The Handshake alternative that only does film night

The problem is not that there is nothing on

Nobody spends forty minutes deciding what to watch on their own. The deadlock arrives with the second person, because neither of you wants to be the one imposing. So you take turns suggesting, you turn each other down politely, and an hour later you are rewatching something you have already seen — only tireder, and less in the mood for it.

That is a coordination problem, not a catalogue problem. Swiping fixes the awkward half of it: you each decide privately and the only thing that becomes visible is what you already agree on. Nobody proposed anything, nobody rejected anything, and a title turns up that you both said yes to.

But swiping alone only works if the cards are worth swiping on. Hand two people a generic top-forty and they will say no forty times, and the app has solved nothing. The deck is the product.

What Dunno is

One thing: deciding which film to watch, between two people or a group. No restaurants, no recipes, no holiday destinations. Everything in it exists to close out one evening.

• Starting takes thirty seconds. One of you opens a live session, a QR code appears, the other scans it from Dunno on their own phone. Neither of you has to sign up to do it — you swipe as guests. That is the difference between settling it on the sofa and leaving it for another night.
• You set the night before you swipe. Your streaming services, and a mood — something cosy, something frightening, something that lifts the evening — and the deck is built from that.
• Groups work like couples. Three to eight of you, with a match landing on unanimity or on a majority, whichever you set.
• There is a way out of a tie. The roulette spins through what you have all already liked and picks one. The decision is made and nobody had to give in.
• It is free. No subscription, no premium deck. Nine languages, and you can export or delete everything you have from settings without emailing anybody.

How the deck is actually built

This is the part that separates these apps, and it is the part nobody writes about. When you ask Dunno for a deck, it does not fetch a popularity ranking and cut it at forty. It scores a wide pool of candidates for the specific people in the session, and then arranges the result deliberately.

• It scores for everyone at once. A card is not ranked for you and then filtered for your partner — every member of the session contributes their own signal to a single score, so what rises is what works for all of you. Two people is just the smallest group.
• One clear no is enough. If any one of you has shown a settled pattern of rejecting a genre, or has a content filter on, the card is pushed to the very bottom of the deck. The evening is not improved by making one person sit through forty horror posters.
• Dislikes weigh more than likes. A genre you consistently reject drags a card down about twice as hard as a genre you like lifts it. Rejection is the more reliable signal, and it should be — you say no far more often than yes.
• It learns from swipes, not from a form. Genre, decade, country of origin and individual actors all build up their own learned affinity as you swipe, and stars you give films you have watched feed a second, slower signal. Nothing you tell it at sign-up outranks what you actually do.
• It uses what you two have already agreed on. A title the other person has already swiped right on climbs sharply, and genres you have matched on before count as proven common ground.
• It explores before it exploits. The first deck of a session is half exploration by design, aimed at the genres you have taught it least about; each following deck explores less and leans harder on what it has learned. That is why the deck gets sharper as the evening goes on instead of looping.
• It refuses to hand you a wall of the same thing. Early decks cap how many cards can share a tone, so the first pass spreads across the range instead of burying you in thrillers.
• It is built for tonight, not for a wishlist. Titles on the services you pay for rise, and with the strict filter on, everything else is gone. Films you have already seen sink. Pre-2000 titles have to be genuinely well rated to earn a place. Cards missing a synopsis are demoted, because a card you cannot judge is a card you cannot swipe on.

Why that beats a generalist's deck

None of this is a secret about how Handshake works internally — we have not seen their code, and neither have you. It is an argument about what an app can afford to build.

An app that has to settle films, restaurants, recipes, wine and travel destinations needs a deck for every one of those. A taste model does not transfer between them: nothing about which thrillers you reject tells it anything about wine. So the sensible engineering decision, the one anybody in that position makes, is to rank each catalogue in a broadly sensible order — popular, well rated, recent — and let the swiping do the work. It is a fair trade, and it buys you an app that also settles dinner.

Dunno pays for the other side of that trade. Because there is only one catalogue, all of the effort goes into it: the joint scoring, the learned per-genre and per-actor affinities, the veto, the exploration schedule, the tone spread, the availability filter. The result is not a better list of films in the abstract — it is a shorter path to yes for the specific people holding the phones, which is the only thing that matters at eleven at night.

It is free, it is on Android and iPhone, and you can try it tonight without creating an account. If you want the longer read on deciding as a couple, it is in the movie app for couples.