Build, Don't Rent: I Stopped Asking What to Subscribe To

I pay one hundred dollars a month for Claude Max, and it is the reason my software spending is going down. On why my default question stopped being what to subscribe to and became whether I can just build it.

Preston Grisham ·

I pay one hundred dollars a month for a Claude Max account. It is the most I have ever paid for a single piece of software, and it is the reason my total software spending is going down.

Here is the math. The two biggest were Squarespace, where my personal website used to live, and SmugMug, which I had paid for years to host my photos. I canceled both and rebuilt the site from scratch on my own backend, so my photo library and this blog run right here now instead of on platforms that charged me every month for the privilege. The rest were smaller: a no-code database I was using to track blog ideas, a river conditions app, and a couple of subscriptions I am almost embarrassed to have been paying for. None of those cost much on their own. Most were under fifteen dollars a month, and that is the trap. Each one is cheap enough that canceling it feels like it is not worth the afternoon, so you keep all of them, and the total quietly becomes a real number.

For years my instinct when I wanted something was to go find the service that already did it and pay for it. River flow data before a kayak trip? There is an app. A place to capture blog ideas before they evaporate? There is a tool for that, and it syncs, and it has a nice mobile widget. Each decision was rational on its own. Added up, I was renting a dozen small pieces of my own life from a dozen different companies, each of which could change its pricing or shut down on a Tuesday.

Somewhere in there my default question changed. Now the first thing I ask is whether I can just build it myself.

RiverFlow is the clean example. I wanted to know if the water was up before I loaded the kayak. The USGS publishes the data for free. The apps that package it up wanted a subscription to show me a number that the government already gives away. So I built an app that reads the USGS feed and tells me what I want to know. It cost me a weekend and zero dollars a month. The blog ideas tool became a small dashboard I wrote myself. The photo gallery is now part of my own site, on my own backend, where nobody can deprecate the feature I depend on.

The thing that makes this possible is not that I suddenly became a better programmer. It is that the last ten percent of a side project, the part that always killed it, got cheap. Anyone who builds on the side knows the pattern. The fun part is the first ninety percent. The part that never gets done is the boring glue at the end: the error handling, the empty states, the settings screen nobody will ever open, the deploy script. Apps that sat at ninety percent for months are finished now, because that last stretch is exactly the kind of tedious, well-defined work that AI is good at and I am not patient enough for.

The irony is not lost on me that I am defending lower software spending by pointing at a one hundred dollar a month bill. This is not free. It is a real cost, and it only makes sense because of how much I actually build. If I shipped one app a year, a hundred dollars a month would be a terrible deal and I would tell you to go pay for the river app. The case for building over renting only works when you build enough that the one tool you keep replaces the many tools you would otherwise rent.

It also is not really about the money, even though the money is what makes it concrete. What changed is the sense of ownership. When I rent software, I am a tenant in someone else's product, subject to their roadmap and their pricing and their decision to sunset the thing I built my workflow around. When I build it, it is mine. It works the way I want, it does not phone home, and it does not send me an email in eighteen months announcing that my plan is being discontinued.

I am not telling everyone to cancel their subscriptions and learn to code. Most people should pay for the river app. But the default is worth questioning. The next time you reach for your card to rent a small piece of your own life, it is worth asking whether the thing you actually need is smaller than the product you are about to pay for. More often than it used to be, the answer is that you could just build it.