The starting point was content, not code
Before Pieter Levels was labeled a model independent developer, he was a content creator. In the Netherlands he ran an electronic-music show called Panda Mix Show, selecting music, editing it, and maintaining an audience—none of which had much to do with writing code.
That experience later became his biggest difference from other independent developers: he instinctively understood that distribution is harder than the product. Most technically trained people finish the product and then worry about finding users; his order was the reverse.
12 months, 12 products
In 2014 he sold his possessions, left the Netherlands, and went to Asia. On the surface this was a choice about the digital-nomad lifestyle; in practice it was a restructuring of his cost base—when monthly expenses are pushed very low, the cost of failure collapses with them.
That September he publicly launched “12 startups in 12 months”: build and publicly release one product every month, regardless of how finished it was.
The clever part was not diligence but using a public commitment to put himself under pressure. Once “one release every month” was written in public, perfectionism had nowhere to hide—it forced him to cut the time budget for polishing to zero.
The result followed the laws of probability: most projects died. But one survived.
Nomad List: starting with a spreadsheet
Nomad List was not originally a product at all, but a public Google spreadsheet he used himself. It recorded internet speed, prices, climate, and safety in different cities so he could decide where to go next.
He posted the sheet online and unexpectedly attracted a great deal of attention. The demand was validated before the product was built, rather than the other way around.
The real lesson is not “build a city database,” but that his most successful product began as an ugly, minimal implementation that solved a real pain point for himself.
Remote OK extended the same logic—Nomad List users kept asking, “Where can I find remote work?” so he casually built a job board. This spin-off, grown out of conversations with users, later became his most stable long-term cash flow.
A technology stack conservative enough to look heretical
His technology choices have long provoked engineers: PHP, jQuery, one server, no modern frontend framework, no microservices.
This is not about technical ability; it is about a different objective function. A team optimizes for maintainability and collaboration efficiency, while a solo developer optimizes for the time from idea to launch. When there is only one person, “all the code in one file” can be an advantage—no communication cost and no cognitive burden from abstraction layers.
The conclusion is not “you should use PHP too,” but: your technology choices should serve your organizational form, not your résumé.
Repeated validation in the AI era
When diffusion models took off in 2023, he launched the AI portrait product Photo AI within weeks, and its revenue quickly surpassed that of products he had operated for nearly a decade.
This validated the core of his method: he never aims to lead technologically; he aims to ship quickly. The technical barrier for AI portraits was not high, and thousands of people had the ability to build something similar. But he was among the first to package it as a paid product and put it in front of an audience—the audience he had accumulated over the previous decade.
A hit is not the same as a durable business
Starting on February 22, 2025, he publicly built the browser flight simulator Fly.pieter.com and said the product reached a $1M ARR run rate after 17 days. That number shows strong distribution and immediate willingness to pay, but it does not mean it had earned $1 million continuously for a year. In Pieter’s own project list, Fly was later classified as “made money but was not sustainable,” rather than a long-term success. Fly build log · Complete project list
This is why this page goes deep only on Nomads.com, Remote OK, and Photo AI: they represent a membership community, a B2B recruiting marketplace, and a high-compute subscription product, respectively, and their business loops have been tested over a longer period. Hotelist, Infinite Slop, and Pieter.com remain representative works, but “widely shared” cannot be written directly as “sustainable revenue.” Hotelist actually launched in October 2024; Pieter.com’s retro-computer experiment also began in 2024 and is a non-commercial project. Hotelist project record · Pieter.com build log
Infinite Slop launched on August 29, 2026, letting viewers decide through chat what the next AI video should play. The next day Pieter publicly said that 37,000 people had watched it; what can currently be verified is the view count and product mechanism. No original revenue disclosure corresponding to “about $15,000/month” has been found, so that amount is not recorded as a verified result. Infinite Slop view disclosure