A few weeks ago, I was talking with colleagues about AI and jobs. Will it take them, won’t it, what should people do. Somewhere in the middle of all that, someone said something that stuck.
“The problem isn’t losing your job. The problem is that most people don’t know who they are without it.”
It names something the entire AI-and-jobs conversation keeps skating past.
The deal we all signed up for
Yuval Noah Harari makes this observation in Nexus that puts things in perspective. In the year 1000 AD, a farmer’s son was a farmer. His grandson was a farmer. His great-grandson was a farmer. Nobody chose a career because there was no choice to make.
Modern schooling changed the options but not the structure. Starting in the 1700s, kids went to classrooms, picked up trades, entered professions. More variety. But the underlying deal stayed the same: spend your first 20 years learning, spend the next 40 applying what you learned.
That deal held for a remarkably long time. Most of us still operate on it without realizing it.
And it’s expiring.
The wrong debate
I’m not going to argue about whether AI takes jobs. That framing misses the point. AI is changing what jobs are. The tasks that made up your role five years ago won’t be the same tasks five years from now.
Harari gives this sharp example. A 45-year-old fighter pilot, one of the best, gets asked to train a drone system. He does it brilliantly. Then the drone takes his seat. Not because he was bad at his job. Because he was so good at it that he made himself replaceable.
Now he has to start over. At 45. With 25 years of expertise in something the world no longer needs a human to do.
The economic problem is obvious. But there’s a deeper one underneath it that nobody wants to name.
The status trap
People don’t resist change because they’re lazy or stupid. They resist it because starting over means losing status.
A senior professional with 20 years of experience decides to learn a new skill. On day one, they’re at the same level as a 23-year-old fresh out of college. They don’t just feel uncomfortable. They feel humiliated. They go from being the person others come to for answers to being the person asking basic questions.
Nobody talks about the humiliation part. But that’s the real barrier. Not the learning curve. The ego curve.
I watched this happen with my father. When computers were introduced at his workplace, he couldn’t adapt. He was in a government job, so he wasn’t let go. But he never got better opportunities either. His career froze while people around him moved ahead.
He didn’t lack intelligence. He lacked the willingness to be a beginner again, in front of people who used to look up to him. Being a beginner at 50, asking a 25-year-old how to do something on a computer, isn’t a skills problem. It’s a status problem. It feels like falling.
So he didn’t try. The incentive structure made it rational to stay put, protect what he had, and hope the change wouldn’t catch up.
But it did. It always does.
And this is how the trap works. You get good at something. You get recognized for it. Your competence becomes your reputation. Your reputation becomes your identity. And once your identity is welded to a specific skill, protecting that skill starts to feel like protecting yourself.
So when something new comes along, you don’t adopt it early. You can’t. Early adoption means admitting you’re not good at the new thing yet. It means risking the status you spent years building. So you wait. You tell yourself you’re being prudent. That the new thing is overhyped. That your fundamentals still matter. But while you’re waiting, other people are learning. And a gap opens up.
That’s the delayed adoption problem. It’s not that people can’t learn the new thing. It’s that learning it costs them something they’re not willing to pay: their position in the room.
And delayed adoption leads to something worse. Identity lag. The technology moves on. The job changes. But your sense of who you are is still stuck in the old version. The world has updated. Your internal operating system hasn’t.
Most of us learned our core skill in our early twenties. Got decent by 28. Got good by 32. And then just… kept doing it. More senior titles, more meetings, but nothing fundamentally new. We ran the same operating system for decades and called it experience.
But experience and learning are not the same thing. Experience is doing what you already know in more situations. Learning is acquiring the ability to do something you couldn’t do before. Most professionals haven’t been a beginner at anything in 20 years.
Why this time is different
What happened to my father with computers was a gentle disruption. It played out over years. People had time, even if most wasted it.
AI is the same kind of revolution at a completely different magnitude. The cycle that took a decade now takes months. The status trap, the delayed adoption, the identity lag — all of it is going to play out faster than people can process.
That’s why “upskill” isn’t enough. Every reskilling program, every corporate training, every LinkedIn post telling you to learn prompt engineering treats this as a skills problem. But the bottleneck isn’t learning the new skill. It’s letting go of the old identity. No training program in the world addresses that.
What actually helps
I can tell you what this looks like from the inside, because I almost fell into the same trap.
I work in product management. For months, I told myself my role was safe. Product thinking is human. Strategy is human. AI can’t do what I do. I had all the arguments lined up.
But there was this uneasiness in my chest that wouldn’t go away. Deep down, I knew that sitting on past accolades while everything around me was shifting was the worst possible thing I could do. I was doing exactly what my father did. Telling myself the change didn’t apply to me.
So I enrolled in a couple of agentic AI coding courses. Went back to being a complete beginner. The first few weeks were humbling in a way I hadn’t experienced since my early twenties. I was slow. I was confused. I asked questions that probably seemed obvious to everyone else.
But something unlocked that I wasn’t expecting. Not just a new skill. A new way of relating to my own competence. I stopped being a product manager nervously watching AI from the sidelines and became a person actively learning something new. The identity shifted, and the anxiety went with it.
The discomfort of being a beginner is temporary. The discomfort of knowing you should be learning and choosing not to is permanent.
Two generations, one lesson
My father couldn’t make the leap when computers arrived. Not because he was incapable, but because the cost felt too high. The status, the dignity, the sense of self he’d built over decades. He chose to protect all of that. I understand why. I nearly made the same choice.
The difference isn’t intelligence or willpower. It’s whether you see your identity as something fixed or something you can update. My father’s generation was never taught that the operating system needed updating. Ours has no excuse.
AI isn’t going to wait for anyone to feel ready. The only way ahead is to get comfortable with being uncomfortable. Not by learning the right skill. By becoming the kind of person who can learn any skill, at any age, without needing to be good at it right away.