I’d love to start a conversation here about attempting to solve this in a very narrowly defined place as an experiment. Since this is HN, I nominate San Francisco.

My hypothesis is that the bandwidth required to present oneself and assess another person is larger than what fits in the self_awareness -> profile_creation -> dating_app_algorithm -> profile_selection -> personality_model_creation -> model_evaluation pipeline.

We already know that a lot of communication happens non verbally. There are other signals we care about: do this person’s friends like them, etc. There are questions about compatible value systems and one’s relationship to risk, money, independence etc.

If there is a technology solution here, I doubt that it looks like a collection of profiles.

Anecdotally, online gaming and niche reddit communities have a somewhat high success rate at creating friendships. One possible way to use this model is a dating app that lets you solve puzzles with another person real time instead of swiping on profile pictures?

I don’t know. I’m throwing ideas here to get a conversation started.