Software is getting cheap. Not just cheaper—the cost curve is collapsing in a way that changes what’s possible for a single person to do.

For most of my career, the bottleneck was always code. Could you build the thing? Did you have the engineers? Could you ship fast enough? That bottleneck is dissolving. What’s left is everything else: figuring out what to build, validating whether anyone wants it, operating it once it exists, selling it, supporting it. The hard part isn’t making software anymore. The hard part is running a business.

So I’m building a system to run businesses. I’m calling it Inc.

The idea is simple enough: one system that lets me discover, validate, build, deploy, operate, market, and sell businesses with as little human effort as possible. Not zero human effort—I’m not delusional—but minimal. I want to find the floor.

I’m planning to use Inc to launch twelve businesses. That number is a bit arbitrary, but it’s a commitment to iteration. I’m not trying to find the one perfect idea and sprint at it. I’m trying to build a machine that gets better at finding and running ideas, and twelve gives me enough cycles to actually learn something.

Here’s the constraint I’ve set myself: no employees. Or at least, that’s the starting position. If something genuinely requires human relationships I don’t have, I’ll reconsider. But the default is automation, not delegation. Hiring is the easy answer to most scaling problems, and I want to see what happens when you take it off the table.

This isn’t a VC play. I’m not raising money. I’m funding this myself, which means the businesses need to be small enough to validate cheaply and profitable enough to sustain themselves. I’m not building unicorns. I’m building precise little machines that solve specific problems and generate enough value to pay for my life.

If one of them breaks out—if something really works and the right move is to raise money and sprint—then fine, I’ll forget everything I just said and go do that. But that’s not the plan. The plan is to stay small, stay flexible, and learn what an effective company actually looks like when most of the work is done by agents instead of people.

I genuinely don’t know what that looks like. Current startup patterns are a good starting point because they’ve been pressure-tested, but they’re patterns designed for humans. Teams, roles, standups, accountability structures—all of it assumes people. As agents get more capable, I suspect a lot of that becomes vestigial. Maybe all of it. I want to find out.

Inc is the system that lets me run that experiment. It might become a product eventually. It might stay an internal tool forever. I’m not deciding yet because I don’t need to.

What I do know is that I want to build businesses without burning out, without managing people, and without pretending I have more certainty than I do. Inc is how I’m going to try.