I’ve been building things on the web for over a decade now, and lately I’ve started to feel like I’m not really a programmer anymore. That probably sounds dramatic coming from someone who still opens a code editor most days, so let me explain.
For most of my career the job had a familiar shape, even when the work itself was hard. A client needed a website. I sat down, opened my editor, and wrote it. Line by line and file by file, I turned an idea into something you could click. The craft lived in the typing almost as much as in the thinking, and I loved that. I still do. There’s something honest and satisfying about making a thing with your own hands and watching it come alive in a browser.
But the shape of the work is changing, and so am I.
These days I spend far less time writing every line and far more time directing. I describe what I want, I set the constraints, I review what comes back, and I decide whether it’s actually any good. I’m orchestrating at least as much as I’m authoring. The center of gravity has quietly moved from my keyboard to my judgment. If I had to pick a word for it, I’d say I’ve become less of a programmer and more of a builder, and on the bigger projects, more of an architect.
A recent project made this click for me. I rebuilt doorcounty.com, moving the entire site off an old CMS and onto WordPress. The catch was that I had almost no source code and no database to import from. A few years ago that would have meant months of slow, manual rebuilding, copying content by hand from one system into another. Instead I designed an intelligent pipeline that paired AI with my own scripting to read the old site, make sense of its content, and bring all of it across into a clean, modern WordPress build the client could edit themselves. I wasn’t the one typing out every listing. I was the one who decided how the whole thing should work, then steered it until it did.
That’s the shift in a single sentence. The value is moving away from how fast I can write code and toward how well I can decide what to build, how the pieces should fit, and whether the result is worth shipping.
Here’s the part I think people miss when they talk about AI replacing developers. Directing AI well is not the same as knowing nothing and hoping for the best. It’s the opposite. You can’t steer what you don’t understand. When AI hands me an approach, I need to know whether it’s sound, whether it will scale, where it will quietly break, and whether it’s even solving the right problem in the first place. That judgment didn’t appear out of thin air. It’s the direct result of a decade plus of writing code by hand, shipping real sites, breaking things, and fixing them at two in the morning.
My years in marketing agencies matter here just as much as the development side. Agency work taught me to start from the client and the goal instead of the code. Who is this for? What does the business actually need? What will a real visitor feel when they land on the page? AI is remarkably good at producing output. It is not good at deciding whether that output serves the person paying for it, or the person using it. That part is still mine, and it turns out the marketing half of my career prepared me for this moment far more than I ever expected it to.
So I don’t think the craft is disappearing. I think it’s relocating. It used to live in my keystrokes. Now it lives in my decisions. The taste, the architecture, the feel for what good looks like, the instinct for what a client really means when they struggle to describe what they want, all of that matters more now, not less. The tools got faster, and that only raises the value of knowing where to point them.
If I’m honest, there’s a little nostalgia mixed in. I came up loving the hands on part, the late nights getting a layout pixel perfect, the small thrill of a stubborn function finally working. I’m not ready to give that up, and on the right project I still happily disappear into the code for hours. But I’m also genuinely excited. I get to think at a higher level, take on bigger and stranger problems, and build more in a week than I used to build in a month. The Door County migration is a good example of something that simply wasn’t possible for me, working solo, even a couple of years ago.
I’ve spent my whole career on the web, watching it change every single year. This feels like one of the bigger changes, and I’d rather lean into it than pretend it isn’t happening. I’m still the same developer who fell in love with WordPress and Tailwind and the plain satisfaction of shipping something real. I’m just doing more directing and architecting now, and leaning harder than ever on everything the last decade taught me.
I used to say my job was writing code. Now I think my job is building, and figuring out what’s worth building in the first place.
Keep building.