§Six whitepapers stand behind this one. They teach a discipline in stages: what product is at a startup and who owns it, how to think about what to build, how to earn insight and turn it into evidence, how to commit to a direction, how to make it close as a business, and how to keep learning after launch. Each is a method, with templates, sequences and rules.

§This one is not a method. It is the reflection that closes the collection, and it has a single job: to stand at the end of all that discipline and ask what artificial intelligence changed about it. The answer is not the one most teams act on, and the gap between the two is the reason this essay exists.

§ 0.1The change everyone sees, and the response most get wrong#

§Everyone sees the same change. Building a working prototype, for twenty years the expensive step, is now nearly free. A person with no design or engineering training can turn a written brief into a navigable app in two days. The tools that do this, the ones that turn a prompt into a working interface, are among the fastest-growing products in the industry, and their existence is the premise of this whitepaper.

§The response most teams have to that change is to build more. If a prototype is free, build ten. If a feature costs one prompt, add it. This is the wrong response, and the collection has spent six whitepapers showing why in specific cases. Here is the general form: the cost of building was never the thing that made products good. It was, if anything, a discipline in disguise, a tax that forced teams to be sure before they built. Remove the tax and you do not get better products for free. You get more products, most of them answers to questions nobody asked.

§ 0.2The argument under all seven#

§There is one argument beneath the whole collection, and this essay is where it becomes visible, because you can now see all six stages at once. The argument is that product work is the discipline of deciding what to build under uncertainty, and that the discipline is a chain of ways to earn the right to build: a problem earned from research rather than assumed, insight earned from behavior rather than opinion, a direction earned by excluding alternatives, a business earned by naming what has to be true, a first version earned by a hypothesis that could be wrong.

§Every stage is a gate. Every gate exists to stop a team from building something it has not earned the right to build. And AI, by making building free, did not remove the gates. It removed the excuse for skipping them, which was that building was too expensive to do carefully. Ethan Mollick's framing of the technology as a tireless collaborator is right, and this essay's addition is that the collaborator is tireless at exactly the parts that were never the point. The parts that were the point are still yours.

§ 0.3Everyone starts at zero now#

§South Park Commons named the phase before a company exists "minus one to zero": the search for what is worth working on, before there is a product to build at all. AI has quietly made that phase the whole game. When anyone can build, everyone is already at zero, and the distance that used to separate teams, the cost of getting something built, has collapsed. What is left is the minus one: deciding what is worth building before you build it, and then getting the first version right. This collection is the product craft of that stretch, from the decision of what to build to the first thing that ships, in a world where building is no longer the constraint.

§That is also why this book is written for anyone who owns what gets built, a solo founder or someone on a team of five alike. The cost that used to make product a specialist's job is gone. The judgment that replaced it belongs to whoever decides.

§ 0.4What this essay does#

§It looks back. Chapter one describes the bottleneck as it was, and why every method you learned was shaped by it. Chapter two tells the day it moved, through one field sprint, in numbers. Chapter three reads the six whitepapers back as a single argument, stage by stage. Chapter four draws out the five principles that recur across all of them. Chapter five collects, honestly, what AI changed at each stage. Chapter six names the one genuinely new role, the owner as the person who maintains context rather than produces documents. Chapter seven names what did not change at all, points to where a team's time should go now, and closes.

§There are no templates here. The other six whitepapers have those. This one has the thing the templates were always for.