The Identity Crisis

Gergely Orosz published a piece recently called “When AI writes almost all code, what happens to software engineering?” and a follow-up about the grief that comes with it. I’ve had many a late night racking my brain about both of them since.

He’s right about the broad strokes. AI writes most of my code now. Not in the “autocomplete saves me some typing” way. In the “I describe what I want and review what comes back” way. I use Claude Code heavily in and out of work. It builds features, catches bugs, scaffolds entire projects.

And I’m good with that. Mostly.

The Part That’s Hard to Say Out Loud

Here’s where it gets uncomfortable. I spent years getting good at this. Not just writing code, but understanding why code works, how systems fit together, how to debug something at 2am when production is on fire and the logs aren’t helping. That knowledge didn’t come from a tutorial. It came from failing over and over again until the patterns stuck.

Now I watch an agent do in seconds what used to take me hours. And it does it well. Sometimes better than I would have. That’s not a hypothetical. That’s my Tuesday.

Gergely talks about the grief of watching your craft get automated. I don’t think grief is the wrong word, but for me it feels like an understatement. There’s something massively daunting about spending over a decade building skills and then watching the floor fall out from under them. You can rationalize it all you want. You can say “well now I’m higher leverage” or “I’m more productive than ever.” Both of those things are true. But they don’t fully address the feeling.

The feeling is: if the machine can do the thing I was proud of doing, what exactly am I bringing to the table?

What I’ve Landed On

I don’t think the answer is “we become prompt engineers” or whatever the LinkedIn crowd is selling this week. That framing misses the point entirely.

What I actually do now is closer to what a senior engineer was always supposed to do. I decide what to build. I decide how the pieces fit together. I evaluate whether the output is correct, secure, and maintainable. I catch the things that look right but aren’t. I make judgment calls that require context the agents don’t have (yet).

The implementation was never really the hard part. We just spent so much time on it that we confused it with the job.

The Uncomfortable Truth

The engineers who are going to struggle aren’t the ones who can’t prompt. They’re the ones who never developed the judgment layer underneath. If you learned to program by following patterns without understanding why they work, these tools are going to be a problem for you. Not because they replace you, but because you won’t be able to tell when they are wrong.

And it is wrong. Regularly. It writes code that compiles, passes tests, and subtly misses the point. Catching that requires the exact kind of experience that feels like it’s being devalued.

The irony is that the skills you need to work effectively with AI are the same skills that took years of manual engineering to foster. Understanding systems. Reading code critically. Knowing what “good” looks like. You can’t skip the reps and go straight to the AI-assisted workflow. Or you can, but you won’t know when you’re shipping garbage.

Where This Leaves Us

I don’t have a clean conclusion. The identity crisis is real and it’s ongoing. I use AI constantly and I’m better for it. I also sometimes feel like I’m watching my own craft dissolve into something I don’t fully recognize yet. I can’t recall a moment in time where I’ve felt so uneasy about my own future. In all honesty, that’s probably the reason why I, and I’d assume others in my field, are obsessive over this stuff. We are all at the bleeding edge of computational intelligence trying to figure out where we fit into the equation.

Gergely asks what happens to software engineering when AI writes almost all the code. I think the more honest question is what happens to software engineers. The engineering will be fine. The discipline adapts. It always has.

The people in it are the ones figuring out who they are now.