9. Ship Small

Shipping is how you learn what you actually built.

Pieter Levels runs Nomad List and Remote OK, two products he built mostly alone. Remote OK was famously a single PHP file called index.php, no frameworks, no fancy stack, almost no libraries, that grew into a business doing more than sixty-five thousand dollars a month. Levels has built dozens of small things since, sold some, killed some. In 2025 he made eighty-seven thousand dollars in a month from a flight simulator he built mostly by vibe coding. He is also the reason I left London in my twenties, but that’s a longer story.

The lesson of the index.php move is not the file. The file is theatre. The lesson is the constraint the file imposes. One file forces hard decisions: what is the promise, who is it for, what do they do next, and how do they pay. There is nowhere to hide. There is no roadmap to point at. There is just a page that either does something or doesn’t.

Easy Mode picks the constraint over the roadmap. Ship the smallest version that can be used by a stranger and paid for by a customer. One page, one promise, one button, one way to pay. Everything else is a feature you have decided in advance is more important than learning whether anyone wants the product at all.

The default failure mode is the opposite. People build infrastructure to avoid shipping. Auth systems. Multi-tenant databases. Admin panels. Email pipelines. A second product nobody asked for. All of it feels like progress because all of it’s the kind of work you can do alone without anyone judging the output. Then the day arrives when there is no choice left but to put the thing in front of a person, and you have not built the thing. You have built tools instead of a product.

I have done my best small ships when I had no other option. The first banner I shipped at sixteen was small because that was all I knew how to make. The first edition of this book was small because I sat down to write it in a few weeks instead of a few years. Draftpen is a single-purpose social media asset maker because I needed one and I built the one. Bordful is an open-source job board because I needed one and nobody owned one I liked. Pynions I called naive in public because it was, and most of what I learned from building it has turned into work I can charge for since.

Small doesn’t mean small forever. Small means you find out fast. Small means the next version is informed by reality instead of fantasy. Small means the cost of being wrong is a week, not a year. Most people are not afraid of failing. They are afraid of failing publicly at scale. Ship small enough that the failure mode is private and cheap, and the upside is that you actually learn something.

A useful shape for “small” in 2026: a landing page with a payment link, manually deliver the thing, see if anyone wants a second one. A one-feature app that does one thing well. A short paid PDF. A spreadsheet that you sell for thirty dollars. A free demo behind an email gate. None of these require infrastructure. All of them require choosing what the promise is.

If you are stuck choosing what to ship: define a ship that takes a single afternoon, ship it this week, and send the link to five people who might care. If you can’t identify five, ship anyway. The link being live is the unblock.

What are you building that nobody can buy, use, or share yet?