Your SaaS Is Tomorrow's Free Plugin

I know a founder in Berlin. Thirty-one years old. Two million euros raised from three venture funds. Eight people on the payroll, engineers and designers and product managers, the right people, the people you are told to hire. Eighteen months of brutal execution on a compliance tool for fintech companies. That product was hard to build. That was the point. Compliance logic, audit trails, the regulatory edge cases nobody wants to read. It took deep domain knowledge and hundreds of thousands of euros and a lot of nights. Then somebody sent her a link. An open-source implementation of most of her product, rebuilt in forty-eight hours with an AI assistant, documented better than what her own team had shipped, free to fork. She was not angry. She sat in a Berlin apartment at six in the morning on her second coffee with the specific expression of somebody watching the floor move. Eighteen months of work did not become worthless because it was bad. It became worthless because it stopped being scarce.

Here is the pattern we were all taught to trust. New technology destroys jobs and creates industries, and the industries are bigger than the jobs. The printing press made publishers. The car made highways, refineries, garages, motels, suburbs. The internet made e-commerce and cloud and a labour market that did not exist in 1990. Every one of those revolutions made replication cheaper. None of them made it free. AI is not opening a door in the same way. AI is walking into the room you are already renting and turning the price of your product into zero. Writing software was the last rented scarcity of the last forty years, and the rent was high precisely because replication was expensive. That is the entire venture capital thesis of the 1995 to 2025 period. Find something hard. Build it. Raise money. Defend the moat. Exit. Somebody with a 20 dollar monthly subscription can now sit in a cafe and rebuild your two million euro fundraise. If your product does anything that can be expressed in code, it is becoming a commodity. Not eventually. This year.

The economics tighten from four directions at once, and they arrived together. The first is the cost of the cities where startups happen. An apartment in the good part of Berlin that cost 600 euros a month in 2010 is 2,000 today. London, Toronto, San Francisco ran the same race, five to six times the 2010 rent for the same rooms. A salary that looked like success in 2015 now covers rent and not much else. The second is the commoditisation of the white-collar skills that were supposed to be safe. Legal research, financial analysis, coding, design, marketing, consulting. All of it was priced on scarcity and all of it is being repriced. The third is the price of money. A decade of near-zero rates let people build companies on thin margins, because capital was free. That era is closed, and with it a funding environment that assumed cheap money forever. The fourth is that the four vectors reinforce each other. Expensive cities push salaries up. AI pushes the value of that salary down. Dear capital removes the bridge. Nobody designed this. It just arrived at once, the way the 2008 crisis arrived on a Christmas Day when I was signing a 20 million euro personal guarantee on a business that was already gone.

I watched this land on somebody in my own office. Berkeley graduate, then a top consulting firm, the exact path the feeder schools advertise. She came to work with us and spent a fortnight watching AI agents do the analysis her firm charges 300,000 dollars for, faster and with fewer typos. She said the quiet part out loud: consulting is finished, the firms are waiting to die, you can build the whole capability in house with three people and a laptop. She was not being dramatic. She was reading a price list. Specialisation is exactly what the models are best at. Taking a narrow body of expert knowledge and making it cheap. Which means the most specialised person in the room is now the most replaceable person in the room, and the most valuable person is the one who can stand between domains and put the pieces together. You cannot get a degree in that. You get it by building things for years and losing some of them.

Thirty years of moving the same way. In the 1990s I was in New York while the internet was minting its first fortunes, and I did not take a job at a bank. Around 2002 I could see finance becoming the centre of gravity and I built a route into those rooms. After 2008 I moved east while the growth did. In every case I was not predicting. I was reading the direction of the river and swimming downstream. The river runs toward cheap cities, remote work and builders who know how to drive these tools. It does not run toward the traditional venture path anymore. Now the number that matters. The probability that a 24-month software roadmap built by eight people gets rebuilt by three people with agents before your Series B closes is not small. I would price it above half. Most founders I meet are pricing it at zero, which is why they are still hiring for the roadmap instead of re-scoping it. I have built companies across twelve countries and restructured a EUR 75 million industrial group. The work was never the software. The work was the translation between rooms, and that is the part that still has a price.

The old deal was simple. Go to work, collect the salary, consume in your own time, and the system carries you. That deal is being repriced and nobody sent a memo. There are two positions available now. Consumer of what machines produce, which is a slow liquidation with a pleasant interface. Or creator, which means deciding what gets built, in what order, for whom, and taking the risk that the answer is wrong. Every tool you need is nearly free. The judgement is not, and it is not for sale. The machine does the commodity work. You decide what the commodity is for. That is the whole gap between the two positions, and it is a choice you make in about a fortnight, or it gets made for you.