The Swipe Is on Trial, and AI Is the New Pitch

The guy who built swiping now says swiping is the problem. Justin McLeod left Hinge and raised $18 million to build something that pointedly is not a dating app. Meanwhile Hinge itself is adding features that lean on your friends, not your feed. The whole category is rewriting its core promise in real time, and it's doing it while users are openly fed up. Let me catch you up.

The founder is selling against his own playbook

McLeod raised $18 million for Overtone, a voice-forward AI service that makes "only the introductions that are worth making." His line is blunt: "Overtone is not a dating app." No profiles that reduce people to stats and photos. No algorithmic feeds. No juggling ten chats at once.

Read that again. The person who normalized the swipe is now framing the swipe as the thing that broke dating. And Match Group, which owns Hinge, is helping fund the company built to replace that model. That is not a small hedge. It's the incumbent paying to see if its own product design was the mistake.

Burnout is the real product problem

The pivot isn't happening in a vacuum. A Forbes Health survey found 78 percent of dating app users felt burnt out, spending about 51 minutes a day on apps without much to show for it. Mashable's own reporting had single women calling dating in 2026 "bleak".

So the bet across the category is the same: kill endless choice, hand the sorting to AI, deliver fewer and better options. Ditto AI sets up a single date instead of a flood of matches, skipping the scrolling and small talk. Overtone's slogan is "all you need is less." The product thesis is that too much choice is the disease, and less is the cure.

Where AI wins trust and where it loses it

Here's the split worth watching. When Bumble's Whitney Wolfe Herd announced an AI matchmaker, the reaction online was "largely negative." People don't want to hand the intimate part to a computer. TechCrunch found the same frustration with delegating conversations to machines.

But look at what McLeod is careful to say. AI narrows who might be a good match, and it "transparently explain[s] why we believe someone is a great match." That's the line. Users seem fine with AI doing the sorting behind the curtain. They recoil when AI writes their messages or fakes the connection. The trust breaks at the moment the machine pretends to be you.

Hinge's answer is people, not more algorithm

While McLeod goes all-in on AI, Hinge is testing the opposite lever. Friend's Take lets your friends add a text, voice note, or image to your profile. The data behind it: 79 percent of users already lean on friends and family for dating advice, per Hinge's internal numbers.

Relationship scientist Logan Ury framed it plainly: "Your friends see you clearly, often more clearly than you see yourself." So one company bets a model can know you better, and another bets your friends already do. Both are chasing the same word, authenticity. They just disagree on where it comes from.

The deep cut

Notice what Overtone is actually selling. Not better software. It's selling matchmaking, the human, high-touch kind that used to cost thousands of dollars, now priced for everyone. The AI is the delivery system, not the pitch. The pitch is curation and a reason for every match.

That's the move to steal. If your product drowns users in options, adding more AI to the pile won't fix the burnout, it'll deepen it. The winning play is fewer choices, each one explained. And be surgical about where AI shows its face. Let it sort and reason behind the scenes. Do not let it speak for the user. The Bumble backlash and the Overtone framing are the same lesson from two directions: trust survives when AI helps you decide, and dies when it pretends to be you.

Three questions for your team

1. Where in our product is more choice actually making users feel worse, and what would it take to cut the options and explain the ones we keep?
2. Every place we use AI, can the user see why it made a call? If we can't explain the recommendation in plain language, should we ship it?
3. Are we using AI to replace a human signal our users already trust, like a friend's opinion, when we should be surfacing that signal instead?