You open your laptop before dawn and check the sign-up notifications that came in overnight. Six months ago you couldn't write a single line of code; now, with AI tools, you can spin up a small web service in a matter of days. Adding a new feature used to mean weeks with a freelance developer — now it's done in a day. Your hands have gotten faster. And yet, three months in, the number in your bank account has barely moved. You're working longer hours than you ever did at a company.
This is now one of the most common scenes in solo entrepreneurship. Better tools should mean a better life, but the extra time those faster hands create tends to get poured right back into shipping more features and taking on more client work. Is building faster and building more really the answer? That question is where today's story begins.
People Who Use AI Fall Into Three Stages
Anyone who starts using AI ends up standing at one of three stages.
Stage one is the person who asks AI questions and gets answers — using a chatbot instead of a search engine to polish writing or organize research. Stage two goes a step further: using AI to build your own tools. Even without ever learning to code, you automate repetitive tasks and launch a small service on your own. Your hands get faster, and alone, you accomplish what used to take a whole team. Stage three is making money from those tools — running a company of one.
Boil the difference between the three stages down to one word each, and stage two is productivity, stage three is profitability. AI has thrown open the doors to stages one and two for everyone. The problem is that many people arrive at stage two and stop there. Output that's faster and more plentiful feels good enough on its own that you don't even notice you've stalled. The faster hands and the unmoving bank balance from the opening scene are exactly what that stall looks like.
Discussions of automation governance draw an old distinction between people who are inside a loop and people who are outside it. For someone inside the loop, AI is just a faster pair of hands — they're still selling their own time. Someone outside the loop designs the very structure the AI runs inside, deciding one level up what actually generates the money. That's the line between stage two and stage three. Using a tool well and designing the value the tool creates are two entirely different things.
What Stalls You Isn't the Ability to Build — It's Judgment
It's tempting to blame the tools for failing to move from stage two to stage three. But the tools are already good enough. What's actually blocking you isn't the ability to build — it's judgment.
What AI has collapsed is the cost of building. The cost of deciding what to build, who to sell it to, what to charge, and when to stop hasn't moved an inch. In fact, the cheaper building gets, the more valuable the judgment of whether to build at all becomes. Once the same tools spread to everyone, the thing that separates outcomes is no longer how well you build, but what you chose to build in the first place.
One distinction helps here. Work is the activity of turning time into output; management is the activity of turning that output into an asset. Someone stalled at stage two has done a lot of work, but hasn't yet built the structure that lets output accumulate as an asset. You write an automation prompt to save time, post your building process to gather subscribers, and take on client work to get paid. These are three entirely different kinds of holdings, yet they usually get lumped together under the same word: "work." Saving time, getting paid, and building trust are valued by different formulas and grown by different methods. Treat them as one and the same, and you almost always end up in the same place: the time-saving pile keeps growing while revenue hits a ceiling.
Push this distinction further and you run into an old blind spot in accounting. International accounting standards bar companies from recording internally generated brands, customer lists, or know-how as assets, on the grounds that their cost can't be measured reliably. For a manufacturer whose balance sheet is filled with factories and inventory, that's a minor omission. But for a company of one, nearly everything of value is exactly that kind of self-generated intangible — your name, your subscriber list, the data you've accumulated, the workflows you've refined. Looked at through the financial statements alone, a solo business always appears empty-handed. That's why anyone running a company alone has to build a second ledger by hand, separate from the financial statements, to track their real assets. This blank space that accounting leaves open is filled precisely by strategy's moat theory and by the capital concepts used in corporate reporting. Fields that have never cited one another arrive at the same conclusion here — a convergence you can't see if you're only looking through the lens of a single discipline.
That's also why conventional management theory tends to fall short here. The division of knowledge into accounting, strategy, marketing, and finance suits large organizations with plenty of people and capital, but it doesn't fit someone working alone. A single question like "how much should I charge for this tool?" involves cost, market, and value all at once — and someone trained in one discipline gets stuck figuring out which drawer to open first. In the industrial era, capital captured the efficiency gains; today, it's whoever designs how AI operates that captures the business. Economist R. Coase's century-old question, "why do firms exist?", comes down to one person's desk as a much narrower one: "what do I do, what does AI do, and what gets outsourced?"
Three Questions to Find Out Which Stage You're In
Whether you're stuck at stage two becomes obvious the moment you write down three things. One sheet of paper is enough.
First, list out everything you built over the past month and label each one: did it save you time, did it build trust, or did it directly earn you money? If nothing on the list directly earned money, the reason revenue has stalled isn't a lack of effort — it's the absence of a revenue-generating asset. You've only been accumulating time-saving holdings.
Second, pick one product you're currently getting paid for and ask three questions: Is there a tool the customer uses to make or get something? Is there a place where the transaction and payment actually happen? Can the value the customer received be captured in a number? Of the three, the one most likely to be empty is the last one — measurement. Because it happens on the customer's side, it rarely fills itself in unless you deliberately design for it. Leave that box empty, and even the same skill set gets trapped under an hourly-rate ceiling. The tool and the market are visible to whoever is building; measurement only exists if you consciously attach it.
Third, divide your bank balance by your minimum monthly survival cost. That number is how many months you can hold out. If it's short, cutting fixed costs comes before trying anything new. If your income is volatile, calculate this number based on your worst month, not your average one.
Once you have all three answers, the one thing to fix next becomes obvious. It's better not to try to fix everything at once. When you're working alone, your time runs in a single stream — take on two things at once and both slow down to the pace of your slowest thinking. Deciding what not to do comes before deciding what to do.
From Productivity to Profitability
What AI has opened up is an era when anyone can build — not an era when anyone can earn. The bridge between stage two and stage three isn't a better tool; it's the judgment to name what you've built, calculate its ceiling, and decide what not to do. Stage three begins when you step back from building faster and more, and instead ask what to build with those faster hands. That shift in judgment doesn't just reshape the business. Change how you decide what to build, what to let go of, and where to spend your time, and your whole approach to work — and the direction of your life — shifts with it.
This series crosses that bridge one piece at a time. It starts from a single manuscript that reassembles standard theory from accounting, economics, management, and investing — without regard for disciplinary boundaries — around the problems of running a business of one. The next installment looks at why a solo company's most valuable asset gets recorded as zero on the books, and how to fill in that blank yourself. If you've gotten comfortable building but stalled out on earning, the real story starts with the next piece.
Appendix: Key Concepts
- Transaction cost theory — Introduced by economist R. Coase in his 1937 paper "The Nature of the Firm." It refers to the cost of finding, negotiating with, and monitoring trading partners in the market; when that cost exceeds the cost of handling something internally, a firm brings the work in-house. As AI lowers the cost of internal processing, the same logic now applies to how one person allocates their own work.
- Intangible asset recognition rules — International accounting standards (IAS 38) prohibit recognizing internally generated brands, customer lists, and similar items as assets. This is why a solo company's core holdings show up as zero on the books.
- The three-stage framework and in-the-loop / out-of-the-loop — The opening framework of the book Running a Company of One, which divides AI users into three stages — user, builder, and operator — and connects them to the concept of human positioning in automation governance.
