For a stretch of my career, I worked as a User Experience designer. At the time, UX had not yet stratified into specialized subdisciplines, so my role required me to do all manner of design, behavioral psychology, and research. My team and I spent our brainstorming and strategy sessions analyzing what made people click, scroll, and habitually re-open an app, then systematically building those triggers back into products on purpose. Of every methodology we tried and tested, instant gratification was the trick we returned to most frequently, because it worked, and it worked well.
Nir Eyal named this pattern in Hooked, which can be distilled as: a trigger sets off an action, and the action (which can be as basic as scrolling) yields an immediate payoff. This builds up a person’s craving—not just for the reward, but for a way to alleviate the discomfort of the craving itself. This cycle is what keeps a person coming back.
You can see how pervasive this finished model is when you know how to identify it in the wild. Food delivery apps let you feed yourself quickly and without expending energy; dating apps can dispense a type of validation; endless video feeds cure boredom on demand; sports betting provide an affirmation of an intuition—and so on.
One could argue that Google pioneered early versions of instant gratification in UX as it trained a generation to expect an answer within a second. Type anything into the box, and something comes back immediately. But what came back was not the answer itself, but several possible answers. There were multiple URLs that might have what you sought, and you still had to read, compare, and decide. Chat-based AI has managed to deploy instant gratification in the sorting, comparing, and consideration steps. It serves you the answer with fierce assurance, instantly. Chat’s certitude is the same payoff instant gratification had already trained people to expect, just moved off product interaction and onto knowledge delivery.
Smarter by Proxy
Before Chat, that kind of confidence came from a person, and a person was easier, and more fun, to trust than a page of search results as text. Influencers earned it the old-fashioned way, through cultural permission and enough real experience behind them to back up their claims. The clear-skinned woman speaking credibly about skincare, the single parent stretching one income four ways sharing how to budget, the guy who filmed his own hundred-pound weight loss explaining which diet actually worked. It read as more real, more human, because a visible person was saying it and could show it to be true.
That kind of trust takes time to build. Followers and fans assemble slowly, and nurturing an audience is quite the public affair. As the ~403 million terabytes of data created and consumed across the internet each day deploy design that rewards frequent task-switching, attention spans have shrunk, and the influencer’s slower method is not nearly as viable. As a result, platforms and creators have compressed education into entertainment instead (I refuse to refer to it as “edutainment”). Short-form video has done this so well that it is no longer unusual to know a show, a podcast, a book, an article, or a movie wholly through clips, without ever consuming its entirety. What people walk away with is thefeeling of having learned something, which is rarely the same ashaving learned it.
Chat takes even that last piece away. It delivers the shape of having learned something, minus the face, the earned permission, and the editorial judgment about what was included and what was excluded. Nothing human stands between you and the answer.
The Illusion of Hard Work
Three books I was reading at the same time—Searches, Hyperpolitics, and Being Human Is an Occult Practice—touch on human thinking and behavior in different ways for vastly different points, but all somehow elicited the same question: what does one actually pay for information that costs nothing to get?Computers in Human Behavior published a study, “AI Makes You Smarter But None the Wiser” which asked a part of a similar question.
In short, a couple of hundred people worked through a batch of LSAT logic problems. Half used ChatGPT, half worked alone. Among the unaided group, the usual Dunning-Kruger pattern held, where the worst performers thought they’d done best, and the best performers doubted themselves. Add ChatGPT, and that pattern disappears. Everyone, regardless of actual skill, came away more confident than their results warranted, and the people who called themselves most fluent with AI were the worst judges of how they’d actually done.
What using Chat produced was a sense of having already done the research. This is a harder mindset to correct than plain ignorance. Someone who knows they don’t know something is still looking for an answer. Someone who already feels informed has no incentive to keep looking.
This is a different kind of instant gratification loop, one using a false sense of competence as the dopamine. The psychological reward of feeling competent is delivered instantly, bypassing the actual cognitive labor required to earn it.
Competence requires knowledge built through inquiry, experience and honest correction over time. I’m pessimistic about the country having much of it. Roughly 130 million American adults read below a sixth-grade level, according to the most recent Survey of Adult Skills. That gap existed long before generative AI did. What’s new is how confident everyone feels despite the deficit.
The End of Inquiry
Friction is precisely what insta-information is eliminating, yet competence can’t exist without it. Real learning requires sitting in discomfort long enough to be changed by it—a process that demands the honesty to be wrong and the courage to accept that you may never be fully right. It forces existing beliefs to bend or break, sometimes in exciting ways and sometimes in disappointing ones. You learn as much from what a thing isn’t as from what it is.
Competence is not free, nor is it instant. The same artificial confidence used to bypass a simple research task works just as smoothly when applied to your personhood, whether it be your identity, relationships, career, or finances. If you rely on this shortcut long enough, it ceases to be a temporary quick solution and becomes what you are built from. A person built on error and honest correction learns to earnestly question ideas and stay curious. A person built on fast, unchallenged answers never has to learn either skill.
In Thus Spoke Zarathustra, Nietzsche pictures a people who have removed every hard thing from life and call the result happiness. But the people can no longer answer real questions about love, about creation, about longing. “We have invented happiness,” say the last men, and give an odd blink. On the surface, the piece reads a bit quirky, but diving into the themes, you find it’s a description of people who no longer have any reason to think for themselves.
That is the exact failure Kant named in his essay, An Answer to the Question: What Is Enlightenment?: self-incurred immaturity. It is the choice to let someone or something else do your thinking the moment it becomes easier than doing the work yourself. Kant blamed laziness and cowardice, not incapacity, writing that it is comfortable to let a book do your understanding, a pastor tend your conscience, or a doctor manage your body so you never have to exert yourself at all. Kant wasn’t speaking of legitimate expertise, but the habit of abandoning our own discernment for the sake of convenience. His answer was Sapere aude—dare to know, and have the courage to use your own understanding without anyone else doing it for you.
Letting technology do our thinking feels like a reward—an instant shortcut to feeling competent without the labor. If the architecture of instant answers masquerading as knowledge has conditioned us out of our courage to know or understand, we can reclaim our own minds by choosing the friction on purpose. Choose to sit in the wonder of not knowing. Build true competence by daring to do the work ourselves, and find the courage to think.