The hotel ceiling was white. The air conditioner hummed. Pieter Levels, not yet thirty, lay on a cheap bed in Chiang Mai and counted what was left.

His YouTube channel teaching music production once paid enough—a few thousand dollars a month, enough to drift through Asia. Then the algorithm shifted and the income collapsed to roughly five hundred. This was not a polished “low point” for a podcast. He had sold his Amsterdam apartment and studio gear, flown one-way to Asia, and landed here as a failure. Later, into a microphone, he said the line that still sounds like a joke and a bruise: he was twenty-seven, staring at the ceiling, calling himself a loser.

His father did not talk feelings. The man who spent years tearing walls out of the family house offered a rougher prescription—get sand, shovel it to the other side, do something physical, do not lie still. Pieter understood: the danger in depression is stillness.

He found himself a mountain of sand to shovel for a year.

Twelve months, twelve startups: shame into muscle

The mountain had a name. Around 2014 he set the challenge that later became famous—12 startups in 12 months: twelve months, twelve sites people could pay for, one due every month.

Not because he had too many ideas—because he feared quitting halfway again. He knew the creator disease: always one breath short of shipping. A hard deadline forced delivery. Some products somehow hit MTV or WIRED; friends forwarded links and he felt like a winner. At night, counting hotel bills with a quiet Stripe account, he felt like he had won nothing.

Fame humiliates gently: it lets you believe you are almost successful while rent still knocks.

Those months forced an unromantic habit: turn an idea into an ugly thing that can take a card in two weeks; a charge means real demand; most things fail, then you build the next. PHP that runs is enough. It sounds like a method. In Chiang Mai it was just a shovel.

He was the first user

The most important sentence in the whole story is short: he did not discover the digital-nomad market—he was that user.

The shocks after landing never appear on “freedom” posters. Wi‑Fi that cannot hold a call. Rent twice what you budgeted. Walking a strange street at night and realizing you never checked safety. Someone on a forum calls a city “amazing”; you fly in and find they meant parties while you needed a quiet corner to code. Notes scatter across Reddit, travel blogs, expired spreadsheets—Bangkok and Lisbon cannot sit side by side: whose net is faster, whose cost of living fits your wallet?

Pieter lived inside those shocks. The public city spreadsheet was not “finding a category.” It was opening his own notes so people in the same mess could edit them. Safety, café density, LGBTQ friendliness… others added columns until the table became a shared brain. On Product Hunt and Hacker News it drew tens of thousands of visitors in a day—in 2014 almost no one treated “where next” as a serious tool. That table was the fourth product in the challenge: Nomad List.

From there he kept walking the same loop: solve your own problem → ship in public → users show up → charge → improve for paying users. Nomad List. Remote OK. Years later, Photo AI.

Traffic arrived. A prettier free leaderboard might have stopped at “cute personal project.” Money depended on whether he saw what users actually lacked: often not another table.

What Nomad List really sells

City data later became free in many places—blogs, lists, even deliberate “Nomad List alternatives” comparing speed and prices without a membership. If you only glance once at which city is cheap, you need not pay. Critics are right about that.

What nomads often fear is not a missing number. It is landing and knowing no one.

You arrive with an Airbnb key. Days work. Nights? Who answers visa questions? Which café will let you sit all day? Fever—who knows a local clinic? Free groups drown in ads and “I want to be a nomad someday” posts. Public data can be copied. “Who is in Lisbon right now and willing to grab dinner tonight” cannot.

Pieter put the business on that layer: browsing cities mostly free as a funnel; payment unlocks people—chat, forums, trips, meetups in the same city. When free groups flooded with spam, he started with a tiny one-time fee; at five dollars people still joined, so he raised to twenty-five, fifty, sixty-five, then annual—around the Indie Hackers interview, membership sat near seventy-five dollars a year. Price rose; people still came. Not because the UI glittered—he admits the product is not trendy—but because people who have paid are more likely to be actually on the road, and answers get useful.

As people accumulate, the network gets more valuable. Trips, friends, city habits settle inside; switching platforms means being lonely again. Interfaces copy; years of member relationships do not. One person shipping fixes fast means “good enough” often beats “pretty but slow.”

Acquisition barely burned cash. Time and HuffPost-type outlets wrote about the odd site; links and search piled up. Facebook ads converted absurdly badly. Membership fees times thousands of people on the road, near-zero marginal cost—servers and his time.

Revenue needs dated labels. An early Indie Hackers piece framed Nomad List near about thirty thousand dollars a month; later on his own blog he described membership roughly in the twenty-to-forty-thousand-a-month band, around three hundred thousand a year—public self-reports, not audits, not the same spreadsheet on the same day. People who only want data call it expensive. People who find friends, work, and fewer traps in a strange city often find the fee cheaper than one wrong long lease. The product welcomes the second group. Pieter once was that group.

People still need to eat: routing traffic onto a job board

Nomads keep asking: where is the next remote job?

Again he productized a problem around him. First Nomad Jobs—cold emails to Buffer, Zapier, and others already hiring remote, begging for listings so the board would not look empty; roles nested into Nomad List so old traffic walked into a new counter. Later companies paid to post; then it grew into Remote OK. Individual memberships are tens or low hundreds; companies pay thicker sums to hire. He learned a colder truth: B2B often feeds you better than B2C.

The pandemic made remote mainstream overnight; the board rose with it. He admits luck—but the board was nailed down when remote was still unfashionable. In May 2019 his blog said Nomad List and Remote OK together cleared roughly eighty-three thousand dollars a month—about a million a year; four years, not overnight. Through it he still refused to become a manager of hires: he wanted to build with his hands and feared going soft. In product terms, small user requests often still landed in that Dutchman’s own commits.

Success here has body heat: not ARR on a chart, but no longer staring at a hotel ceiling counting five hundred dollars.

When a giant hits him, he is afraid too

In 2022, cheesy AI avatars could blow up overnight. His blog recorded about one hundred thousand dollars in AI avatars roughly ten days after AvatarAI.me launched; on the Lex Fridman podcast he looser recalled a peak that felt like “about a hundred fifty thousand in a week”—his own numbers, not a third-party audit. Then pleasure thinned: apps like Lensa, with users already installed, followed and pushed. In the same interview he guessed they made on the order of thirty million dollars—his spoken estimate, not a filing. On camera he admitted sadness: he rarely met an opponent that fought this well, and they did it well.

The fear is concrete: a one-person shop against a machine that can push notifications.

He did not “work harder” into a wall. Avatar felt too kitsch—turning yourself into Barbie; the hype would cool, and it did. He moved to something people already pay photographers for: respectable photos—LinkedIn, résumés, faces for formal rooms. Photo AI charged from day one, wired existing model APIs, and listened to paying users complain where faces looked wrong. Old followers still watched him ship on Twitter; cold start was not from zero—same loop: step in the hole yourself, ship public, improve for people who pay.

Giants win installed base and marketing. He wins by quitting fast enough, plus a decade of “I’ll fix it for you” trust. Sometimes the light in a person is not never failing—it is admitting, after failure, “this is not my battlefield,” then planting the shovel in the next patch of sand.

After a decade

He can lie flat. He can take his father’s sand. He can be the unlucky nomad before the product. He can use a five-dollar gate against spam. He can beg companies for job posts. He can get hit and refuse a doomed fight. He would rather write PHP alone than let his hands go dull.

Nomad List gets paid because real people on the road fear choosing the wrong city—and fear landing alone even more. Remote OK and Photo AI are the same hands across different decades: solve your own problem first, then let paying users say what still hurts.

The ceiling is still that ceiling. He just no longer has to count those five hundred dollars.