Building for fun again

I started programming because I thought it was magic. Not metaphorically. I literally thought being able to tell a computer what to do and have it obey was the closest thing to magic that existed.

I was maybe twelve or thirteen. I didn’t have a plan for what I was building. I didn’t have a career goal. I just liked making things.

Then it became my job.

This is such a common story among programmers that it barely needs telling, but I’ll tell it anyway, because the ending is different than I expected. You start out programming for fun. You get good at it. Someone offers to pay you to do it. You think, great, getting paid for my hobby. And for a while it is great. But slowly, imperceptibly, the thing you loved becomes the thing you do. Obligation just has a way of draining the joy out of almost anything.

For me it happened gradually over maybe five years. The technical problems at work were interesting enough. But the impulse to just make something, for no reason, with no spec, no deadline, that impulse went away. I didn’t even notice until it was gone.

I spent over a decade in that state. I was a professional dev. I was good at it. But I had completely stopped building things for fun. If you had asked me why, I would have said I didn’t have time, which is what everyone says when the real answer is they don’t have the desire.

The spark

Then, in late 2022, OpenAI released ChatGPT.

I remember the first time I used it. I typed in some half-formed prompt and got back something that genuinely surprised me. Not because it was good (it wasn’t, really), but because it felt like the computer was doing something I didn’t explicitly tell it to do. That feeling, the computer surprising you, was exactly what I’d been missing for twenty years without knowing it.

Same thing I’d felt as a twelve-year-old. Something between wonder and possibility. The sense that there are things here I haven’t figured out yet, and figuring them out would be fun.

Within a week I was building things again. Not for work, not for any particular purpose. Just tinkering, staying up too late because I wanted to see if something would work. I hadn’t done that since my twenties.

What changed

I think it comes down to the ratio of idea to implementation.

When you’re a professional developer, you spend most of your time on implementation (if you’re lucky). The idea is the easy part. The hard part is the months of work to make it real. That’s fine when someone is paying you, but it’s exactly the wrong ratio for hobby projects. I built things as a kid because I had ideas and I wanted to see them exist, and my ideas were small enough that the implementation was trivial. As an adult, my ideas outgrew the time I had. So I just stopped.

AI changed that ratio. Suddenly I could go from idea to working prototype in an evening. Not a crappy prototype, a real, functional thing that actually did what I wanted. The gap between “wouldn’t it be cool if…” and “here, try this” collapsed from months to hours.

AI handles enough of the mechanical work that you can stay in the part where building is actually fun, the part where you’re thinking about what to make instead of fighting with boilerplate. And there’s this thing that happens when you’re building for fun that you never get at work: you follow your curiosity wherever it goes. You go down rabbit holes. You change direction because something unexpected caught your attention. You can’t do that on a work project. There are requirements and deadlines and people depending on you. But that kind of undirected exploration is what makes building feel like play instead of labor.

AI makes it viable again for adults with jobs and families and limited free time. You can follow a rabbit hole for an evening and come out the other side with something that works. If it turns out to be a dead end, you’ve lost one evening, not one month.

The nerd awakens

I’ve been building things for fun again for several years now, and I’ve shipped more personal projects in that time than in the previous two decades combined. None of them are world-changing. Some are useful only to me. A few are things I’ve shared that other people actually use. The point isn’t really the output though. It’s that somewhere along the way, I started wanting to build again. And the thing that rekindled that? Honestly, it was AI.

Which makes the rest of this complicated. I’m not going to pretend the anxiety around AI isn’t real. I feel it too. The same tool that gave me back my love of building is also the one that makes people wonder how long their jobs will last. Both things are true at the same time, and I don’t think you have to pick one. But I can tell you what’s true for me right now: I’m building things again just because I want to see if they’ll work, staying up past midnight for no good reason, and it feels exactly like it did when I was thirteen.