Why I Am Still a Programmer in the AI Age

If you are not a professional software engineer, you may not realize how drastically and quickly AI has changed our craft. At times I feel like a dinosaur in the months after the meteor slammed into the Yucatán. The lush tropical environment I evolved in is gone, and it’s too early to tell if I am a member of a species doomed to extinction or if I will successfully evolve into something new. I suspect motivation will influence my survival odds; without it, I may simply stop trying. In January 2024, before AI was a factor in daily programming, I wrote about two reasons I loved the craft. It’s time to revisit them and see how they hold up now that AI has changed how programmers work.

The first motivation I talked about was a love of summoning things into existence using nothing but words, like wizards from fairy tales. This is even more awesome in the AI era, and the “wizard” metaphor has become even more apt. When programming in the year 2024, you could be confident that your development tools would do exactly what you asked for, no more and no less. In the year 2026, this is no longer true. AI programming tools like Claude Code have a mind of their own — this is what makes it so much easier and faster to build things, but it also introduces a whole new way things can go wrong. Like the Sorcerer’s Apprentice, your tools can go rogue. You better hope you’ve built a strong enough binding circle¹ to keep your programming demons contained.

The second motivation I wrote about: I love learning, and programming helps me learn. I wrote back in 2024: “One of the greatest pleasures I get is when I understand how something works…. Writing the precise instructions needed to get a computer to do anything provides the same kind of feedback of ‘do I really understand this’ as trying to explain it to another human.” On this front, AI programming tools are a mixed bag. On the one hand, AI is a superpower that helps me quickly become productive in new codebases and with new tech stacks. On the other hand, for the first time in the history of this industry, you can ship code without understanding how it works. Using AI, you can learn faster, but nothing forces you to master concepts the way we had to in the hand-crafted-code era.

That is the biggest change for me. Before AI agents, learning was bundled into the work. If I spent a few hours building a feature, I was also strengthening my mental models of how the system works along the way. Now those two outcomes can separate. I can ship the feature and still have only a shallow understanding of what changed unless I deliberately slow down and rebuild the mental model for myself.

So far, I’ve been able to use AI diligently, with an equal focus on “building” and “learning.” It has been a huge positive boost to my motivation. However, I see a risk that companies start treating learning as optional. Modern software companies fetishize speed. If people believe they can ship features faster by delegating the “understanding how it’s built” part to Claude, who will be able to resist?

One counterargument: Don’t skip the “understanding” part of computer programming, because if you do, the speed gains are illusory. Today’s cognitive debt² becomes tomorrow’s fragile, buggy, and late project. I am optimistic companies will accept this one, as it speaks their language: “What helps us ship faster?”

There is another argument I find more appealing. It speaks directly to my motivation, though I am not sure employers will find it as persuasive. “Successful programmers are addicted to learning. These are the people you want to have on your team, as this trait makes them adaptable as technology changes. If you want these people to thrive, you must continue to provide an environment that encourages learning and mastery. Don’t sacrifice that for speed in the short term, as it will cost you talent in the long term.”

I won’t try to predict what AI will do to the software industry over the next couple of years, or how employers will respond to the changing landscape. For now, my personal motivation remains high. I feel myself adapting. My scales are changing to feathers.

A modern-day dinosaur. Photo taken in my back yard.