The tightening in your chest when a problem finally clicks — you probably remember it. Not the answer arriving, but the moment just before: the slow assembling of something difficult into something you understood. There was a particular texture to that feeling. A little heat. The slight surprise of your own competence. It’s worth pausing to ask when you last felt it. Not the version where you typed a question and read the response and thought, yes, that’s right — but the version where the
thinking itself was yours, where the friction of not-yet-knowing gave way to something you built from scratch. For a lot of people who use AI tools daily, that pause produces a longer silence than expected. This isn’t an argument against using these tools. It’s something quieter and more specific: a look at what that small, reliable hit of satisfaction actually is when you paste a result into a document, cross a line off a list, and feel, briefly, capable. Because the feeling is real. The
brain is genuinely registering something. The question is what it’s registering — and whether that something is the same thing it used to be. There is a version of you, probably not very many years ago, who would have spent forty minutes drafting that email. Who would have pulled a book off a shelf, or opened three browser tabs, or called a colleague to think through a problem out loud. Who would have felt, at the end of it, a specific kind of ownership over
the result. Not pride exactly — more like the quiet satisfaction of a muscle that had been used. What the feeling is actually made of Psychology has spent decades studying the relationship between effort and reward. What researchers in this field have consistently found is that the brain doesn’t just reward outcomes — it rewards the process of achieving them. The struggle matters neurologically. The wrong turn that gets corrected. The moment of confusion that resolves. These aren’t inefficiencies the brain tolerates on the way
to the good feeling. They are part of the good feeling. Remove them, and what remains is a thinner signal. There’s a term behavioral researchers use for this: the effort heuristic. We tend to value things more when we perceive that effort went into them — including our own work, our own thinking, our own solutions. It’s why the meal you cooked tastes better than the one that arrived in a bag. Not because the ingredients are superior, but because something of you went into
it. When you finish a task with an AI assistant, the brain still fires its small completion signal. The to-do list item disappears. The document exists where it didn’t before. Something was accomplished in the most literal sense. But the effort heuristic has been quietly bypassed. The struggle — the part that used to make the outcome feel yours — wasn’t there. And the brain, which is not always a precise instrument, doesn’t always notice the difference in the moment. It registers done. It does
not always register: but you didn’t do it. Why does cognitive patience disappear so quietly? Here is what makes this harder to see clearly. The shift doesn’t happen in a single afternoon. There is no moment where you sit down and decide that the part of you that used to work through problems is no longer needed. It happens across dozens of small decisions, each one completely reasonable. The email was genuinely urgent. The research genuinely would have taken hours. The draft genuinely needed to
exist by morning. Every individual choice has a defensible logic. But accumulated over months, those choices amount to something. A capability that was once exercised regularly is now exercised rarely. A particular kind of cognitive patience — the willingness to sit with a half-formed thought until it becomes a whole one — gets less practice. I’ve noticed this in myself: the slight impatience that arrives now when a problem doesn’t resolve quickly, a low-grade restlessness that didn’t used to be there. It feels like distraction.
It might actually be atrophy. The 1990s had a version of this with calculators in schools. The argument was always the same: the tool is faster, the outcome is identical, why insist on the slower method? And the counterargument was always the same, and always slightly difficult to articulate: because the slower method builds something in the person doing it that the faster method doesn’t. Not just skill. Something more like confidence in one’s own thinking. A working relationship with difficulty. What’s different now is
the scope. A calculator handled arithmetic. These tools handle reasoning, language, synthesis, judgment. The category of things a person might once have worked through alone — and felt something about having worked through — has expanded enormously. And the retirement of the effort, when it comes for reasoning itself, is a different kind of loss than the retirement of long division. What happens when you want to share the result? There is one detail in this worth sitting with: the impulse to show someone. You
finish the thing, and there’s a brief pull toward telling a colleague, sending the link, mentioning it in passing. Look what I made. That impulse is genuine, and it’s worth taking seriously, because it’s the part of the brain that still believes the accomplishment was personal. It hasn’t caught up yet. It’s still running the old software, the version that connected finished work to the self who finished it. That gap — between the completion signal the brain fires and the actual source of the
work — is where the confusion lives. Not guilt, exactly. Not dishonesty. Something more like a slight blurring of the line between tool and author, between assisted and achieved. A blurring that, repeated often enough, starts to feel like the normal texture of getting things done. I’ve watched colleagues in their thirties describe a vague dissatisfaction they can’t quite locate. They’re more productive than they’ve ever been. Their output is cleaner, faster, more polished. And yet something feels slightly off, like a song played in
the right key but at the wrong tempo. What they’re describing, I think, is the absence of the friction that used to tell them the work was theirs. What gets lost when effort leaves the room There is something that only comes from figuring something out the slow way: a specific kind of self-trust. Not confidence in your credentials or your output, but confidence in your own cognitive process — the knowledge, built through repetition, that you can sit with something difficult and eventually understand
it. That knowledge is not abstract. It lives in the body, somewhere around the sternum. It’s the feeling of having been confused and then not confused, enough times, that you stop being afraid of confusion. When the tool resolves the confusion before you’ve had to sit with it, that trust doesn’t get built. Or rather, it doesn’t get maintained. It was built earlier, in school, in earlier years of work, in all the times before the shortcut existed. But like most things that aren’t practiced,
it fades. Quietly. Without ceremony. Without a goodbye party, as the headline of this piece puts it — and that phrase is more precise than it might first appear. There is no moment of farewell. Just a gradual absence where something used to be. This is not a case against efficiency. It is not nostalgia for difficulty for its own sake. It is something more specific: a suggestion that the satisfaction you feel when you finish a task with AI assistance is worth examining rather
than simply accepting. Not because the feeling is wrong, but because it might be pointing at something it doesn’t fully understand. The brain is filing the experience under accomplishment. It may be filing it in the wrong drawer. Somewhere in the middle of a Tuesday, between the third task completed and the fourth begun, there is a version of you who used to take longer with things. Who got frustrated, and then got somewhere. Who felt, at the end of a hard afternoon, like they
had been genuinely used. That person hasn’t gone anywhere. But they’ve been given fewer reasons to show up. And one day you might reach for them, for that particular steadiness, and find that the muscle has gone quiet from disuse — and that the last time you really needed it, you opened a browser tab instead.
ChatGPT, AI productivity, effort heuristic, cognitive patience, self-trust, psychology of reward