The Dispatch

Continued from Page One.

When the largest app in the category spends its first keynote in fourteen years rebuilding toward intent, events, and structure, that is not competition news. That is the market conceding the thesis. We thank the flagship for the validation, note that a $60 million product-development budget was spent arriving at our founding premise, and return to work. Launch is coming. The premise ships whole down here.

LOS ANGELES — This city produces two things in surplus: events and people with no one to attend them with. The show is Friday. The group chat has been "so down" since March. And at T-minus four hours, the texts arrive in their ancient sequence — the sitter fell through, Mark's exhausted, next time for sure — leaving our citizen with a spare ticket, a full heart, and the special arithmetic of deciding whether the thing is worth the specific loneliness of attending it solo.

The Dispatch has covered the Third Wheel's plight before. What deserves its own spotlight is the fix. Event Friends is a MOOD — a declared intent — for exactly one thing: I want company at events. Not romance (there are MOODs for that). Not a best friend by Tuesday (there's a MOOD for that too). Company. For the concert, the crawfish festival, the game, the museum's late night, the comedy show that everyone's "so down" for in theory.

The mechanics are almost embarrassingly direct. Event Friends isn't a grid of faces; it's a grid of events — the shows, festivals, and games inside your radius. Find the one you're already eyeing (or post your own), join it, and the event's own group does the rest: "Two tickets, Friday, doors at 8 — who yells the lyrics?" The event gets attended. The extra ticket gets used. The edge of the crowd goes unstood-upon.

And note what the declared intent quietly deletes: the is-this-a-date fog that poisons every attempt to do this on a dating app, and the flake-rate that kills every attempt to do it in a group chat. Both parties opted into the same plain premise before contact. The only remaining question is whether you're a sing-along person, and that, dear reader, is exactly the kind of question a first meeting should be spent on.

PORTUGAL — Slate ran an essay this January observing something remarkable: on the apps, visible imperfection has become an authenticity signal — proof of humanity in a sea of AI-smoothed profiles. The supporting data, from the industry's own Singles in America research: 61 percent of singles believe profiles have become less authentic. Add the Norton finding that most daters suspect they've chatted with a machine, and you have a market where the products of the platform no longer trust the inventory of the platform. In any other industry we would call this a recall.

As an engineer, I resist the moral reading — that daters became liars. Systems produce the behavior they reward, and the incumbent system rewards performance. Your profile is not read by a person first; it is scored by a ranking model, which decides your distribution, which determines your options. Rational users therefore optimize for the model: the tested photos, the algorithm-friendly prompts, and now the AI-polished bio, because the AI knows what the ranker likes. Each user optimizes; all users converge; the catalog becomes, as the Slate writer put it, identical prospects in different fonts. Authenticity didn't die of dishonesty. It died of convergent optimization. Everyone climbing the same hill arrives at the same summit, wearing the same smile.

The fix cannot be a "be yourself" campaign atop the same ranking stack. The fix is changing what the system rewards. On MOOD, distribution is not won by performing — you are visible to your declared MOOD, full stop. There is no ranker to seduce, so there is no summit to converge on. And the first message is an Offer — a plan, not a performance — which is very hard to fake, because its content is a Tuesday.

When the machine stops grading the mask, people stop wearing it. This is not idealism. It is incentive design.

LAFAYETTE — Every pitch meeting, someone eventually asks about our decline button. Why make rejection a first-class feature? Isn't that... negative? And every time, I give the same answer: the decline button is the most pro-human thing in the product, and I will defend it to the last slide of the deck.

Consider what the incumbent architecture does with "no." It has no vessel for it. A user who isn't interested has two options: compose an unsolicited breakup letter to a stranger, or vanish. Millions choose vanishing — not because they are cruel, but because the system priced honesty at an essay and silence at zero. The result is the ghosting epidemic this paper has covered at length: the majority of daters wandering the same haunted house, each of us both ghost and ghosted, everyone tired.

Now consider what a structured no does. An Offer arrives — trivia Thursday, coffee Saturday. It can be accepted. It can be countered: not Thursday, but how about Sunday? Or it can be declined — one tap, no essay, no cruelty, no three-week fade, no midnight forensics on the other end about what went wrong. Both parties reach a known state. The account closes. Everyone keeps their dignity and their evening.

People sometimes hear "intent-first" and imagine a colder product, all transactions and terms. They have it exactly backwards. The coldest thing in modern dating is the silence — the unanswered message aging in an inbox like a letter to nowhere. Warmth is not the absence of no. Warmth is the presence of an answer. A community where no is sayable is a community where yes means something.

We built the decline button first and the confetti second. I'd do it in that order every time.