Why are some of my most experienced people the ones struggling most with AI?

A brilliant developer told me he loves coding. It’s his passion. But little by little, he said, he won’t be able to do it as a job anymore. Because of AI. He’d tried the vibe coding thing. It just didn’t feel like him. And he could feel it — the slow drain of something he’d always known where to find.

Something more nuanced than fear.

You’ve probably felt a version of this. Maybe not about coding. Maybe about writing, or analysis, or the particular kind of thinking your work used to require. The thing you spent years getting good at. The thing that, somewhere along the way, stopped being just a skill and started being part of how you understood yourself.

That’s the part nobody’s naming yet.

We keep calling it anxiety about AI. Fear of replacement. Worry about the future of work. And those things are real. But the word “anxiety” is doing a lot of work to cover something more specific: the moment when the equation you built your identity on stops adding up.

The equation most of us never had to examine. You work. You produce. You deliver. You matter. It was so embedded it stopped feeling like a belief and started feeling like oxygen. Nobody questions oxygen.

Until something makes the air thin.

The brain does something specific when identity is threatened. Neuroscientists call it predictive processing — your brain isn’t reacting to reality, it’s predicting it, constantly, based on models it has built over time. One of the most deeply held models is the one it built about you. Who you are. What you do. What that means.

When AI starts producing what you used to produce — faster, without the years, without the effort — the brain reads that as a prediction error. Not a workflow disruption. A who-am-I disruption.

And faced with a prediction error, the brain does one of two things: it updates the model, or it pulls back toward the old one. Updating is risky. The old model has years of evidence behind it. So the brain resists. Because it’s doing exactly what it was designed to do.

This is not only happening inside individuals.

The same mechanism runs through teams, organizations, whole industries. Identity is not only internal, we partly know who we are through what the world reflects back. Colleagues, culture, the way a room responds when you speak. All of it confirming a model.

When AI transformation arrives inside a company, it doesn’t just ask people to learn new tools. It asks them to update their identity model inside a social system that is still running the old one. The environment keeps confirming who they used to be. No wonder it stalls.

We’ve been calling it a technical problem. Then a change management problem. The tools aren’t landing, so we add training. The training isn’t landing, so we add incentives. And still something resists.

Because the block was never technical. It was never even behavioral. It’s identity-level. And almost nobody in an AI strategy has budgeted for that.

There’s a developer somewhere in your company right now doing the same math as my friend. Using AI. Getting faster. And quietly asking whether the thing that made him worth something is still his.

That question doesn’t show up in a usage report.

What we’re measuring tells us how much. It doesn’t tell us what it’s costing. And what it’s costing — in trust, in meaning, in the particular kind of contribution that only comes from someone who feels grounded in what they bring — that’s the number nobody’s tracking.

We were always going to arrive here. The question of who we are when the output is no longer the proof. AI just made it impossible to postpone.

And that question, sitting with it honestly, inside companies and inside ourselves, that’s where transformation actually begins.