In Search of The List

I just want a list. Simple. Let me describe what's on this list. I just want like three or four things, in some order, that AI will not get better than me at in the area of software engineering and knowledge work. Easy, right? Good, then give me the damn list.

But what if there is no list?

First it was coding. I distinctly remember the feeling of "oh shit, it's better and faster than me". Ultimately, not a huge deal. The models still weren't very good at complex features across a codebase and certainly couldn't do things across the whole product surface. Then they came for that. It's fine, I have system-architecture chops and domain modeling. Then the models improved there. Sure, but they can't run a browser and interact with an Xcode Simulator. Wait. What's that? They can? Crap. Okay, but I have taste and judgment. The models definitely can never have that, right? It's hard for me to imagine what the models won't be better at than humans in terms of skills and tasks in general knowledge work. I alternate between being inspired and terrified with each new model release.

If we can't rely on our ability to deliver better work than the model, what is left? Does knowledge work become the new Marx textile factory where we are alienated from the craft we used to care about? It's a deeply uncomfortable question and feels a bit like staring into the abyss. The question often isn't at the forefront of my mind, but it sits just beneath the surface of my consciousness like a slightly sunk lily pad in a pond. Every now and again, though, the question leaps into view and forces me to confront it. What does an engineer do now? Where am I providing value? What am I going to be doing in a year? Two?

I'd love to be able to arrive at a place of certainty and comfort. You know, wrap the whole thing up with a fun anecdote and a few principles to live by—but I don't think this is one of those situations. The reality is, there's so much uncertainty around what happens next, and we simply don't have a clear picture of how knowledge work will change. But I think it's important to recognize that all of the things that the models are taking over are skills and tasks. Even judgment, taste, nuanced reasoning and decision making fall into that category. But there are human ways of being, like taking responsibility, having ambition, and being curious enough to learn and grow. A model can’t do those things in my place (or yours). The people I already see thriving with AI are those who are being curious and ambitious and running all sorts of experiments in service of learning and growth.

So, back to my opening request. I've come to realize that I won't get that list. And I still don't know what I'll be doing in a year. Or two.

But I’m starting to see that uncertainty leaves room for things to turn out better than I can currently imagine, too. There are things I want to build, problems I care about, and ideas I can now try that would have been out of reach a year ago. That makes me curious about what comes next, even when it scares me.

Being curious, taking responsibility, and paying attention won’t guarantee me a comfortable future. But they give me somewhere to start, and a way to help shape what happens next.

I don’t have the list I wanted. I do have things I want to try.