why dating apps feel exhausting.

People who use dating apps for long enough describe the same feeling, and the language is consistent across cohorts. Drained. Numb. Vaguely angry. Surprised at how tired they are given that the actual time spent in the app was modest. The exhaustion is not a moral failure or a sign that someone has become 'bad at dating.' It is a structural output of the container, and once you see the three mechanisms producing it, you can predict it.

This is an essay about what the exhaustion is actually made of. I work on Byvibration, which is built around a different container. Toward the end I will say what that container is. Most of the piece is mechanism.

mechanism one: the photo-first sort forces snap judgments at scale

The default surface on Tinder, Hinge, and Bumble is a face. The interface offers two affordances, accept or reject, and they are designed to be cheap. The cost of any single decision is one thumb movement. The expected number of decisions per session is somewhere between forty and three hundred.

The cognitive load this places on a person is not 'looking at faces.' It is 'rendering a verdict about a stranger, hundreds of times, with a one-second budget per verdict, all evening.' The brain doing this work is being asked to summon a microscopic version of the social-judgment circuitry that humans normally run once or twice a week, and run it three hundred times in a row.

The work is invisible while you are doing it. It only shows up afterwards, when you close the app and notice that you cannot concentrate on a book, and you do not particularly want to talk to anyone. You have not lost anything specific. You have just been doing the cognitive equivalent of speed-rating strangers for forty minutes and the system that runs that judgment is now flat.

This is not a metaphor. This is what the photo-first sort, at the volume the apps require to feel populated, is asking the brain to do.

mechanism two: the message inbox guarantees low-signal volume

If you do match with someone on a photo-first app, the conversation begins in an empty text box. The signal that crossed the matching boundary was a face plus a few lines of bio. There is no shared question, no shared piece of writing, no shared object to react to. Both parties are now responsible for generating context from scratch, inside a chat window that punishes pauses.

The honest thing for anyone in this position to say is 'I do not know enough about you to write a real first line, so here is a polite low-stakes one.' The container makes that pattern rational. Three to five exchanges of pleasantries are the substrate that the chat surface guarantees, regardless of who the two people actually are.

Across an inbox of ten or fifteen open threads, this is the load: ten or fifteen conversations that have not yet earned the right to be interesting, each demanding a small but ongoing amount of attention, each one possibly the one that goes somewhere. The user is being asked to keep ten low-signal channels alive on the chance that one of them resolves. The exhaustion this produces is named 'ghosting' in the cultural conversation, but the real source is not the ghoster. It is the inbox shape that made fifteen empty channels exist at the same time.

mechanism three: variable-ratio reinforcement is doing what it always does

The third mechanism is the most discussed and the most accurate. Swipe-driven apps deliver matches on an unpredictable schedule. Some sessions yield none, some yield five, some yield one. The schedule is variable-ratio reinforcement, which is the schedule that produces the strongest, most persistent behavior in every species that has been tested for it, including us.

The match is the reward. The session is the variable-ratio trial. The behavioral system that this trains is identical to the one a slot machine trains. The user is not weak for opening the app a fourth time tonight. The reinforcement schedule is the strongest schedule there is, and the user is responding to it the way the schedule predicts.

The combined cost is not 'I played a slot machine for forty minutes.' It is 'I played a slot machine in a domain where the reward is supposedly the start of a relationship, and that disconnect between the slot-machine feeling and the supposed purpose is the part that hollows you out.'

what follows from the mechanisms

If exhaustion is the structural output of those three mechanisms, then the way to remove it is not 'use the apps better' or 'swipe less.' Both of those advice patterns try to fight the container with willpower. The container wins, because the container is built on the most reliable reinforcement schedule in psychology, attached to the cheapest possible decision affordance, attached to a chat surface that guarantees low-signal volume.

The remedy has to be a different container. Specifically, one that:

• Sorts on something that takes more than a second to evaluate, so the brain is not asked to render three hundred verdicts in a sitting.
• Generates a non-empty first message before any conversation begins, so the chat does not start in pleasantry-substrate territory.
• Removes the variable-ratio match drip by either matching slowly or matching once a day on quality rather than rapidly on quantity.

Byvibration is one attempt at exactly this container. The matching layer cannot see photos at all (we wrote the engine so it would have been a type error for it to try; the code is at github.com/donnowyu/soulmate-core). The conversation surface uses a 24-hour reply cadence (see /essays/letters-mode-is-mercy for why). And the system surfaces a small number of matches a day, on prose and voice signal, not on rapid swipe streams.

We are not the only people building toward this design. There are other slow-dating products, voice-first products, friendship-first products. The shared insight across all of them is that the exhaustion has a mechanism, and the mechanism is fixable by changing the container. The argument of this piece is that anyone who is tired of dating apps is not doing anything wrong. They are responding correctly to three forces that any product built on photo-first swipe at volume is going to produce.

If a different container interests you, you can read more about the architecture at /essays/why-matching-layer-is-physically-blind, the conversation cadence at /essays/letters-mode-is-mercy, or visit byvibration.com.

I work on Byvibration.