Everyone keeps saying AI will turn us all into entrepreneurs. I don't buy it. I think it'll turn a lot of us into builders, which isn't the same thing.
Building has got cheap
A few years ago a small business needed a developer, a designer, someone on support and someone doing ads. Now one person with a laptop and a few AI tools can do most of that over a weekend.
I've done it myself. Since last November I've built four AI products on my own, mostly with Claude Code. One of them, Miru, is a WhatsApp assistant that helps with everyday UK stuff like train times, fuel prices and sorting school messages. Building it was the easy bit. Getting 20 people to sign up took a couple of months, and getting someone to pay is still the bit I'm working on.
And I'm clearly not the only one. By last year, over a third of new US startups had a single founder, up from under a quarter in 2019, according to Carta's Solo Founders Report.
Finding customers hasn't
So building isn't really the hard part anymore. Getting people to pay for what you've built is, and then doing it again and again. That's sales and marketing. Most of us would rather avoid it, and AI hasn't changed that much.
The App Store shows it well. Around 560,000 new apps went live in the first half of this year, almost as many as in the whole of 2025, but downloads only went up 2% (Sensor Tower data via 9to5Mac). Loads more apps chasing the same people. One developer the New York Times spoke to got 1,300 downloads and made about $500 from it.
If you can build something in a week, so can fifty other people.
People also say we're just going back to how things used to be, when most people worked for themselves. That's true, but a lot of them didn't have much choice, and plenty were pushed out by bigger players.
The 9 to 5 isn't the safe option either
None of this means the 9 to 5 is the safe option. It feels safe because the salary turns up every month and someone else worries about finding customers.
But in the first five months of this year, US employers blamed AI for nearly 88,000 job cuts, more than 2024 and 2025 combined (Challenger, Gray & Christmas, via Outlook Business). Those are the reasons companies gave, and some of them may be using AI as cover for cuts they'd have made anyway. Either way, a job can go with one email.
Both paths carry risk. It just shows up differently.
Two questions before you go solo
If you're thinking of going it alone, I'd worry less about whether you can build it and more about these two:
- Who already trusts you enough to buy from you?
- What do you know about the problem that a stranger with the same tools doesn't?
These aren't only questions for founders. Most of us in jobs have never had to find our own customer, because the company does that for us. They're worth asking even if you never leave.
If you've got good answers to those, there's never been a better time. If you haven't, a steady job is still a perfectly sensible choice. Just don't assume it's guaranteed.
I'm building small AI products under Human Agency. More of my work is at mekalav.com.