From Zero Code to a One-Person Company

Scroll through short video apps for ten minutes and you feel it before you can name it: AI is everywhere now, and underneath the excitement there is a low, steady fear. Programmers are being laid off in waves. Data processing, contract review, legal drafting — work that used to take a trained person a week now takes a model a few minutes, and it does not get tired or ask for a raise. I am not going to pretend this is not happening. It is happening, and I watch it happen every week.

Here is the strange part, though. The same force closing doors behind other people opened one in front of me. A few months ago I could not write a single line of code. Today I build websites for the factory, ship tools my colleagues actually open every day, and take on outside clients who pay for them. I am still not a programmer. I just stopped waiting for permission, and started learning AI instead.

What I actually built

The list is short and unglamorous, and that is the point. A website for the workshop. Internal tools that replace the copy-paste work nobody wanted to do. Pages that bring in customers. None of it is revolutionary in isolation. What is revolutionary is that I could do any of it at all — a person who, not long ago, thought "software" simply meant hiring someone else.

People assume AI in business means images and copy. That is the floor, not the ceiling. Yes, I use it every day for the most basic e-commerce work: product images, listing copy, short video scripts. That part is table stakes now — everyone can do it, which means it is worth less than it was last year. But in the same months I also shipped a WeChat mini-program, a mobile app, and a desktop tool. Three years ago each of those would have needed a team, a budget and a good deal of luck. I needed a laptop and a willingness to be bad at it for a while.

The eighty percent

The clearest number I have is about invoicing. Ours used to be manual: the same fields typed again and again, tax numbers cross-checked by eye, one person losing an entire afternoon to forty invoices. I built a tool with AI that reads the order sheet, matches the tax ID, checks the classification code and fills the declaration template. It now takes roughly a fifth of the time.

That is not a productivity statistic. It is an afternoon turning into twenty minutes, every single time. And the person who used to spend that afternoon is now doing something a machine still cannot do: talking to customers, solving the awkward cases, deciding what we should sell next. The tool did not take their job. It took the part of their job that was quietly eating them.

The OPC bet

Which is how I ended up inside the OPC wave — the one-person company. The structure is not new, but AI finally makes it real. One person with taste, judgment and a clear idea can now operate at a scale that used to require a small company. Not because they suddenly work faster, but because the boring eighty percent is handled, and what is left is deciding what to build and whether it is any good.

That remaining part is still mine, and it is still stubbornly human. That is what pulled me in: not the promise of doing less work, but the discovery that I could finally build the things I had been carrying around in my head for years. AI let me turn ideas into products without asking anyone's permission first.

Two futures, held at the same time

I am optimistic about what this does for civilization and uneasy about what it does to individuals, and I have stopped pretending those two are contradictory. Both are true at once.

If machines take over most of the work, then for the first time the link between survival and labor breaks. Scarcity stops being the default condition. Some people describe this as a path toward something like communism — not the political slogan, but the plain version: basic needs met, nobody starving, work becoming optional. On a civilizational scale that is a leap upward rather than a decline. We would finally have the time to climb toward a higher rung.

Carbon-based robots

But every tool cuts both ways, and this one cuts at something fragile. If food, shelter and safety are guaranteed and nothing is asked of you, a certain kind of person will simply stop. Not cruelly, not dramatically — they will eat, scroll, play, sleep, and repeat. No ambition, because nothing is scarce enough to fight for. No reason to talk to anyone, because nothing requires it. And human emotion is not a permanent fixture; it is a muscle. Stop using it and it quietly atrophies.

I have started calling them carbon-based robots. Biologically human, running on the same chemistry as the rest of us, but with the drive, the curiosity and the social instinct switched off. They eat, they breathe, they consume content. What they no longer do is want anything.

Where I land

I do not have a solution to a society-level problem, and I am suspicious of anyone who claims they do. What I have is narrow and personal. The invoicing tool did not make me lazier; it made me want to build the next thing. The mini-program was not the end of my ambition, it was proof that I could afford to have more of it.

So I try to use the freed time rather than just enjoy it. Build something. Keep the wanting alive on purpose. Keep talking to people even when it is inconvenient, because that is the part no model can do on my behalf.

A few months ago I could not code. Today I run a one-person company, and the honest summary is this: AI will very likely replace a great deal of human work. Whether it also replaces human drive is still, at least for now, up to us.

How this all started — an old Android phone and no plan — is written up in From Old Phone to AI Platform. The factory side of the story is in A Bag Factory's AI Transformation.