There's a specific kind of vertigo that comes with watching something you spent years building become... ordinary.
I remember the first time I got a compliment for writing clean, well-structured code. It felt earned. Not because I was born knowing how to do it — but because I'd put in the time. From my college days, I was the person who stayed up debugging when others gave up, who read documentation other people skimmed, who slowly built a mental model of systems that felt like a private language only I could speak.
That private language, it turns out, is now available as a prompt.
I'm not catastrophizing. The code AI writes isn't always good. It hallucinates, it misses context, it can't hold the full shape of a complex system in its head the way a senior engineer can. But here's what I keep coming back to: for a large slice of what I used to call "my skills" — the kind that took years to develop — a determined amateur with the right prompts can get 80% of the way there in a weekend.
That's a strange thing to reckon with.
I've watched engineers respond to this in two ways. The first response is to shrug and keep going — maybe lean into AI a little more, get faster at shipping, collect the productivity gains. Nothing fundamentally changes. This path is comfortable and arguably rational. There's no shame in it.
The second response is more unsettling, and more interesting: it forces the question of what you were actually building toward. Because if the skill is no longer the moat, then what is?
I've been sitting with that question more than I'd like to admit.
What I'm slowly discovering is that there are things AI hasn't absorbed yet — and they're not the things I expected to matter.
The ability to ask a better question in the first place. The judgment to know when the technically correct solution is the wrong one. The instinct for what a user will feel when they hit an edge case, not just what they'll see. The organizational wisdom to know that this feature needs to be built, but not now, not by this team, not in this way.
These aren't soft skills in the dismissive sense. They're a different kind of hard skill — harder to build, harder to demonstrate, harder to hire for. And they compound in strange ways. The engineer who understands why a system is designed the way it is will always be more valuable than the one who can only describe what it does.
The year AI made my skills ordinary forced me to start asking: which skills are actually mine, as opposed to the ones I inherited from repetition? I don't have a clean answer. But the question feels important — maybe the most important one I've sat with in a while.
If you're a builder in any form, I suspect you know the feeling. And I think the fact that it's uncomfortable is worth paying attention to.