In 2010, straight men were embarrassed to pay for premium memberships or features on online dating because doing so was admitting you were desperate.
Now in 2021, I am just starting to have female friends that are subscribing to hinge and having that same internal conversation about desperation and struggle with their ego that me and my straight male friends had with ourselves a decade ago in our early 20s when we spent our first $5-10 a month to get unlimited swipes.
People are realizing that spending money on trying to find love is worth it in a world where many spend so much on Christmas presents, restaurants, bars, vacations, etc.
I am just waiting for governments to realize failing birth rates are such an important factor in the economy that it behooves them to try to actual launch their own free services to try to improve matchmaking. Or to regulate these marketplaces more seriously to encourage these services to optimize more for lasting matches than optimizing for monetization and revenue.
I believe Japan is the first to try actually spending government funds specifically on dating apps, both funding dating app startups and building their own, although other countries have been funding different initiatives to encourage people to start families like tax schemes or other monetary rewards.
Machine learning has a lot to contribute, but at the same time it can be weaponized against us by private for-profits that want to return money to shareholders.
But its theoretically possible to try to find more quality matches with actual reciprocated interest with machine learning, and to find other ways to tweak the rules and systems of online dating to discourage people from matching people they aren't actually interested in dating and responding to those they have matched, to decrease the 'grass is greener / what if the next person on the app I swipe on is better than that person I went out with last friday night' phenomenon, and solve all the other cultural problems created by commoditization of dating prospects, etc.
I think Hinge was doing some interesting stuff with ML that was giving it a tangible feeling of surfacing better potential matches than other services, but with their acquisition by Match Group and move to put their better matches behind a paywall and encourages spending I assume that's changed (especially since the 'roses' update).
What frustrates me is that there seems to be very little way for users to actually provide feedback to these apps learning mechanisms. When the ML is right, it's great. But when the ML is wrong, it's absolutely awful.
For example, to report back 'the matches you have been providing were appealing to me because of A, B, and C, but were unappealing to me because of E, F, and G', where the particular feedback factors could be 'physical appearance', 'job', 'political beliefs', 'children or family goals', stop showing me people with terms like 'poly, ethically nonmonogamous', or alternatively start showing me more of those people, etc.
I have had friends here who get nothing but handsome but unemployed men showing up in their feeds which for them was a dealbreaker whereas I don't care much at all about my partner's job. And I myself have 'don't want kids' as a boolean flag in my hinge profile but because I once added 'don't want kids' as text in one of my bio prompts, I was getting tons of people who _do_ want kids in my feed. I think the ML algo just saw the text 'kids' in the text part of my bio and ignored the 'don't want kids' boolean part. Myself and my friends have noticed things like if we stopped swiping based on attractiveness and started only looking at job title and for a couple weeks swiped right on people with higher status jobs even if they weren't our type, we got shifted into what seemed like entirely different archetypes and categories of people being sent to us. Like suddenly instead of my female friends getting cute skater boys they were getting wall street traders that had never even been available or visible to them previously. Which raises the question of - is it possible for the dating app ML and our swiping to put us into local maximums (based on our last 1000 swipes) but global minimums?