@futurebird Before the internet I had the notion to provide a dating service for weirdos. I wrote up a long questionnaire with questions about creativity and special interests. I hung up flyers all over Minneapolis, got a P.O. box, and left stacks of the questionnaire at coffee shops. I got six respondents, none of whom were a match for each other, and the project stalled out there. Part of the problem was I had no money and no outreach skills of my own.

@clayfoot

He's just saying the new buzzword. Matchmaking isn't as complex as people want to make it. The most important thing is that the people want the same kind of relationship. So they have a shared vision of being together that makes sense.

"What would it be like to take care of this person when they are sick? What will it be like to find their dishes in the sink?" These questions are key.

It's pretty normal to want a partner, they won't change who you are or solve your problems.

@futurebird

Gen ai (as in generative AI) can't do anything and is just a buzzword here.
General AI would certainly be able to perform well, but it doesn't exist (and we have no idea if it ever will exist).
AI in general may be helpful though, provided it gets enough data.
I mean, a lot of data, as to be able to point to patterns that a human wouldn't see.

@clayfoot

@lienrag @clayfoot

But, it already has ... all of the data possible. I think people don't understand just how huge "training data" (train is a poor word, it should be just called "data" or "data to be optimized into a relational fine-grain net with a ton of computing time")

@futurebird

Oh no it doesn't.
Not about the specific people it needs to match, and probably not about previous successful matchmaking.
At least not until it can get a lot of personal data that is not voluntarily provided to dating applications.

@clayfoot