Whose computer are you building on?

I write code in more languages than I can read, and I ship it anyway.
I’m 35. I started with Python a few years ago, copying out of tutorials, then HTML and CSS, then fell hard for TypeScript, the one language I know best. The rest I write now with a lot of help, in languages I’m still learning, on a field that won’t hold still. Here’s what gets me: the tools get about a year better every few months, and I get a few weeks better. I’ll finally have some idea down cold, and that same week the tool is writing my feature in a language I’ve never opened. You can feel capable and obsolete in the same afternoon.
I don’t say that to complain. These tools are the only reason I’ve built anything that stuck, like the Zo Cookbook, a searchable app with 1,162 recipes that strangers actually use. The question I keep circling, usually past midnight, is who owns the thing I’m building all this on.
You’re renting the means of production
Strip the magic off these tools and they’re the means of production for the work I do. The factory floor, rented by the month, in a building I’ll never see. The intelligence I build with doesn’t live on my machine. It lives in a data center owned by people who answer to investors who need the number to climb forever. I don’t own any of it. I rent access, and the rent is built to find the ceiling of what I can pay.
An economist named Yanis Varoufakis calls where this goes technofeudalism: a handful of platforms own the cloud capital everyone has to route through, and we pay for it twice, once in money and again in the data and free work we hand over every time we use it, which trains the next version we’ll also rent. Closer to tenant farmers than customers. I’m not sure he’s right about all of it, but the word landlord shows up in my head every time I open one of these apps.
And landlords have habits. The first stretch is cheap, free if they can swing it, because the point is to get you moved in. Then your work grows roots, the rent climbs the way rent does, and leaving means tearing out everything you planted. Read the lease, too: most of these products keep the right to feed whatever you make into their next model. You’re the tenant and the raw material at once. That never makes the homepage.

So who should you build on?
Once you realize you’re renting your tools, you start seeing it everywhere. The one I keep coming back to is Zo Computer. It’s a personal cloud computer, your own server with an agent on top, and it’s where I do most of my building, including the Cookbook. I pay for it, sure. Paying for a good tool has never bothered me. What bothers me is a tool that keeps your work when you walk, and Zo doesn’t: the whole pitch is that what’s on your computer is yours, and it behaves that way. I can reach the people who make it, too, which is rare enough to matter.
I reach for Dyad for a simple reason: it runs on my own machine. My files stay mine, and the guy who makes it builds it mostly on his own. I can walk away and take my work with me.
This isn’t a purity test. Set these tools up right and they earn their keep. I get useful work out of them every day, the big ones included, and I’m glad they exist. The slop people complain about comes from lazy use, not from the tools being junk. Cost has never been my worry. Getting stuck is, and the deeper a tool reaches into my income and the work I can’t afford to lose, the more it matters whether I can pack up and leave.
Why you don’t hear this much
The reason almost nobody in my corner says this out loud is boring. The big companies are also the sponsorships, and the API half of everyone’s product is bolted onto. Calling the landlord a landlord goes badly when you need the landlord to like you. I’m not in that bind, because I’m not building a career on their goodwill. I’m trying to help the people standing where I stood, late to this and with no net under them, get usable work out of AI without quietly signing the deed over to whoever built the smoothest signup screen.
You don’t have to quit the big companies’ tools tomorrow. You just have to know the deal before you’re in too deep to leave it. Most people building right now never ask who owns the room they work in, and by the time it matters, the rent has already gone up. So the next time you lean on a tool for something you can’t afford to lose, find out who keeps what you make, and what leaving would cost you. Most of that answer is sitting in the pricing page and the fine print. The work is yours. The tools you build it on should act like it.
By Jeff Kazzee · Subscribe · X · LinkedIn · GitHub
Jeff Kazzee writes about AI, careers, and learning to think clearly. Subscribe on Substack for the next essay. Find him on X, LinkedIn, and GitHub.