Article
Prague, 13 May 2026. On my way to work I started thinking about something that stayed with me for days. If most routine work on a computer disappears in the next ten years, and a large share of repetitive manual work disappears with it, what happens to the flow of money?
Who pays whom for what? Which economic layers will exist, how large will they be, and what relationships will run between them?
The Question
Around thirty per cent of those who do routine work on a computer lose their position in the medium term. Around seventy per cent in the long term. Across the whole workforce — including jobs that cannot be done on a computer — around fifteen per cent in the medium term, and around thirty per cent in the long term, lose their original job permanently. I have written about that shift before in Will AI Take My Job?.
Add to that a large share of repetitive manual work taken over by robotisation. The wage economy as we know it does not survive these numbers. Money has to keep flowing somewhere — earned, spent, taxed, redistributed. Take wage income away from half the population and the cycle breaks at several points at once. The question is what replaces it.
Why the Old Answers Fail
The first reply is that universal basic income will solve it. The problem is that UBI requires the state to tax someone, and the people who would have to be taxed are mostly the same people who write the tax laws. Politicians, lawmakers, central bankers — most of them sit in the top five to ten per cent of capital ownership themselves. Asking them to tax their own capital is asking the carp to drain the pond. France tried a wealth tax. Capital left within years. Corporate tax rates in the EU have fallen for three decades. The largest tax havens in Europe — Ireland, Luxembourg, Cyprus — are EU member states. The OECD global minimum corporate tax sits at fifteen per cent, far below what would be needed to redistribute much.
The second reply is that the market will adjust. But the market can shrink to a hundred million customers and still produce more revenue than today. Apple, Tesla, LVMH, Microsoft can survive serving the top one per cent with custom products and recurring services priced ten or twenty times higher than current consumer models. If a billionaire pays a million dollars a year for a personal security system with thirty robots, a thousand such customers produce the same revenue as a million iPhone buyers — at far higher margins. The middle market is not protected by inevitability. It is a current arrangement that can be abandoned.
The third reply is that pressure from the street will force change. This was true for the last three hundred years of political economy. It is no longer true. With drones, robotic policing, and AI surveillance, the cost of holding power against an unarmed population drops sharply. Capital and political elites no longer need consent to maintain order. They need the technology, and they have it.
The Brakes That Still Hold
None of this means the worst possible outcome is the only outcome. Some brakes still hold.
Technological dependence. Drones, AI, robotic systems all rely on physical infrastructure. Taiwan Semiconductor produces around ninety per cent of the world's advanced chips and depends on twenty-three million Taiwanese people. Power plants need operators. Rare earth materials come from China, Australia, and Myanmar. Datacentres need maintenance crews, engineers, and supply chains. If the upper class discarded everyone else, they would break the systems that hold their own position.
Elites are not united. There are several elites — a US elite, a Chinese elite, an EU elite, a Russian elite, a Gulf elite — each with conflicting interests. Geopolitical competition forces every bloc to keep its population at least loyal on paper — for soldiers, scientists, voters, consumers, and population numbers. No bloc can afford a fully discarded population while the others have not done the same.
Demographic decline. China is at a fertility rate around one. The European Union sits at one and a half. South Korea is below one. The real demographic problem within twenty years is not surplus population but shortage. The fight is more likely to be over how to retain and produce more humans, not how to discard them.
These brakes will not produce utopia. But they will keep the worst version from happening completely.
The Likely Outcome — Gated Luxury Capitalism
The most likely shape of the economy by the mid-2030s is what I would call gated luxury capitalism. A narrow capital class — perhaps one per cent of the world population — owns the AI, infrastructure, land, brands, and distribution channels. They live in protected enclaves with private healthcare, private education, and robotic security. Around them sits a service caste of half a billion to a billion people who keep the system running. Beneath that is a much larger subsistence population supported by some form of minimum income. And outside the formal money flow is an enormous discarded population in regions where the state has failed or has never functioned.
This is not a forecast in the strict sense. It is the direction the current arrangement points to if no major political force changes it. And no such force is currently visible.
Six Layers — The New Map
The clearest way to map it is in six layers. These are approximate ranges that overlap and shift at the margins.
Capital. About one per cent of the global population, roughly eighty million people. Owners of AI, infrastructure, real estate, brands, distribution. Geographically mobile. Concentrated in a few dozen enclaves worldwide.
Tech Service. Four to five per cent, around three to four hundred million. Engineers, technicians, datacentre staff, energy operators, biotech researchers, top-tier financial professionals. Located near critical infrastructure.
Local Service. Six to eight per cent, around five to seven hundred million. Hairdressers, therapists, carers, teachers, cooks, couriers, local doctors, physiotherapists, tradesmen. The most stable layer, because what they do cannot be automated within the next decade. They exist wherever there are people to serve.
Subsistence on UBI. Twenty-five to thirty-five per cent, two to three billion. In functioning states. Supported by basic income and a healthcare floor.
Join the Library
Full access to my thoughts, personal stories, findings, and what I learn from the people I meet.
Join the Library — €29.99 per yearSummary
Common questions on this article's topic
What happens to the economy when AI replaces most jobs?
Will universal basic income solve AI job loss?
Where will the wealthy live in the AI economy?
What are the six layers of the AI economy?
What is the hybrid strategy for surviving the AI economy?
Can pressure from the public stop this?
More articles
I am building an AI system to predict the S&P 500. It runs on my own machine, uses free public data — yfinance, FRED, the Shiller dataset — and grades every forecast against reality. This series documents the build itself: the decisions, the methodology, the mistakes. What I will eventually share from the running system is a separate question, and an honest one.
Yesterday I could not tear myself away from the computer. When I lifted my head, it was half past eight in the evening. I had been sitting alone upstairs for about three hours.
Will AI take my job? A certified Google trainer told me in June 2024 that my profession would cease to exist. Twenty-two months later, my job title has not changed — but ninety percent of what I do during the day is different. I have delegated more of my thinking to AI agents than I thought possible. I am not afraid. This is why, and what it means for anyone asking the same question.
One hour. Fifty-five minutes. That is how long it took to build what a Czech software firm had quoted at over €50,000. I built it with Claude Code. Not a prototype. Not a proof of concept. A working tool — the one the company actually needed. By the evening of the same day, it was running on staging. This is not about Claude Code. It is about what Claude Code exposes.
I have conducted roughly one hundred and fifty practical interviews over the past four years. Fifty for data specialist roles. A hundred for advertising and performance marketing specialists. Almost every one of them involved sitting down with a candidate over a practical task — something close to a real problem we actually need to solve at the company. Not theory. Not trivia. Applied problem-solving. Over time, I started noticing a pattern.
Before you can teach AI to understand anything, you need to see what it is hiding from you.
The moment other people needed access to it, the problem changed completely. It was no longer about whether the agent could learn. It was about who gets to teach it.
I wanted to build an agent that doesn't just assist. One that acts.
This is what I learned about local vs cloud AI, and why I switched to Claude Code.
What happened — and how can it be reversed?
Four days in Catalonia. No computer, no AI, almost no social media. I bought this notebook so that I could write down what I would think about, and what I would come across and learn on the trip.
