How would you redesign dating apps?

Completely random, I know.

My main experience is with Tinder, and I find it interesting how it clearly recognises the shortcomings of the "Match" model, but uses it as a way to get money from users.

Want to signal to somebody that you are really into them? Pay for superlikes. Want to see people who liked you so you can see if it's worth giving them a chance? Pay for Tinder Gold. Are you not having much luck with matches? Pay to be shown more at the top of the queue. It's even worse if you are aware of the internal mechanics of how the queue prioritises people through likes that lead to matches, so you're incentivised to "economise" in your likes and constantly consider if handing out a like to somebody you're unsure about could reduce the amount of people you're shown to. And I'm not even getting into the whole gender imbalance issues or moderation.

I'm also wondering how one could develop an open source alternative, more so for reducing/removing the profit motive from the system design, than necessarily free software values. I think it would make a lot of sense to let people see who liked them as it's an effective self-selection tool, but also keep the shuffled queue as a "discovery" feature. A search feature based on self-assigned traits could be cool too, I guess. I also think a more radically different feature involving mutual friends would be cool, but I don't have anything concrete on that front yet.

Though it's a common excuse by corporations, I don't think giving more power to the users present any significant privacy concern. Everything in a Tinder profile is already public in some way, and security by obfuscation is never a good tactic.

And finally, I think the system should normalise unmatching. In real life, it's very common to find somebody, talk for a bit and lose interest before any dates. Tinder makes unmatching feel like a failure rather than just a normal part of the dating experience.

Edit: elaborating a bit on the open source bit, since it's more technical. Making the client open source but maintaining the exact same design and network architecture does not necessarily change either the maintenance costs (requiring some form of cash influx) nor the control (as the server API defines the available features). Originally I thought about making it self-hostable, but I'm now wondering if it'd be possible to connect users over a p2p protocol. One benefit is that inactive users would be filtered out automatically, but it'd be too dependent on either people being online, really good distributed caching, or a lot of background battery and network usage. I'm not well versed in mobile development or less mainstream networking to even provide an educated guess on the feasibility of this.