What the Machinist Lost Wasn't the Job

The wheel on my office chair broke a few months ago. One of those small plastic casters that you don’t think about until it splits in half and you’re sitting crooked. I looked online for a replacement, couldn’t find the exact one, gave up after ten minutes. Then I opened Fusion, measured what was left of the old wheel with calipers, modeled a new one, printed it. It didn’t fit. I adjusted, printed again. Fit perfectly. The only things I bought were filament and a pack of wheel bearings.

Ten years ago, that fix would have looked completely different. Aluminum tubing, a hacksaw, a drill, a bench, the right kind of plastic to turn down on a lathe. An afternoon, easy. Now it was an evening at my desk and a printer humming in the corner.

I love that. I really do. I love that I can think a thing and have it in my hand the next morning. But something about the chair wheel has been bothering me, and I think I finally figured out what.

A while later my motorcycle had a stripped thread in the oil filter housing. The threads in the engine block, not the bolt — which is the worse of the two, because you can’t just buy a new block. The fix is a helicoil. You drill the stripped hole bigger, cut new threads with a tap, screw in a steel insert that becomes the new thread. It’s a delicate job. You’re drilling into an engine. If you wander, you wander into something expensive.

I did it in the parking lot in front of my building. It worked. The bike runs.

And here’s the thing I keep turning over in my head. I wasn’t scared. Not because I’m some kind of mechanic — I’m not. I was unscared because I’d done this kind of work before, years before, in shop class, on a lathe, on scrap metal that didn’t matter. I’d drilled crooked holes and learned what crooked feels like in your hands before it shows up in the part. I’d cut threads badly and seen what badly looks like. By the time the engine block was in front of me, my hands already knew things I couldn’t have written down.

The chair wheel didn’t teach me any of that. The chair wheel was a lovely, frictionless solve. The helicoil was the bill coming due for years of unglamorous practice.

For most of the twentieth century, machining was a craft. Years of apprenticeship. Math and metallurgy in your head, sure, but also a feel for metal that you couldn’t really teach — only stand next to and absorb. Then, slowly, over decades, numerical control arrived. The machinist stopped guiding the cutter with their hands and started writing instructions for a computer that guided the cutter instead. The output got better. The parts got more precise. The work got faster.

When people argued about it at the time, they argued about jobs. Would the machinist be replaced? Would the craft die? The answer turned out to be more interesting and more uncomfortable than either side wanted. The job didn’t disappear. It changed. The machinist became someone who oversaw machines that broke in complicated ways, and that turned out to be a skill of its own, and the pay was probably better.

What got lost was something quieter. The hands. The thing you couldn’t put on a résumé. The years of standing at a manual lathe before you ever touched the automated one. The new generation of operators didn’t have that. They didn’t need to, mostly. Until they did, and they couldn’t say why something felt off, because the feeling never had time to develop.

I’m a developer. I write software for a living. And I’ve been using AI to help me write code for a while now, and it’s good, and it’s getting better, and I love it the same way I love the 3D printer. I can think a thing and have it running by lunch. The friction is gone.

But I learned to code the slow way. I wrote the bad version first. I spent weeks on bugs that a model would have spotted in a second. I built the wrong abstraction and lived with it long enough to understand why it was wrong. None of that was efficient. All of it is the reason I can look at what the model gives me and know, sometimes without being able to say why, that something is off. That a function is going to be a problem later. That this isn’t quite the shape the system wants.

That’s the helicoil moment, for me. The chair wheel is most of my work now, and it’s wonderful. But every so often something goes wrong in a way that needs the old hands. And the old hands only exist because I spent years on the lathe.

I don’t know what to do with this, honestly. I’m not going to pretend the slow way was better — a lot of it was just suffering, and a lot of the suffering taught me nothing. I’m not going to tell anyone to ignore the new tools. I use them every day. They’re extraordinary.

But I think about the kid who’s going to learn to code now, with a model whispering in their ear from day one, and I don’t know what their helicoil moment looks like. Maybe they’ll have one I can’t picture yet. Maybe the feel just attaches to whatever loop you grew up inside, and theirs will be different from mine but no less real. Maybe.

Or maybe they’ll get to the engine block one day and find that the hands they need were supposed to come from somewhere they never went.

I’m not sure yet. I’m watching.