The work is done. The document is finished, the email sent, the outline built, the research summarised into clean bullet points. It is 4:40 on a Tuesday and you have, by any reasonable measure, had a good afternoon. You close the laptop with the small click of something completed. And then — nothing. Not relief. Not satisfaction. Something closer to the feeling of reaching for a glass of water and finding it already empty. A kind of low-grade flatness, like a room where the music
just stopped and no one has spoken yet. You sit with it for a moment. You check the document again. It’s good. It might even be better than what you’d have produced alone. So why does it feel like you weren’t quite there for it? This is the experience that a lot of people using AI tools are quietly carrying around, usually without a name for it. They assume — reasonably — that they’re tired. That they’ve been staring at screens too long, or that
the work itself was draining. Some wonder if they’re becoming dependent on something they don’t fully understand. The word that surfaces most often, when people describe this feeling to friends or in the small hours of a group chat, is burnout. It sounds right. It has the correct texture of modern exhaustion. But burnout, properly understood, is the result of prolonged depletion — of giving more than you have, repeatedly, over time. What this feeling is, is almost the opposite. It is the strange hollowness
of having received output without generating effort. It is not tiredness. It is the quiet absence of something your brain was counting on. What the brain is actually looking for There is a particular kind of satisfaction that comes from thinking something through — not from arriving at an answer, but from the process of getting there. The false start. The paragraph you delete. The moment you realize the argument you were building doesn’t hold, and you have to go back two steps and rebuild
it differently. Cognitive psychology has long observed that this friction — the resistance, the effortful processing — is not incidental to thinking. It is thinking, in the most meaningful sense. The struggle is the signal. When you work through a problem with genuine mental effort, your brain encodes the process, not just the product. There’s a felt sense of authorship that gets laid down alongside the output — a kind of internal receipt that says: you were here, you did this, this thought passed through
you. It is not glamorous. It often feels like wading. But that resistance is how the brain registers that something real happened. What AI tools do — and do extraordinarily well — is remove that resistance. The friction disappears. The output arrives clean and fast, like a door that opens before you’ve finished reaching for the handle. And the brain, which was bracing for effort, finds itself in a room it didn’t walk into. The receipt that never prints I’ve noticed this in my own
work: the difference between an afternoon spent wrestling with a piece of writing and an afternoon spent refining prompts to produce something similar. Both afternoons can yield the same document. Only one of them leaves me feeling like I thought something. There’s a useful way to understand this. When we learn something through difficulty — through the specific texture of not-knowing-yet — we tend to retain it differently than when it arrives pre-digested. The same principle seems to apply to the feeling of having done
work. The brain uses effort as a proxy for engagement. Friction is the evidence. Without it, the output exists, but the internal record of having produced it is strangely thin. This is what creates the empty-satisfaction feeling. It is not that the work was bad. It is that the work arrived without the cognitive fingerprints your brain expected to find on it. You reach for the memory of having thought through the problem, and it isn’t quite there — because the thinking, in the traditional
sense, wasn’t quite required. The unfinished feeling is real, and it is pointing at something accurate. Your brain is telling you, in its quiet way, that something is missing from the ledger. Why this matters more than it sounds There is a version of this conversation that turns quickly into a screed about technology, and this is not that. AI tools are not the villain here, any more than a dishwasher is the villain in a story about someone who misses the meditative quality of
washing up by hand. The tool does what it does. The question is what we lose track of when we stop noticing what the friction was for. Because the friction was doing more than one thing. It was building the thought, yes. But it was also building the thinker. The slow accumulation of effortful work — the paragraphs wrestled into shape, the arguments tested and revised, the ideas that didn’t survive contact with the page — this is how people develop a particular kind of
confidence. Not the confidence of having produced good work, but the quieter confidence of knowing how they think. Of having a relationship with their own mind that was forged through use. When that process is consistently outsourced, the output remains. The development doesn’t, quite. And this, I think, is the deeper source of the flatness. It is not just that the afternoon felt thin. It is that the person who sat at the desk for three hours is not entirely sure what they now know
that they didn’t know before. There’s a kind of loneliness in that — like a radio playing in the next room, present and producing sound, but not quite reaching you. Research on career satisfaction suggests that when we don’t feel genuinely engaged in our work processes, we can experience a form of professional disconnection that resembles the hedonic treadmill effect. What does the flatness actually reveal? The vindication in naming this correctly is not small. Because if you’ve been calling it burnout, you’ve been trying
to solve the wrong problem. You’ve been resting when what you needed was friction. You’ve been stepping away from the screen when what your brain actually wanted was to be put to use. This doesn’t mean the answer is to abandon AI tools, or to perform difficulty for its own sake. It means learning to notice which tasks are worth the friction — and protecting them. The first draft of something you care about. The thinking you do before you know what you think. The
argument you need to make in your own words before you’ll actually believe it. These are not inefficiencies to be optimised away. They are the places where the thinker gets made. There is a real skill, emerging now, in knowing when to reach for the tool and when to stay with the resistance a little longer. Not as a moral position. As a practical one. Because the brain that never struggles begins to lose its sense of what it’s capable of — and that loss
is quiet enough that you might not notice it until you reach for a thought and find the shelf unexpectedly bare. The flatness after a ChatGPT afternoon is not a sign that something is wrong with you, or with the tool. It is a sign that your brain is honest. It knows the difference between output and thought. It is telling you, in the only language it has, that it wanted to be in the room for this one. That is worth listening to. Not
with guilt. Just with a little more attention to what you’re handing over, and what you’re keeping. Like those who maintain cognitive engagement as they age, preserving some mental challenges for yourself isn’t about nostalgia — it’s about maintaining the relationship between effort and satisfaction that keeps thinking sharp. The document is still good. Close the laptop. And tomorrow, maybe, leave one hard thing for yourself to do the slow way — the wading-through way, the crossing-things-out way. Not because it’s virtuous. Because you’ll remember
you were there.
ChatGPT, AI tools, burnout, cognitive friction, authorship, productivity, mental engagement, career satisfaction