You can't game your way to a real education

The nyt ran a headline last month that said you can't game your way to a real education and I haven't read the article but I don't really need to because the headline is doing all the work and it's also the entire thesis Lexie has been built on since the first version of it existed on my laptop.

Most edtech right now is built like a slot machine with a learning skin on top, where the kid taps something, hears a sound, sees a coin, sees a streak, sees a little animated character do a little dance, and then moves on having learned approximately nothing but feeling like they did. The parent sees the streak and feels good. The kid sees the streak and feels good. The app retains everyone. Nobody is learning anything but the vibes are immaculate.

The part that gets less attention is what those reward cycles do to a kid's brain over time. Every tap, every chime, every small dopamine hit, the brain adapts to expect the next one on roughly the same schedule, and the window of tolerance for anything that doesn't deliver on that schedule starts to close. A kid trained on six second reward loops cannot sit with a hard question for a minute. Not won't, can't. You rewired the thing. And then the same kid gets handed a chapter of a textbook and we act surprised when they bounce off it.

The boring truth that almost no edtech company wants to put on a landing page is that learning is supposed to feel slightly uncomfortable, because the discomfort is the mechanism. When you sit with a question you don't immediately know the answer to and your brain does that small frantic search through everything you've ever read, that search is the thing that builds the memory, not the answer at the end of it.

Spaced repetition works because it forces the search to happen at the exact moment you were about to forget. Active recall works because typing the answer yourself is harder than recognizing it in a multiple choice. Desirable difficulty is called desirable difficulty because the difficulty is the desirable part, which is a sentence I wish more people building study apps had read.

The leaderboard part of this is older and somehow still everywhere. Ranking kids in real time produces kids who optimise for the rank, which sounds obvious until you notice the small horror underneath it, which is that the metric becomes the subject. The kid isn't learning french, the kid is learning the french leaderboard, which is a related but distinct discipline with very few use cases outside the app. And the social cost is the part that doesn't show up in the metrics at all, the kids who fall behind on the game scores quietly deciding they're the kind of kid who isn't good at this, which is a story they then carry around for the rest of their education.

You cannot shortcut this. You can put a cartoon fox on top of it and the fox will not help. The fox is in fact part of the problem.

Lexie is built around the part most study apps have spent a decade engineering out. You take a photo of your notes and the app gives you back the kind of questions a good teacher would, the ones you actually have to think about, and when you get something wrong you don't get a sad noise and a retry, you get a hint that nudges you toward the answer without handing it over, and then you try again. When you write a half formed explanation of how mitosis works the app reads it and tells you what's missing or unclear, not whether it was a good effort. There are no streaks. There are no coins. There is no anthropomorphic animal anywhere in the product threatening to cry if you skip a day.

The kids who actually use it aren't having fun in the way edtech has decided fun has to look. They're having the other kind, the kind where you close the app and realize you understand something you didn't understand twenty minutes ago, which is quieter and doesn't make a sound when it happens but is the only kind that matters. Parents have started picking up on this, which surprised me a little, but the family annual plan has a churn rate of zero so somebody is noticing something.

None of this is accidental. Gamification is not what happens when well meaning people accidentally make learning fun. Gamification is what happens when product teams optimized for retention figure out that real learning is bad for retention. A kid who actually understands fractions closes the app. A kid two flames away from a ninety day streak does not. The loop is engineered around the fact that education, properly done, ends. You learn the thing, you leave. Apps that need you to come back every day forever cannot afford for you to learn the thing.

The actual cost of all this doesn't show up while the kid is still in the system, it shows up later, when they hit the first thing in life that isn't gamified, which is most of life. An open ended college essay. A job nobody is grading hour by hour. A hard conversation. The question of what to do with the next ten years. None of these come with a right answer to tap and none of them award a streak for showing up, and a kid who's been trained for a decade to expect both arrives at eighteen with no equipment for any of it. They got good at the game and not at the thing the game claimed to be teaching.

The only test worth running on a study app is whether the kid still knows the thing once the app is closed. Most of these products cannot pass it, which is why the streaks exist in the first place. Take the rewards away and there is nothing in the kid's head to bring them back, and that is the one fact every one of these companies has organised the entire product around not letting parents notice.