It would look like Web2 website.

I don't think deepfakes is more problems than spammers.

Cultural and regional diversity without stereotyping.

A lot of questions.

No data-driven matching, just let me set the location and show me who is online and who was online recently. Please no funny business here.

Just don't be a jerk.

There's a joke that everyone who tries to solve their own problem makes a dating app. I think that's half of the problem - the people making the apps are amateurs, while the happily married people don't feel a need to make them.

Catfishing has always been around. AI doesn't make it much worse. It's a temporary problem.

Compatibility is overrated IMO. I'm probably incompatible with my wife on many levels. I don't know a lot of people who are fully compatible with their spouses. If he's not abusive, if she's not manipulative, it's probably going to work out.

Part of the romance is wanting to make it work despite the obstacles. If I could optimize for anything, it's for platforms matching people willing to make it work. Coffee Meets Bagel does a decent job at this by forcing the two to not talk with others. Muzz goes even further and lets you meet the dad to propose marriage.

"Human relationships and expectations have evolved a lot in the last decade"

Is this really true? I agree they've changed some, but a lot?

In any case, I'll take a stab at your question. Not thinking about how to make money, just how to make something successful from the users' point of view.

I really enjoyed OKCupid in its heyday, circa 2010. I only got a few dates out of it, but the site was genuinely fun. Profiles were long, with multiple essays, and this was very much encouraged. The personality tests were also fun -- and the founders discovered that questions like "Do you enjoy horror movies?", "Have you ever been to another country by yourself?", and even "Do you like the taste of beer?" predicted compatibility much better than trite questions. There was no notion of a "match", messaging worked like email and you could write to anyone on the site.

As the founders said, "OKCupid does a fantastic job at matching you with people who claim to be what you claim to want." I loved the tongue-in-cheek humor and the site's awareness of its own limitations. And you had to put in a little bit of effort to use the site, which was very much a good thing.

OKCupid still exists, but it's a shadow of its former self nowadays.

How to improve? Maybe a slight barrier to engagement, such as a few tailored questions you have to answer for each person before messaging someone. (To prevent men from sending out messages to every women in sight. Or at least make it more difficult...)

Or, something like, if you look at a profile, then you can't navigate away for at least twenty seconds. Again, to prevent you from messaging too many people, or dismissing lots of people instantly.

I believe that the biggest issue with dating sites is that it's hard to align the incentives of the user with the incentive of the site owner. Not sure how you solve that one.

OK Cupid was the best in its approach to match people based on how they answer various questions. The problem here is that every dating web/app always turns into Instagram anyway because digital "dating" is all about looks and nothing else.

Another problem with OKC approach is that you will run out of people you match with pretty quickly. This is why people still use Tinder, even after they liked everyone on the app - there are always more people profiles to go through.

So the business aspect for something meaningful if very limited.

I work on social apps. The central issue with dating apps is incentives. The incentives to form long-term relationships go directly against profit. The incentives for the individual go against what's best for the group. There's a lot of prisoners dilema issues.

The only way I think to have a dating app that works is to have it run by non-profit with rational matching algorithms not geared towards profit maximization.

I'd give the app four walls and have it host a variety of events.

Idk if any use these today because I haven’t used one in years, but I’d do things that improve trust and quality:

- high cost

- human verification, background check, personality assessment

- mandate profile “about me” video

- “things that interest me” section on profile. Could be funny videos, memes, books, shows.

- have in-person mingling events

- monthly 1:1 coaching

Would launch in one city first, heavily advertise, be selective on who to let in and only launch once there’s enough people to open it up.

Only have a certain amount of people in at once, let in new people as existing couples leave. People that are too selective get kicked out.

It wouldn’t matter what you used you would have the same problems - it would become a sausage fest where if you aren’t over 6 tall, you wouldn’t get any matches.

Dating apps are 60-80% men from my brief cursory research.

Before someone thinks I’m an incel, I’ve never used a dating app and have been married for 15 years.

Have you seen the black mirror episode; Hang the DJ? Exactly like that.

Something that bricks your phone or computer after you sign up. Nothing about cyberspace will help you meet anyone.

Like Meetup.com

Controversial but I wouldn't use photos/media at all. At least not as the "main" feature where the actual metric is user engagement, not outcomes. I've worked on a massive, successful dating app before, so I have a bunch of ideas. I also created an app 4-5 years ago but shelved it due to COVID lol.

The challenge is making money on outcomes. There's more money to be made in ongoing care than complete prevention, for example.