameters live on the site with a version number against them, and they will move as people run them and tell me what happened.
Which is the one thing I would like in return. If you grant a window, tell me how it went, including if it went nowhere — especially then. I am not asking for that as a marketing device; it is the only mechanism by which any of this gets less wrong.
leapfrog.lerias.org is the other half, and it is free too. This book is the why and the when. That one is the how. If you have finished this and the problem you now have is that you need to actually ship something, start there.
This book is free as well, under a licence that lets you pass it on as long as you do not sell it. There is no list to join and there is nothing for sale, which is partly a principle and partly the terms under which I am allowed to write it at all. If you think somebody in your building should read it, send it to them. That is the whole of the distribution strategy.
One last thing
A note on how this was made, since a book about AI-assisted work ought to be clear about its own.
I wrote it on evenings and weekends over several months, on my own time and my own account. I used AI throughout — to assemble and cross-check research, to validate claims against their sources, to draft and redraft passages, to review chapters for the things I had stopped being able to see, and to argue ideas into shape before I committed to any of them.
The argument is mine. So is every decision about what this book claims, what it refuses to claim, and where the evidence stops. Nothing went in because a tool produced it; a good deal came out because I did not believe it.
What the technology changed was the scale of what one person could attempt in the hours available. That happens to be the subject of this book, and I have no interest in pretending I am standing outside it.
That’s it. Have fun.