In 2023, AI could solve 4.4% of the problems in a standard software engineering benchmark. By 2024, that number was 71.7%. A Stanford Digital Economy Lab study, using payroll records from millions of American workers, found that employment for software developers aged 22 to 25 has dropped nearly 20% since late 2022. Across all AI-exposed professions — marketing, customer service, accounting — early-career hiring fell 13% relative to fields like nursing and maintenance work.
A 2026 Gallup survey found that workers at companies that have already adopted AI are more anxious about losing their jobs than those at companies that haven't. The people most unsettled are not the ones who've been laid off. They're the ones still employed — watching the tools improve, still doing the work, still getting paid — but no longer sure what their work will mean in two years..
The practical advice is everywhere. Upskill. Learn prompt engineering. Diversify your capabilities. Adapt or be left behind. All of it makes sense. None of it seems to touch the thing that's actually bothering people.
The familiar shape of this
If you're old enough to remember the 2008 financial crisis, or the dotcom crash before it, or even just a major restructuring at your own company — you've felt this before. The specifics were different. The feeling wasn't.
It's the feeling of the ground moving. Not that something bad has happened, but that the thing you were standing on — the assumption that your work will continue to exist in roughly its current form — turns out to have been less solid than you thought. The world shifts, and you realise you'd been leaning on something without knowing it.
What's interesting about this feeling is how little it has to do with the specifics. In 2008, the fear was about markets. In 2020, it was about a virus. Now it's about AI. The surface changes. The feeling doesn't. It arrives with the same weight, the same low hum of uncertainty, the same quiet question at 2am: will I be okay?
This pattern is worth noticing. Not because it makes the current disruption less real — it is real, the numbers are clear — but because the fear itself seems to predate its object. It was already there, looking for something to attach to. AI is just the latest surface.
What the fear is about, and what it isn't
Ask someone what they're afraid of and they'll give a practical answer. Losing my income. Falling behind. Not being able to provide for my family. Having to start over at forty-five. These are real concerns. They describe real consequences. Anyone dismissing them is not paying attention.
But there's something underneath the practical layer that's harder to name. You notice it when the practical concerns are addressed and the unease doesn't go away. You've taken the course. You've updated your skills. Your job is, for now, secure. And still — something.
It's not just about money, and it's not just about identity, though it includes both. It's closer to this: the discovery that the life you'd arranged — the career, the expertise, the daily structure, the sense of knowing where you stand — was resting on an assumption you'd never examined. The assumption that the world would continue to need you in roughly the way it needs you now.
That assumption isn't about AI. It's about permanence. It's the belief — usually unconscious, usually unspoken — that the ground will hold.
Everybody carries this. Not just knowledge workers watching AI demos. The farmer watching climate patterns shift. The factory worker watching automation creep across the floor. The parent watching their children grow up in a world they don't recognise. The specific threat changes. The underlying vulnerability doesn't.
The ground that wasn't there
Evan Thompson, a philosopher at the University of British Columbia, has spent years working at the intersection of neuroscience and contemplative traditions. In Waking, Dreaming, Being, he makes a case that would unsettle most people if they sat with it long enough: the self is not a thing. It's a process.
Not a metaphor. Not a philosophical position staked out for the sake of argument. A description of what the neuroscience actually shows. What feels like a continuous, stable "you" — the person who woke up this morning, who remembers yesterday, who plans for next year — is closer to a running pattern. A habit of self-reference so seamless it feels like a fact. But it's generated moment to moment. It doesn't persist between those moments any more than a wave persists between crests.
This matters for the AI conversation because it changes the shape of the fear. If identity is a process — something the brain is doing, not something the brain contains — then the feeling of instability that AI provokes isn't new. It was always the case. The ground you were standing on was never solid. It was a process pretending to be a floor.
AI didn't create the instability. It interrupted the pretence.
An old vulnerability
David Loy, a philosopher working across Zen Buddhism and Western thought, describes this as the self's "sense of lack" — a persistent, low-level feeling that something is missing, that identity is never quite complete. Not a psychological disorder. A structural feature. The self, because it is constructed rather than given, always needs the next thing to hold itself together. The next achievement, the next validation, the next role.
You can hear this in the way people talk about their careers. Not just as ways of earning a living — though they are that — but as ways of knowing who they are. The job provides more than income. It provides location. It tells you where you stand, how you matter, what you're for. Remove it, and something opens up that has nothing to do with money.
This is why the fear around AI feels different from ordinary career anxiety. It's not just that the work might go away. It's that the work was doing something for the self — anchoring it, justifying it, giving it weight — and the prospect of that anchor dissolving reveals how much was depending on it.
But this dependency wasn't created by AI. It was there when you got your first promotion and felt, for the first time, like you'd arrived. It was there when you lost a project and felt, not just disappointed, but diminished — smaller, somehow, than you'd been the day before. It was there every time a change in circumstances shook your sense of who you were.
AI is the current trigger. The vulnerability is older than any technology.
Sitting with the question
There's a simple experiment that most people avoid, and for good reason. It's uncomfortable.
Sit still for a few minutes. No phone, no task, nothing to do. And instead of following the thoughts that arise — the plans, the worries, the commentary — try to find the one who is thinking them.
Not the thoughts themselves. Those are easy to locate. They appear and disappear like weather. But the thinker. The one behind the thoughts. The person to whom all of this is supposedly happening.
You'll find sensations. You'll find memories. You'll find the thought "this is me." But the me that the thought refers to — can you find it as a thing that exists independently of the thought?
This isn't a trick, and it isn't a therapy exercise. It's the oldest inquiry there is. And it's directly relevant to the AI question, because the thing we're afraid of losing — the stable, continuous self whose future is under threat — may not be there in the way we assumed.
Not that there's nothing here. Obviously something is here — reading, thinking, feeling. But the solid, fixed entity that needs protecting? That might be a story rather than a fact. A very convincing one. One that updates itself so quickly it never seems to have gaps. But a story nonetheless.
What changes when you see this
Nothing about the practical situation changes. The layoffs are real. The economic disruption is real. The need to adapt is real. People who need to retrain should retrain. People who need to change direction should change direction. This essay isn't advice to sit on a cushion while your industry transforms.
But something shifts in the relationship to the fear itself. When you see that the self being threatened was always a process rather than a fixture — always assembling itself from conditions, always needing the next thing to hold together — the fear doesn't necessarily disappear. But it loosens. It becomes something you can feel without being organised by it.
The difference is between being afraid and being swallowed by fear. Between feeling the ground move and believing that you are the ground.
Most of the conversation about AI stays on the surface: which jobs will go, which skills to learn, how to stay relevant. All important. But underneath those questions sits a quieter one that rarely gets asked, because asking it changes the entire frame:
Who is the one who is afraid?
Not what are they afraid of. Not how should they respond. Who are they — and are they as solid as they feel?
If the answer is that they're a process, not a fixture, then what's here when the process is seen for what it is isn't emptiness or nihilism. It's the awareness in which the whole drama — the fear, the career, the disruption, the self — has been unfolding all along. It was here before the job title. It's here now. It doesn't need protecting, because it was never in danger.
Not because the world is safe. Because awareness isn't the kind of thing that can be made redundant.
Sources and further reading:
- Brynjolfsson, Chandar & Chen, Stanford Digital Economy Lab — on AI's measured impact on early-career employment
- Evan Thompson, "Waking, Dreaming, Being" — on the self as a process in neuroscience and contemplative traditions
- David Loy, "Lack and Transcendence" — on the constructed self and its sense of fundamental incompleteness
- Nisargadatta Maharaj, "I Am That" — on identity as process rather than entity