Some of the most useful software I own has exactly one user.

Me.

Lately, I’ve been building little applications with artificial intelligence to solve particular problems unique to me. It’s an expansion on the automation I’ve done my whole life, but now "build an app" is a legitimate tool next to "build a script." The more of these I make, the more I appreciate the possibilities of software built for an audience of one.

For example, I’ve built an application that turns my Apple TV into a custom status board. It shows my calendar, local weather, camera views, and other information I want to see. I get to decide what belongs on that screen because I’m the person looking at it.

Yesterday I got thinking it would be nice for my status board app to have a linear calendar (where each month gets its own row) for the year showing major events. After three prompts I had it working. But it was showing the calendar year. I wanted it to start with the current month and show the next 12 months. That took another prompt.

I’ve also built an application that drives my Vestaboard with custom postings. On my Mac, I’ve made an application that lets me interact with my Obsidian data in ways Obsidian itself doesn’t offer. I’ve built several Obsidian plugins, too.

These are small, specific projects, and I use them. That feels like a pretty good definition of success.

Building for myself makes a lot of the decisions easy. I already know which Mac this will run on and how my data is organized. When I’m deciding what a button should do, I can work from the way I actually use the computer.

The software only has to work on my systems. Supporting every possible configuration can stay off the list. When I need a feature, I can work on adding it. When something doesn’t matter to me, I can leave it out.

That limited scope helps enormously. As I get better at explaining what I want, I’m finding that AI does a remarkably good job of helping me put these little systems together.

Of course, I still have to test what the AI builds. Getting an app to launch doesn’t tell me whether it handles my data properly, and I’m responsible for what I run on my computers. But now, when I think, "I wish my computer could do this," building it myself is an option.

I’m hearing about all sorts of these projects in the Robot Builders Club. People who haven’t developed software before are making apps for themselves. They can explain the problem they want to solve because it’s something they already deal with.

And they have a good way to judge the result. Does the app do the thing they wanted it to do? They can try it in their own work and find out.

Then there’s the question of selling what you’ve made.

My inbox is filling up with pitches from people building productivity apps. I understand the enthusiasm. Once you’ve made something useful, it’s easy to start thinking about who else might want it.

But I’m reluctant to recommend an app when I don’t know whether the developer will still be working on it a year from now. What happens when an operating system update breaks something? I want to have some confidence that readers will get help, especially if they’re trusting the app with their work.

Selling it means taking responsibility for other people’s use of it. If you’re excited about that work too, great. It’s worth asking yourself before you put a price on the download.

If your app makes a daily task easier, you already have a reason to keep using it. You can change it when your needs change and spend as much time on it as you like. The fact that someone else might never want it (or you’re not interested in the extra burden that comes with distributing it) doesn’t make it any less useful to you.

I like that I can decide my calendar should start with the current month and then make it work that way. Someone else might prefer January at the top. Mine suits me.

Start with something small enough that you can explain what you want it to do.

What’s one small thing you wish your computer did just for you?

An audience of one is enough.