There's a moment I keep noticing with AI tools. The model can draft, plan, code, summarize, argue either side of anything you like, and the cursor just blinks in an empty box. The bottleneck is not the machine. It's the person in front of it, who has run out of things to want from it.
I think that blinking cursor describes more lives than we'd like to admit: enormous capability, waiting for an assignment. The missing piece has a name, agency, and my growing suspicion is that it's the most important quality almost nobody is deliberately building.
What I mean by agency
Agency is the capacity to originate: to want something from your own center and then actually move toward it, rather than only responding to whatever arrives. That definition has two halves, and they are not equally hard.
The hard half is the wanting. Most of our wants are inherited. They come from family, from whatever our peers happen to be chasing, from a feed that has spent years optimizing itself to show us the life we're supposed to crave. Sorting the wants that are actually yours from the ones that were installed is slow, uncomfortable work, and I'm not convinced anyone ever finishes it. Plenty of people spend decades pursuing a goal they never chose and only find out at the moment of arrival, when the feeling they expected doesn't come.
The other half is movement, and about that there is less to say: wanting without moving is dreaming.
Things that impersonate it
Busyness is the big one. The person whose calendar is packed and whose hands never stop can be the least agentic person in the room, because every one of those hours was assembled out of someone else's requests. I notice it most on days that start in the inbox: the day gets built for me, message by message, and by evening I've been impressively productive on a dozen things I never decided mattered. Motion is easy to confuse with direction, and the confusion is flattering, so it persists.
Anxiety-driven work impersonates agency too; a lot of what reads as drive is fear that has learned to dress well. So does compulsive control, the need to manage every detail and every person, which if anything signals the opposite, since someone clear about the few things that matter can afford to let go of the rest.
The default setting might be passivity
There's an old line of psychology research on this, and its afterlife is instructive. In the 1960s Martin Seligman and Steven Maier ran the experiments that gave us the phrase 'learned helplessness': animals exposed to shocks they couldn't control later failed to escape shocks they could. For fifty years the standard story was that helplessness gets learned. Then in 2016, Maier and Seligman published a reappraisal arguing, on the basis of decades of neuroscience, that the original story was backwards. Passivity, they now think, is the default response to prolonged bad events. What gets learned is control.
I want to be careful here, because famous psychology has a habit of being tidier in the retelling than in the data, and this is one reinterpretation of a complicated literature, not settled fact. But if they're even roughly right, it reframes agency in a useful way. It isn't a factory setting that some unlucky people had knocked out of them. It's an acquisition. You get it by having the experience, repeatedly, of acting and seeing the world move, and you can go a whole life without accumulating much of that experience if your days keep getting assembled for you.
Why it becomes the bottleneck
Herbert Simon saw the shape of this in 1971, writing about information: 'a wealth of information creates a poverty of attention.' The same trade seems to be arriving for execution. When anyone can produce, analyze, build, and generate on demand, capability stops being what distinguishes people, and the scarce thing becomes knowing what is worth doing and caring enough to point the machines at it.
Building an assessment platform has made me notice how one-sided our measuring habits are. We have entire industries of instruments for what people can do, and almost nothing for whether they have anywhere they want to go. Capability is easy to test. Direction barely shows up in the data, which is roughly how it gets treated by schools and employers too.
The dramatic version of this argument says the world will split into people who direct and people who are directed. I distrust how neat that sounds; the world rarely splits neatly, and most of us will keep living somewhere in the middle, directing a little and being directed a lot. But the pressure feels real to me. A tool of enormous capability doesn't fix an absence of wanting, it amplifies whatever wanting is present, and if yours is missing, someone else's will happily route through you.
The honest objection is that all of this is easier to say than to live. Most people have never had much room to originate. The rent is due, the shift starts at nine, and 'figure out what you truly want' is classic advice from people with slack to people without it. I don't have a tidy answer to that. The one thing I'd point at is that the cost of acting on a want, of testing it, prototyping it, learning what it requires, has just dropped further than at any point I can think of, and it dropped for everyone at once.
A week's worth of evidence
If you want to know where you stand, I don't think introspecting in the abstract helps much. Look at the last week instead, concretely, hour by hour if you can bear it. How much of what you did traces back to a decision of yours that it mattered? How much was reaction, assignment, obligation, the current carrying you where it was already going?
My own answer, most weeks, is worse than I'd like. I don't think that's shameful. The current is strong, it's supposed to be, and entire industries work on its behalf. What I'd suggest is small: before next week gets assembled for you, decide one thing it is for, on purpose, in advance. It will feel slightly awkward. As far as I can tell, that awkward feeling is what practicing agency is like at first.