Apps for meeting nearby keep dying.

Every few years somebody builds the app that helps you meet the people physically around you, raises a spectacular round, and dies. Color raised $41 million before launch in 2011 and was gone within two years. Highlight was the app of SXSW 2012, faded as soon as the conference crowds went home, and was eventually acqui-hired by Pinterest. Yik Yak owned American campuses, was valued around $400 million, and shut down in 2017. Foursquare retreated from consumer discovery and rebuilt itself as a location-data company. The pattern is consistent enough that investors gave the category a name: a tarpit. Smart people keep walking in because the idea seems obviously good, and the tar keeps winning.

Our flagship product is an app for meeting the people physically around you. So the graveyard is not an awkward subject for us. It is the most useful document in the category, and we read it before we wrote a line of code. What we found is that the standard conclusion, proximity social does not work, is a misreading. A verdict on the category lets every company die of the same vague cause, and nobody has to say which door each one actually died at. Read separately, the autopsies show three doors.

The first door is trust. Every app in the graveyard ran on a radius: GPS said a stranger was somewhere within a few hundred meters of you, and the product asked you to treat that as a social fact. But a radius is a rumor. Nothing happened between you and that person. You never saw them, they never saw you, and the map cannot tell the difference between someone at the next table and someone three buildings away. Acting on a rumor feels random at best and unsafe at worst, so people mostly did not act, and a discovery app where nobody acts is a list of strangers. The fix is not a better radius. It is replacing the guess with a fact: you appear only where you actually are, to people who were actually there, and the moment your path actually crosses someone else’s becomes a record you can both act on later. Verified presence turns who was near me from a rumor into a reason.

The second door is density. A city-scoped feed needs something like a million users before it feels alive, and no consumer app survives the years of empty screens it takes to get there. This is the door most of the graveyard actually died at: not because nearby was wrong, but because nearby was scoped to a city. Scope it to a room instead and the arithmetic flips. A bar needs thirty people in it tonight. A conference needs one hall. Density and intent arrive together the moment you walk in, because everyone in the room chose to be there, and the product is useful at a scale we can actually bootstrap, one venue at a time.

The third door is consent. The graveyard apps broadcast people to strangers by default, and users felt it immediately: the product read as surveillance with a social skin. The industry treated that as a messaging problem and wrote reassuring policies. It is an architecture problem. Either the system is built so that nothing identifying is ever broadcast, every field of a profile is disclosed by the owner’s choice, and blocking someone removes you from their world completely, or it is not. No policy page rescues an architecture that creeps people out. Consent is the difference between a room that feels like a party and a room that feels like a lineup.

Three doors, three decisions. But there is a deeper reason the graveyard generation could not have won even with better decisions, and it is the reason the timing is different now. Everything those apps showed you was self-report rearranged: profiles, check-ins, statuses. The signal that actually decides whether two strangers connect never entered their systems, because it could not be measured outside a laboratory. That changed recently. In 2022, researchers ran real blind dates at real events and measured everything; visible behavior predicted nothing, and what predicted attraction was two bodies physiologically in sync. And the instruments that can read that signal have quietly gone mainstream: nearly one in three American adults already wears a sensor. The graveyard was built before the honest signal was reachable. We get to build after. That is not cleverness. It is a date of birth.

It is also the deepest difference between us and everyone in this story. Our product has three layers: the room shows you who is open, the encounter keeps the people you actually met, and the third layer, You, is the one we are building next, everyday wearables reading whether it clicked. The graveyard competed on feeds, and a feed can be copied. An instrument for chemistry, and the dataset it learns from one room at a time, has to be built the slow way, by the kind of team that has spent years on exactly this. That layer is why we are here at all. The rooms are the entrance. You is the moat.

None of this makes us safe. The category punishes vagueness, and a list of the old failure modes is not immunity to new ones. We could get the room mechanics right and still misjudge how often people want a room at all. We could win nights and lose mornings. What the graveyard gives us is narrower and more useful than safety: it tells us exactly which mistakes are already spent. If we die, it will have to be at a new door, and we have written the three old ones on the wall where everyone, including you, can hold us to them. The full argument, with the studies and the numbers, lives at the case for our flagship. The doors are open. That was the easy part. Now we have to get people to walk through.