Automated dating noise drove users to human matchmakers
Many users endorse the idea that online dating increases demand for human matchmakers by citing travel advisors and progress reframing problems, while a few dismiss the claim for ignoring economies of scale or merely shifting the grind.
Somewhere between Jevons Paradox and Sturgeons Law (90% of everything is crap) there is maybe something like:
Technology that makes X more efficient increases demand for X, but crap scales linearly with the production of X, stimulating demand for jobs in the category of [cleaning up X’s mess]
@DKThomp Has anyone considered whether the matchmaker platforms are aligned? Like are they optimizing many matches or good matches?
@cceichhorn1 i like jason crawford's spin on this , which is that progress is the replacement of one set of problems with a better set of problems (which is different than the elimination of problems)
@DKThomp Hank Green is a lefty techno optimist, and he described it this way: technology solves problems and creates new ones. Over time, the new problems are smaller than the old ones.
@DKThomp Apparently travel advisors are doing just fine too despite Expedia etc.
A simple model of this could be that people are uncertain about their valuation of the premium service, and experiencing the basic service helps them resolve that uncertainty and drives demand to it.
@DKThomp how do you know the demand for match makers is higher?
@DKThomp We can observe same with Online social media platforms too @grok
@ModeledBehavior @DKThomp More compelling arg here seems to be that the problem is that the platforms newly facilitiate *rematching* if you decline prior matches. Upstream behavior changes when rematching goes from very costly to not very costly.
@ModeledBehavior @DKThomp This is Keeper's business model
@DKThomp This doesn’t seem right at all. More like economy of scale limits competition and profit making requires online dating companies to find new ways to keep users engaged, swiping, paying for upgrades. An even more expensive option of human matchmakers avail to well off.
@DKThomp Having a transparent, efficient, universally applicable process increases the craving people have for bespoke personalization.
I see it in my career in private K-12 education. A good teacher feels even more valuable in an era of YouTube education videos and ChatGPT studying.
@DKThomp For awhile I thought the Silicon Valley VC's insistence that employment isn't going away was just them being snake oil salesmen, trying to neutralize opposition to something that will make them even more rich. But this observation you're making is pretty on point. 🤔
@DKThomp Reminds me of the idea that the wealthy will pay for human centered support while lower income will use the automated technology driven. Like Schwab rolling out AI assisted services while higher income will continue to work with people.
@DKThomp Of course only works if "crap scales linearly with the production of X" is accurate. Demand for cleanup jobs will scale linearly with the production of crap, but if technology really is making X more efficient then why would we expect crap production to scale linearly?
@DKThomp This could also just be a result of the declining social interaction trend and just so happens to coincide with the rise of dating.
@ModeledBehavior @DKThomp They should be non aligned because good matches take two people out of the system, potentialy forever.
@DKThomp I wonder if this may increase demand for human-made art in the medium-to-long term. The proliferation of AI slop images and video will make people yearn for handmade art.
@DKThomp It’s always the same story, tech just shifts the grind instead of killing it.
I've got an article about this exact dynamic in my drafts! Should definitely throw in this example.
WRT technology changes, people see direct costs that will go down, and often make predictions that treat that visible change like that's the only cost category affected. You can see it most cleanly in articles about what the Internet -> crypto -> AI will do to transaction costs.
The problem is treating "transaction costs" as one number and start treating it as a profile of categories. Dating apps torpedoed search costs but inflated screening, measurement, and cognitive costs, and made switching costs low enough to introduce perverse incentives.
Matchmakers aren't exactly what they used to be, and search cost reduction is no longer their biggest value-add (though they take advantage of Internet's reduced search costs). They now sell screening/evaluation and cognitive cost reduction and much better principal-agent alignment. Net market is bigger, people get better matches, but the overall market structure is very different.
Many similar patterns apply wrt AI/Jobs, but also AI shifts a much broader set of costs, and AI sits on both sides of the dynamic (and on both sides of most cost shifts) -- it creates new credence goods to be screened/evaluated *and* does screening/evaluation. It adds cognitive load *and* reduces it. Dating is symmetric between two principals, and matchmaking adds an agent in-between, while employment is asymmetric and principal-agent. Some of AI's cost-shifting will increase demand for things humans are particularly good at, but they'll also increase demand for AI in many of the same areas.
@ModeledBehavior @DKThomp Didnt http://match.com get really good at matching but then found out it made them no money since couples didnt come back to the site