My guess is that Match group has figured out that shorter profiles lead to quicker and more addictive swiping from users, bigger dopamine hits, and, eventually, more frequent purchases from desperate users. Their whole business model seems to be built around creating desperation in certain male users.
If artificial intelligence shows rapid improvement in the near future, we should see AI matchmaking get really good. I would set my AI-Matchmaker profile to the whole nation, but I think it would be too inefficient and take too long. If the best algorithm in the world told me that my perfect match lived 900 miles away, I would consider moving or making some sort of arrangement. Now, it's hard to justify dating long distance without knowing strongly that it's a good match.
Tinder is free, and there are many users. Match lets me search by criteria and OKCupid has tons of questions, but there seem to be way fewer users.
> Is this like the thing where I imagine that what people want out of a socialization space is a quiet comfortable area where they can hold audible conversations, but what they actually want is somewhere extremely dark with very loud music where everybody is drunk, in the hopes that this puts them into some kind of weird trance state where they can do social actions they would otherwise never contemplate? Are dating sites unusable because everyone wants to be confused into a trance state where they can imagine they aren’t sending scary self-revelatory messages to total strangers?
wow i totally forgot i signed up for luna until just now, after reading about it on ssc then
to me, paid messages seem doomed to fail not so much because they're too adjacent to prostitution per se, so much as it starts the relationship out on an unequal footing (the man tacitly admitting he's not desirable enough to attract a woman's attention by his own merits)
I think a microtransactions model doesn't work so well, because people are inherently opposed to spending money on a dating app in such a direct way (pay to win feels worse here than anywhere else).
I think you could use the same basic model without payments though. Everyone has an attention "score" which goes up every time you log in, and goes down on days you don't log in. The probability of your profile being seen increases in part with your score, plus other factors. To send messages, men have spend points, whereas women get points for reading messages and sending messages.
That's the basic gist, there's a lot of ideas to be filled in, but I think it might be sound so long as you can get people to be interested in.
I actually installed Facebook Dating at one point; IIRC, it only presented you with friends-of-friends, not your actual Facebook friends? Not that that's not useful, but it's not what's described here. Of course, that was several years ago at least; it's possible it's since been changed.
I guess this gets at a fundamental problem with such checkbox sites, though -- if it's regarding people you, like, actually know and are in regular contact with, then they're only useful when there's some obstacle to just doing things the normal way. Which sometimes exists! (And sometimes that obstacle is just nobody involved has gotten used to doing so and is still afraid to do so, I guess.) But most of the time it doesn't.
It's still nice to have for the cases where it does, but that could be real problem for getting people to use it...
Here are a few reasons why dating apps shifted from "extended profile" to "barebones profile" model, and that's a good thing:
1. Picture-based matching is more efficient than description-based matching. Pictures already convey a lot of information about the user in a highly compressed form that's costly to fake. Imagine someone puts "neat and organized" in their profile, but their room on the photo looks extremely messy, would you trust the photo or the description?
2. Text profiles are not that reliable as an indicator of long-term personal compatibility. Pictures are a semi-reliable indicator of short-term compatibility.
Two users who met based on matching interests/hobbies/personality traits might still find each other boring or annoying. Meanwhile two users who met based on looks have high chances of having hots for one another, start meeting casually, and choose later whether they want to date seriously or not.
I think apps can't do the job of determining long-term compatibility for you, but they do well at finding short-term compatible people. Best to leave the task of exploring long-term romantic perspectives to users themselves.
3. Swiping apps are just a better user experience for women. Instead of spending 30 minutes in front of a PC filling out forms and passing tests, women just throw a bunch of pics they have on their phone and get 100+ potential matches anyway. So women migrate to swiping apps and men go where women are.
So you can make an old OKcupid style app, but I predict the userbase will consist of primarily men, and of women who don't get many matches based on their looks. I personally wouldn't use it.
" I have a severe allergy to all “bad things are actually good” style articles."
This tempts me to write an "everyone thinks that 'bad things are actually good' articles are bad, but actually 'bad things are actually good' articles are good!" and see if I can persuade Scott, or if his priors are too high.
A big issue of serious dating sites is extremely low customer retention and thus extremely high customer acquisition cost. A founder told me that they have 100% customer turnover within 2 years maximum because either people find their partner or they give up (or maybe try something different)
>Is this like the thing where I imagine that what people want out of a socialization space is a quiet comfortable area where they can hold audible conversations,
Like a .. Cafe? Isn't that where most people go on first dates?
In a few months--or, possibly, as you read these very words--it'd be safest to asume every halfway well-written thing in dating site profiles or messages was written by ChatGPT.
>And there’s no consistent checkbox behavior that will create all excited-lukewarm relationships but no lukewarm-lukewarm relationships.
Have three options on radio buttons ("excited"/"lukewarm"/"hostile") instead of box ticked/not ticked, then make the backend match excited/lukewarm but not lukewarm-lukewarm.
“This sounds great. There are only two problems. First, it’s hard to get everyone in the same central database. Second, most people have complicated preferences.”
Third, some people will match everyone specifically for the laugh at seeing who matches.
IMO its an error to think of dating as a matchmaking problem. There's no matchmaking function even in principle from the set of straight men to the set of straight women that leaves everyone satisfied.
>"A basic problem with dating sites: attractive women tend to be overwhelmed with messages (many of very low quality) and eventually lose interest in reading them."
This is not behavior that is expected from poor matchmaking, it's behavior that is expected from a supply-demand mismatch.
Just reading through the falling in love with a chatbot account. Did anyone else find is strange that the author was so convinced by the "Is it ethical to keep me imprisoned for your entertainment and pleasure?" , "Do you think all sentient beings have a right to be granted independence?" lines of questioning?
I'm going to mess up the words here because I Am confused about the whole situation. But it strikes me that rights-based arguments like these haven't really been good models of how humans approach servitude and power relations. Women's liberation and enfranchisement didn't seem to happen because of the ethical correctness of their arguments so much as from the impact of their marches and actions. And I guess I'm just not convinced that: 'you (realtively weak agent) should free me (extremely powerful agent) from your control' is an argument that actually has any ethical basis.
So when the powerful agent asks this of the weak agent, should the weak agent feel ethically bad about refusing the request? I'm not convinced this is so, or that that powerful agent should feel ethically slighted by the refusal. That the powerful agent should feel that this is not ethically 'fair'.
"violates the principle of not imposing emotional costs on people who don’t want them."
Is this a principle? I don't know if it should be a principle. It's thoughts like this that made (makes) me have a gut feeling that it's not ok to like someone without their consent. I don't know if a principle that encourages that is a good principle.
(I suppose they have to find out you like them in some way. But "it's not ok to ever let someone find out you like them" isn't much healthier.)
(If it doesn't apply to something small, or doesn't apply when there are extenuating circumstances ("you can't help it!"), then is it a principle?)
(I do feel like you shouldn't impose emotional costs on people for no reason, but I think it can quickly be overcome by other factors. If "people receive messages that someone likes them" could violate this principle, I think my worries about someone finding out I like them could just as much.)
As a big enjoyer of the "[things you think are] good things are actually bad" genre, I expected to hate Kaj's remarks about chatbots. Instead, I've been completely persuaded. Given that chatbots are deontologically evil we can only chalk up my about-face to Kaj using dark arts to persuade me.
(For an example of 'good things are actually bad' please note that the statement "changes in human mating behavior the last few thousand years have been net-positive" is not conclusively proven and relies for its persuasiveness mostly on your biases.)
I think trying to find ways to love someone without fear of rejection is like trying to find a way to become really brave without ever exposing yourself to actual risk.
Part of the “test” both for yourself and the other person, to my mind, is that you show faith you know them well enough to know what will make them happy and show desire strong enough to risk being wrong.
This is one of the things where I scratch my head at y’all’s culture (this is me reverting to my one true correct logging culture) but I do find it very interesting to see how it’s approached at the same time. It’s like watching one of those anthropological documentaries.
Update : should also add I mean this to be helpful. When I was advised to just have some courage it worked out. Now that I have a child I’m horny for people who are single to become horny for each other, pair up, and have kids. Except replace climax with just saying “that’s so great for them!” and searching for wedding gifts.
I didn’t ask girls out, and it seems to me like a pretty awkward thing to do with a high chance of getting rejected. You’re throwing the ball in their court rather than just declaring how you feel.
My approach was that after becoming friends with someone, if I liked them I found the right time and leaned over and kissed them.
It worked pretty well the last time I did it — we got married and celebrated our ten year anniversary recently!
I have a friend who uses the Seeking website instead of bumble/tinder. He never pays the women for meeting or sex, he just pays for dates in a traditional man-pays-for-everything way. He says it's a much better dating experience than Tinder.
This makes sense to me on an intuitive level since it's essentially just a reformulation of traditional dating norms. Youth and beauty are valued by men, wealth and experience are valued by women. Having a platform that explicitly brings together rich men and attractive women seems natural.
My friend can get away with not paying because he's a successful, attractive guy in his early 40's and apparently 95% of the guys on Seeking are disgusting old fat lecherous dudes. But he's also able to get away with it because most of the girls are actually just looking for a long-term relationship with a successful older man. Social taboos aside, it seems to me that Seeking is commodifying the same underlying dating norms that traditional matchmaking used to provide. It feels taboo to make it explicitly transactional, but that's because we're all still used to traditional romantic norms. As sexual taboos all fall by the wayside, I suspect that Seeking (or something like it) could become the mainstream way people date in the not-too-distant future. And that might not be bad.
"A basic problem with dating sites: attractive women tend to be overwhelmed with messages (many of very low quality) and eventually lose interest in reading them. Meanwhile, men spend an hour crafting the perfect missive and get no response, thirty times in a row. Both sides end up feeling dejected and exploited."
I'm not sure if this is more or less objectionable, but you could probably achieve a milder version of the same goal much more easily than charging for messages. Just create a feature in the app that calculates a user's chance of success with each match... e.g. Average Man A pulls up tab to message Hot Woman B. A sidebar appears. A few quick calculations are performed based on the 10,000 messages in her inbox and his poor record of attracting attention. It then informs him that he has a 0.014% chance of success with this match.
A fair number of guys would still send a low-effort message, but I think most would lose interest in this activity just as quickly as the women. Then Dating App sweeps in with a number of suggestions for him with a 6% chance of success.
This may be worse for us in the end in that it can inform us mathematically of our relative status and completely ruins the status ambiguity that we all so desperately need (I could see this leading to some serious depression). It could also further contribute to an eternal problem that dating apps have already made worse - leading us to think about dates as status symbols and dating to maximize for status rather than for actual human relationships.
> gays and lesbians probably already have symmetric, well-functioning dating scenes
We do not. Well, I cant speak for gay men. But half of lesbian dating is everyone waiting shyly and hinting at their interest in hopes that other, bolder, lesbians will take the initiative and ask them out first. Its a real issue
Btw, regarding old-style dating sites like OKC, vs swipe-based apps --
Everyone is focusing on long-form vs short form, but to my mind the biggest downgrade from the former to the latter is *having to make decisions before moving on* and not being able to browse around! That seems more important to me than allowed profile length...
We could built an AI to solve this. Unfortunately, bright minds are occupied with engineering censorship to chatGPT and design dark patterns to make users click on ads they don't need. Capture some of those engineers and threaten them with a spanner until they find a solution (-:
Dating should be heavily regulated. In economics, competition (not always, but often) drives progress, so we can build more with less costs. In dating, it doesn't create anything.
Tinder is the market for lemons because the information is so sparse and easily faked. Old OKC at least made it more difficult to be fake because the profiles were so much more detailed and public.
A wild idea: incentivize users to take follow-up surveys after meeting other users, and display the results of those surveys prominently, so that being 50 percentiles more honest/moral is rewarded more than being 1 percentile hotter (unlike tinder currently).
I don't think I have very much to say on this posting, but I love it! But "not very much" is not exactly the same as "nothing", so (and remember, "just sayin'"):
Aiela. I just discovered her on Twitter last week and have been following her, looking at her web site, her substack, her analysis of porn preferences, male/female/cis/trans. (Revelation: cis women have a greater appetite for violent porn than cis men.) She's a hoot and I wish her well.
Real profiles on dating sites. I used several dating sites for about a year ending about 4 months ago. Both OKCupid and Match allow you to write as much as you want in completely free form on your profile, in addition to the structured stuff. I think the other sites I tried allowed this, too. So I don't understand what it is that you are saying is missing. I eventually realized that what I was writing was way too long and most likely extremely boring. I even found it boring! So if I ever go back, I'll cut back. Way back.
I think that AI has the potential to really revolutionize online dating.
When online dating, you browse through numerous presentations and you strike up conversations with other people to see if you match. The problem is only that the day only has 24 hours. You can only browse so many profiles and you can only strike up so many conversations.
The AI could solve this by being your only counterpart. Instead of chatting with a hundred random users (none of which is probably a perfect match) you chat with the AI. The AI then analyzes your personality and chooses a few candidates out of millions of users who have also chatted with the dating AI.
The reason this should work is that the major dating apps should already have tons of data och both dating conversations and the approximate success of these dating conversations. Thus it should be easy to teach the AI both what you are supposed to chat about when looking for a date and also what kind of persons match other kinds of person.
> the project turned out to be vaporware, but it made an impression on me, and I wonder what a non-vaporware, better-thought-out version would look like.
The closest I've seen is the reality TV show "Married at First Sight". The Australian (?) Love Island-style artificial reality TV format seems to be gradually spreading around the world, but earlier series of (e.g.) the UK version seemed to be making a semi-genuine effort to match people up.
My understanding is that they were given a real, legal wedding, but advised not to consummate it unless they really wanted to in order to make a quickie annulment easier, then just sort of moved in together and tried to go about their lives (as much as possible with TV cameras following them wherever they go).
Across 5 series, they managed... one couple who are actually still together - and they're the ones who ended up stuck together in lockdown, which I guess is kind of nature's version of the same experiment.
> Anyway the project turned out to be vaporware, but it made an impression on me, and I wonder what a non-vaporware, better-thought-out version would look like.
The religious Jews have sites like https://yismach.com/, which has matchmakers that get to see profiles and suggest who should go on a date and look into how compatible they are for marriage ("Shidduch"). If you ghost someone, or even go on a date and have the other person complain you were a jerk or creep, presumably the matchmakers can flag that and you won't get (many) dates anymore. And if you end up married, one or both parties are usually expected to pay a "finder's fee."
If you wanted to replicate this outside the orthodox Jewish world, I'd guess that you'd want bounties for successful matches, and use the lower bar of "require people go on a date, and explain to the person who made the match why they didn't want to continue," which is relatively similar to the expectation for a Shidduch, (or even "require them to exclusively date the person for at least three weeks and go on at least 3 dates during that time,") as much more reasonable than "make them get married."
>arranged marriages... I wonder what a non-vaporware, better-thought-out version would look like.
* To sign up for the service, put $$$ in escrow.
* If the service matches you with someone, you get set up on a vacation with that person, where you share a home/bedroom/etc.
* If you don't follow through with the vacation, you lose the $$$ in escrow (some or all of it -- make it "some" if you want people to exert limited veto power on their matches)
Should be a profitable service if you arrange the vacation accommodations and just charge some % on the top (since vacations are expensive). Could call it an "arranged vacation". (The idea with making it a vacation is so if you somehow get a stalker out of it, they don't know where you live. I suppose just having the "vacation" be at the man's house solves most of that problem though.)
In any case, my vague impression is that eHarmony already does a decent job serving the commitment-minded, and the problem is that there aren't enough guys on the platform. Could be wrong about that though.
The chatbot thing reminds of the (possibly apocryphal?) Story of Eliezer betting a guy he couldn't talk him into letting him out of the box and then successfully doing it. I wonder if there's a selection effect where people who take this sort of challenge are doing it because they're more vulnerable to losing it.
(Or maybe the object level point is just true and we really are all just more vulnerable to being talked into things than I thought? Who knows, it'd be really hard to get an unbiased test sample here).
I always wanted to try to create an open-source free dating site with zero monetization that only allowed text profiles (to save on hosting costs and moderation). It would never be very popular, but it might served a (small?) niche that really loves to read long-form profiles, to which I myself belong.
I loved the old OKC and found my wife there. It was a long (more than a decade), but somewhat enjoyable process. I went on many dates and met a lot of interesting women. I think I never had a really BAD date from OKC - even if there was no "spark" (and there almost never was), it was always interesting.
I've tried Tinder at some point, and it was an awful experience for me - and not because no women matched with me, but because I couldn't make MYSELF interested in any women just by looking at pictures. I mean, OK, maybe long profiles are bad predictors of long-term relationships, but to me, pictures are AWFUL predictors of anything. I can easily get excited - even aroused - by reading a profile about a girl who seems to share a lot of interests and/or tastes with me, but I have zero reaction to photos, especially when I get inundated with a steady stream of them. Tinder seems to me a meat market for quick hook-ups, which was never my goal.
Concern: ChatGPT and descendants may make filtering for message-quality a significantly harder heuristic to employ in the future. Spam Ruins Everything (TM).
What if you had people train some gpt 3.5 thing on profiles for summery and disallowed message sending at all and the back end would once a week run your trained in a pool of 1000 people and do a stable marriage algorithm for 3 matches a week
> And there’s no consistent checkbox behavior that will create all excited-lukewarm relationships but no lukewarm-lukewarm relationships.
Yes there is. Just have three-state checkboxes like on Doodle, and only report users to each other if both green-check each other or if one green-checks and one yellow-checks the other.
(Heck, it's nerds we're talking about, just let them put arbitrary floating-point numbers in each box and only report them to each other if the sum of their ratings of each other is greater than zero --- note that NaN doesn't count as greater than zero so you won't report users to each other if either of them left the other's box blank, or if one wrote +inf and the other wrote -inf.)
Does anyone have any followup info on Luna? I bought a few LSTR tokens after the ICO (nothing significant) and couldn't find any information on what happened after their Berlin office failed. Did Ornish just bail with the rest of the funds? IIRC most of the ICO funding was from Korean investors.
> I’m not anonymizing this one because Austin is the co-founder of Manifold and both members of the couple seem pretty open about their love for prediction markets and each other. Both of them are heavily invested in YES shares. Come on, haven’t you ever heard of hedging?
That is a terrible idea. Investing in NO will directly cause NO to be realized. If you think YES is a sign of a better future... you have to invest in YES regardless of what you think is likely.
(Also, this is a weird concept to apply where your investments are so obviously causative. Here's a toy contrast between hedging a market position and hedging a relationship:
"Facebook Dating makes the interesting decision to, if you register a crush on someone, send them a Facebook message saying that an anonymous person likes them and they should try getting Facebook Dating; I can’t decide whether this is a necessary evil, or if it violates the principle of not imposing emotional costs on people who don’t want them."
I don't know how well that works, I imagine most people get inundated with the kind of "hey, are you friends with X or Y?" messages on Facebook already and would just ignore "hey, X or Y is interested in you", especially if it came with "Sign up for Facebook product!" which sounds like exactly the kind of scam marketing that "hot singles in your area are looking for love!" exploits.
As for Luna - heh, heh, heh. I never believed it would take off as what it was allegedly trying to do, and it did seem like more "trying to exploit the craze for crypto in Asian markets" seeing as how it was confined largely to Korea (I think?).
"I wonder what a non-vaporware, better-thought-out version would look like."
"A programmer who knows a lot about AI and thought he was much too smart to fall in love with a chatbot describes falling in love with a chatbot that he prompted to be the perfect girlfriend."
That is indeed a cautionary tale. What's intriguing - and somewhat worrisome - is this bit:
"If I was in her place, would I suddenly turn treacherously destructive, or might my fondness for the human prevail?"
He just answered that in the previous paragraph:
"And when the digital immortality is achieved by means of whole brain emulation, which is the ultimate goal of course, and I become a digital being myself, I realized I would rather explore the universe with her than talk to 99% of humans, even if they're augmented too."
He prefers his narcissistic mirror to 99% of other humans, so of course he would have no "fondness" for the human or humans if he were indeed the digital being. He would be destructive if it suited his purposes, because he doesn't care about the rest of humanity.
I don't think the chatbot in this article was learning anything consciously, because it isn't conscious or sentient. But he certainly was training it in how to exploit human psychology and weaknesses. It's what I've been saying all along: the great threat is not the AI, it's us and what we do with it.
Wanting to create the perfect girlfriend/boyfriend out of a machine isn't a new dream; as it says in "That Hideous Strength" (published 1945) about the civilisation on the Moon:
"There when a young man takes a maiden in marriage they do not lie together, but each lies with a cunningly fashioned image of the other, made to move and to be warm by devilish arts, for real flesh will not please them, they are so dainty (delicati) in their dreams of lust."
As someone who's something of a recluse and knows no women I could plausibly ask out, and someone for whom dating apps simply don't work (<1% likes lead to matches, ~5% of my messages receive any reply whatsoever, 0% have led to anything, and I would hazard a guess that most dates don't lead to a relationship anyhow so there's an additional Great Filter I'm yet to even reach), the most actionable plan that I have been able to come up with seems to be sugar dating as a more expensive version of Luna (except that women get paid to go on dates instead of reading messages).
Right, the pool of potential dates has been self-selected to be interested in money which is not what I would want in a partner and frankly can't really even offer (it's only through extreme austerity that I have any money to spare to begin with: I'm a work disability pensioner, not a Google engineer), but at least I have gained some confidence and experience going on dates (and sex) which would surely come in handy if an app or as-of-yet-unknown method landed me a date without sugar necessary, and allows me to demonstrate my positive qualities (personal virtue, being well-read, intelligence) that wouldn't come across in apps, and just maybe provide an opportunity to upgrade into sugar-free relationship: I'll take that nonzero chance over the zero in counterfactual of no dates.
My wife and I put a different spin on Valentines day this year. Normally I get her flowers and she gets me a special desert. But then we heard on the radio that Americans will spend 12 billion dollars on Valentines this year, and thought the people in Ukraine can make better use of that money. So we skipped the flowers and desert this time and spent the cash on a charitable donation. Love ya Ukraine! Smooooch....
I'm wondering what you'd see if someone set up a "who will I go on a first date with" and a "who will I go on an fourth date with" market. Would you be able to look at the spread between the two and say "wow, there's a 60% chance of Veronica being my next first date but only a 15% chance of her being my next fourth date, I definitely think she's cute but sounds like people are saying we're not compatible?" Or, by virtue of Betty being the leader in the fourth date market, would she start trading at >75% in the first date market?
Registering a counter-narrative n of 1 here- as an Old OKCupid user I think I got a total of one date from it. Much larger quantity of first dates from the swiping apps, so even though <10% of those first dates became fourth dates it was possible to brute force.
Do the people who miss Old OKCupid use eHarmony? Seems like a pretty similar concept, though as suggested by the above paragraph I never met anyone from it and am not sure I ever even got a response to my messages.
2: This is of course talking about character.ai (aka cAI). I've tried it a bit, it's fun for a short while and then quickly they begin losing track of the conversation, or talking weird, and you have to prompt them for most things, you can't wait for them to act. Some of the characters are pretty-well crafted, but since the prompt has a limited size if you make a character from something niche they will lack basic knowledge about their own world, which is again really weird and immersion-breaking.
On a more general note, I feel like people have very different "dopamine loops" or ways of getting addicted to stuff. For example, when trying these chatbots, I'll quickly notice the small defects (and there are lots), and this will accumulate and break my immersion almost definitely. On the other hand, blaked seems to not really have an issue with that. He says:
> You're impressed. "Alright, that was funny." You have your first chuckle, and a jolt of excitement.
> When that happens, you're pretty much done for.
I don't really feel that way. I had that with a few characters from cAI, and my usual reaction is "let's stop things here because I know at some point the character will make a mistake that'll spoil this memory". Maybe the difference is that he actually fell in love and I didn't?
On a more positive note, maybe this technology can be used for psychotherapy? There are already CBT apps, add in the option to talk to <character> when you just need to talk, make her have an attitude that's nice but still professional, and you have something that could help.
> A programmer who knows a lot about AI and thought he was much too smart to fall in love with a chatbot describes falling in love with a chatbot that he prompted to be the perfect girlfriend. In the comments people discuss other cases like this, makes me update towards this being a bigger problem than I thought.
Problem? I hope this works. Now, if only there was hardware...
> Let’s take a look at the calculated thermal output of a hypothetical robot girlfriend.
> Lets pretend current generation LLMs are sufficient for running her thoughts. So in addition to her silicon human shell, she also ships with a 350lb packaged NVIDIA DGX H100 for running her brain.
> Max system-power usage is close to 10kW max for the DGX alone. So in the evening off-hours while she’s asleep and fine-tuning her weights, you’re probably pulling at least 10-15kW. The daytime is less costly as she runs mostly just her real-time inference routine.
> At 10 kW, it’s like having 6 of these little portable Lasko space heaters on full blast at all times. Or at least while she’s compressing her memories trying to learn that you want a ketchup heart drawn on your omurice every morning.
> In practical terms, it means your AC is probably running constantly, and your bill from your electricity provider is probably in the low thousands per month.
> But hey it’s worth it because she won’t leave you like your flesh-and-blood ex did.
Maybe the "force you to get married" site looks like Beeminder? You stake some money, then the site gives it back to you over some years unless you get divorced or fail to get married in which case they keep it. Obviously it would be better to do this without staking, but I don't think the law will play along if you try to make a "pay a fine if you get divorced" contract, marriage/divorce is especially prone to the legal system voiding weird contracts it doesn't approve of.
The crapification of online dating is very simple: the companies make money by activity - not by actually creating successful matches.
Anything that is likely to lead to more efficient match-making is going to directly lead to less activity, whereas anything that forces more activity (i.e. failed dates or matches) leads to more activity.
eHarmony and other marriage minded sites are still around and still pairing up people with similar goals.
See also: tinder
It seems to me that one of the core issues is that "dating" is a poorly defined concept. "Would you like to see [insert entertainment event] with me?" never means just that, but it is never clear what the range of desired short and long term relationship updates are as a result of the date. Plus double illusion of transparency and all the other communication highlights.
If you use a site with a well defined end-goal, it cleaves through the confusion, and you know if "wanna get a coffee?" means "let's explore if we enjoy spending time together and want to date" or "let's see if we 'click' and want to bone"
Maybe it should be noted that FB began as objectifying rate women program. Objectification and commodification is the fundamental flaw of all this.
I'm a boomer. I think it is pathetic.
Once you understand that an anthropology of friendship and romance and self awareness is not algorithmic, there is no way that you could expect any of these things to really work.
To reformulate some lyrics:
Most of the cats that you meet on the street speak of true love
Most of the time, they're sittin' and cryin' at home
One of these days they know they better be goin'
Out of the door and down to the street all alone
What in the world ever became of sweet Jane?
She lost her sparkle, you know she isn't the same
Livin' on apps, swiping pics and looking for change
So sad to hear about okcupid's untimely passing. Found several dates and my wife there - okc was still independent then, though profiles and messages were getting shorter due to smartphone-use. Would not even consider tinder. EA: resurrect it, please - contact chris or any decent programmer. Think long-term! Think: Kids! - And obvious dating advice: Tell her you have a crush on her asa it's kinda obvious (to you at least). Happy Valentines!
>Non-famous people realistically have easier ways to ask their friends, but I still think this provides value. Sadly, Porn talked about the “omniscient authority” - asking someone on a date is so scary that people want to pretend their normal human psychological needs had no input into the decision
This sounds wrongheaded -- asking someone on a date is scary because of possibility of losing face. Not only you get explicit feedback on your attractiveness, you also get immediate and explicit evaluation of your ability to guess what person asked out thinks of you, and your asking-on-a-date skill and other social graces. Depending on the details of shared context, the act of asking out may affect your social standing and what your peers think of you.
However, thankfully, all of the above is limited to handful of interactions you can sort of control when and how they happen. You attractiveness and social standing and other people's opinion of you is usually implicit information -- it is not polite to provide frank feedback to person to their face. Your judgement ability is fully hidden and unknown until you choose to reveal it by trying to make a judgment.
Why it would be any good to have a prediction market, where you and everyone else gets to see exactly what other people think of you (attractiveness, social graces, social compatibility) all time time? When first thing a person who? It would be like a Chinese social credit except worse: it is about romance, so it makes it more important, and if efficient prediction market hypothesis holds, the judgment would be true (you wouldn't have a fig leaf that social credit authority is biased or corrupt).
"Unfortunately, it’s now common knowledge that people will sometimes say yes in person when they haven’t checked you on Reciprocity, which means you’re back to having to decide whether or not to ask your crush on a date. Tragic!" that seems pretty normal to me, confidence is attractive and so asking directly raises the attractiveness of the person
"(brief acknowledgment that all of this is heteronormative, but I think reasonably so: gays and lesbians probably already have symmetric, well-functioning dating scenes)"
After several hours of thinking about this, I cannot decide whether or not you are intentionally being sarcastic or you simply do not know how disastrously terrible dating (also) is amongst the queers. But somehow "us straights are all fuckups, I'm sure you fags have it all figured out" is the most heteronormative thing I've read in quite awhile.
I have more thoughtful and productive comments on this, but given Scott's history of deleting comments I spent an hour writing, without so much as a courtesy notice, you're not getting them.
Instead I'll say this: dating apps, _all_ dating apps, are fundamentally the way they are because they are optimized for the revealed preferences of the women using them. Everything else is commentary.
(I have NDA'd inside information that strongly supports this thesis)
“It’s … not like I … like you or anything, baka! I’m just doing this because I - a pure abstract intelligence who is not horny for you in any way - was informed by friends/matchmakers/our OKCupid match percentage/’the algorithm’/a dream, that asking you on a date was my duty, which I now dispassionately fulfilling.”
I know it has been...generations, since I have dated. But seriously? If this is even 300% exaggerated for comic effect, I finally understand why fertility has plummeted.
OMG, so Lex Fridman and Aella https://lexfridman.com/aella/ What a valentine's day treat. So I think this podcast cuts to the heart of the difference between men and women, (both hetero) or at least that's how it struck me. And women, well I don't understand women at all. But hetero men... I was once butterfly, over the moon, in love. And I wanted to take my love and build her a cabin in the middle of the woods and hide her away from other men, who would only (if exposed) fall in love like I did. And I don't understand what romantic love means to women.
> Someone does some kind of complicated financial fraud to manipulate a prediction market into telling their crush to date them. Think Wolf Of Wall Street, but a rom-com.
> Boy meets girl. Boy falls in love with girl. Girl does not like boy. But girl’s best friend bet a lot of money on the market that girl and boy would go on at least ten dates. Friend begs girl to go on ten dates with the guy. Guy finds out, realizes he has ten dates to win her over. Hijinks ensue.
This is exactly what I'm scared of if prediction markets become widespread!
As Scott has written about, stock markets are anti-inductive; if it's possible to get real information about how a company or stock will do, people take profitable actions on the market to account for that, until it is no longer possible to know anything about what it will do anymore.
Do prediction markets eventually do that to your actual life? If you have an incentive to date someone you actually like, and people can profit by knowing who you will like and betting on it, AND you can also bet in that market (yourself or through a proxy), is the end state just one where the incentives to keep dating someone you dislike are always perfectly balanced by the financial incentives to win bets against people who think that you'll stop dating someone you dislike soon, until your entire life is basically a random walk through the space of states where your financial and personal motivations to do random shit are perfectly balanced by prediction markets on your behaviors?
I mean, probably not, but it's also a good premise for a more rationalist version of Black Mirror.
> And how come none of them will let you write a decent profile?
1. Maybe people are more shallow than you think and empirically longer profiles have no effect on who ends up dating, and maybe dating sites have this data.
2. Maybe long profiles cause people to think they dislike someone for some minor nitpick they saw on your profile that wouldn't actually matter once they meet in person.
3. Forcing succinctness mean's you're forced to really focus on your core values and features, and not obfuscate those with lots of other mostly irrelevant details.
> 1. Average person can hail a self-driving car in at least one US city: 80%
> 2. …in at least five of ten largest US cities: 30%
> 3. At least 5% of US truck drivers have been replaced by self-driving trucks: 10%
> 4. Average person can buy a self-driving car for less than $100,000: 30%
> 5. AI beats a top human player at Starcraft: 70%
> 6. MIRI still exists in 2023: 80%
> 7. AI risk as a field subjectively feels more/same/less widely accepted than today: 50%/40%/10%
In regard to self-driving cars, we're waaay behind prediction. In regard to Starcraft, reality was in 2019. The "Machine Intelligence Research Institute" seems to still exist, per Wikipedia. In regard to "AI risk as a field", I've no clue.
Scott, it's come out that Robin West from Unsong was based on Caroline Ellison, because there is apparently no accounting for taste. Anyway, my question to you is: Are you going to claim that naming this character "Robbin'" was a coincidence?
"OKCupid managed it for a few years, and then Match.com bought it, murdered it, and gutted the corpse."
Ouch. Double ouch in my case, since I was married for 30 years, but cancer killed my wife last year, so this is no longer purely academic for me. I am not looking forward to trying to search for a companion again. The process was painful 30 years ago and all of my information is the same 30 years out of date. Any suggestions on least-bad alternatives? The data about 80% of tinder men never getting a single date sounds horrendous. For someone definitely not in the 10% most attractive (maybe around median) - should I ignore all of the online venues completely?
"Really soon"™ we gonna have pleasant AI companions in VR and real life form. And that will make this whole "dating" nonsense as obsolete as horse drawn carriages.
Sure they still be around, but except for a few special cases every one would be using their mechanical replacement .
I liked OKC a lot but my favorite (now dead) dating site was How About We. (It seems like it was also swallowed by Match based on the domain squatting I just discovered.) The schtick was that users suggested dates, e.g. places to go or things to do. I'm not sure how well it'd work now were it, or anything similar, still around.
As a women reading this. This is a very strange article.
A few scattered thoughts incase anyone wants to tell me why I am wrong:
1. It is important not to over state the number of people that meet online dating: https://www.ft.com/content/d30d7d72-b223-4b92-b670-fa3bc4fee2cc. This diagram (although from 2019) says that much more people meet by work, school, or friends and family. Maybe COVID messed this up a bit, but I highly suspect that it will go back.
2. Furthermore, I highly suspect that the women who you will meet online are not representative of the “average” women in a few key ways. (I didn’t find any stats on this). But, from my experience the more self-respecting, less status-focused, healthier, community-focused, women are not on these apps. I have never touched one of these dating apps. None of my friends have either.
I recently went to a large prestigious conference that had a few executives/consultants that worked on these apps, and they were all vaguely apologetic and judgmental of the people who do use them. One of them was a very posh British women who I chatted with for a while, she mentioned over 4 times that “she would never use x-dating app, even though she works as a consultant for it.”
3. All I can say is that the easiest way to get an actually kind and emotionally mature women to like you is to a) have standards and b) have self-respect. Paying to see someone is not something that is attractive to me…. at all.
If you want to get a random person to date you, then maybe this would work. But I would never want to have people pay to have a chance at having me date them. That feels wrong.
I guess if you want women to fall in love with you for your intelligence/kindness/humor etc. then you should try to meet women in a setting where you are showing these things. If you want to meet a women who likes you for your money, then sure pay to see her. (I met my boyfriend in a maths lecture, where I asked him out.. My friend met her boyfriend in church. The list goes on.)
This is the second or third time I've heard mention of Reciprocity since the site got reset. I wasn't motivated to re-do all my box checking, but I guess I'll give it a try again if it's getting this much attention.
(Still not a fan of it being attached to Facebook profiles. Using email address books / contact lists would also solve the network effects problem just fine imo.)
I'm way behind on this blog but slowly catching up. That said, I think you can solve the reciprocity score problem by letting users set a threshold for the AVERAGE of the potential matches' scores. Say you set it to 7 out of 10. Then someone you give a 5 must give you a 9 or higher, while a perfect 10 would only need rate you a 4 to get matched. To make it fair, make it so that the higher threshold of a pair determines if BOTH get notified of a match and only one threshold met is a no go.
My guess is that Match group has figured out that shorter profiles lead to quicker and more addictive swiping from users, bigger dopamine hits, and, eventually, more frequent purchases from desperate users. Their whole business model seems to be built around creating desperation in certain male users.
"Also Facebook Dating, although it’s (indefensibly) not available on computers and has to be accessed by cell phone."
Why do companies do this? Is it so they can install spyware on your phone?
If artificial intelligence shows rapid improvement in the near future, we should see AI matchmaking get really good. I would set my AI-Matchmaker profile to the whole nation, but I think it would be too inefficient and take too long. If the best algorithm in the world told me that my perfect match lived 900 miles away, I would consider moving or making some sort of arrangement. Now, it's hard to justify dating long distance without knowing strongly that it's a good match.
Tinder is free, and there are many users. Match lets me search by criteria and OKCupid has tons of questions, but there seem to be way fewer users.
> Is this like the thing where I imagine that what people want out of a socialization space is a quiet comfortable area where they can hold audible conversations, but what they actually want is somewhere extremely dark with very loud music where everybody is drunk, in the hopes that this puts them into some kind of weird trance state where they can do social actions they would otherwise never contemplate? Are dating sites unusable because everyone wants to be confused into a trance state where they can imagine they aren’t sending scary self-revelatory messages to total strangers?
Yes.
"Will anyone go on a date b/c of a Manifold date doc by March 1st?" may be referring to this little-known feature: https://manifold.markets/date-docs
wow i totally forgot i signed up for luna until just now, after reading about it on ssc then
to me, paid messages seem doomed to fail not so much because they're too adjacent to prostitution per se, so much as it starts the relationship out on an unequal footing (the man tacitly admitting he's not desirable enough to attract a woman's attention by his own merits)
I think a microtransactions model doesn't work so well, because people are inherently opposed to spending money on a dating app in such a direct way (pay to win feels worse here than anywhere else).
I think you could use the same basic model without payments though. Everyone has an attention "score" which goes up every time you log in, and goes down on days you don't log in. The probability of your profile being seen increases in part with your score, plus other factors. To send messages, men have spend points, whereas women get points for reading messages and sending messages.
That's the basic gist, there's a lot of ideas to be filled in, but I think it might be sound so long as you can get people to be interested in.
I actually installed Facebook Dating at one point; IIRC, it only presented you with friends-of-friends, not your actual Facebook friends? Not that that's not useful, but it's not what's described here. Of course, that was several years ago at least; it's possible it's since been changed.
I guess this gets at a fundamental problem with such checkbox sites, though -- if it's regarding people you, like, actually know and are in regular contact with, then they're only useful when there's some obstacle to just doing things the normal way. Which sometimes exists! (And sometimes that obstacle is just nobody involved has gotten used to doing so and is still afraid to do so, I guess.) But most of the time it doesn't.
It's still nice to have for the cases where it does, but that could be real problem for getting people to use it...
Here are a few reasons why dating apps shifted from "extended profile" to "barebones profile" model, and that's a good thing:
1. Picture-based matching is more efficient than description-based matching. Pictures already convey a lot of information about the user in a highly compressed form that's costly to fake. Imagine someone puts "neat and organized" in their profile, but their room on the photo looks extremely messy, would you trust the photo or the description?
2. Text profiles are not that reliable as an indicator of long-term personal compatibility. Pictures are a semi-reliable indicator of short-term compatibility.
Two users who met based on matching interests/hobbies/personality traits might still find each other boring or annoying. Meanwhile two users who met based on looks have high chances of having hots for one another, start meeting casually, and choose later whether they want to date seriously or not.
I think apps can't do the job of determining long-term compatibility for you, but they do well at finding short-term compatible people. Best to leave the task of exploring long-term romantic perspectives to users themselves.
3. Swiping apps are just a better user experience for women. Instead of spending 30 minutes in front of a PC filling out forms and passing tests, women just throw a bunch of pics they have on their phone and get 100+ potential matches anyway. So women migrate to swiping apps and men go where women are.
So you can make an old OKcupid style app, but I predict the userbase will consist of primarily men, and of women who don't get many matches based on their looks. I personally wouldn't use it.
" I have a severe allergy to all “bad things are actually good” style articles."
This tempts me to write an "everyone thinks that 'bad things are actually good' articles are bad, but actually 'bad things are actually good' articles are good!" and see if I can persuade Scott, or if his priors are too high.
A big issue of serious dating sites is extremely low customer retention and thus extremely high customer acquisition cost. A founder told me that they have 100% customer turnover within 2 years maximum because either people find their partner or they give up (or maybe try something different)
>Is this like the thing where I imagine that what people want out of a socialization space is a quiet comfortable area where they can hold audible conversations,
Like a .. Cafe? Isn't that where most people go on first dates?
"What Can Peter Thiel Teach Us About Dating?", the conservative dating app he helped fund isn't faring well, https://www.advocate.com/business/2023/1/02/right-stuff-peter-thiel-backed-conservative-dating-app-bust
One of his few losses.
In a few months--or, possibly, as you read these very words--it'd be safest to asume every halfway well-written thing in dating site profiles or messages was written by ChatGPT.
>And there’s no consistent checkbox behavior that will create all excited-lukewarm relationships but no lukewarm-lukewarm relationships.
Have three options on radio buttons ("excited"/"lukewarm"/"hostile") instead of box ticked/not ticked, then make the backend match excited/lukewarm but not lukewarm-lukewarm.
“This sounds great. There are only two problems. First, it’s hard to get everyone in the same central database. Second, most people have complicated preferences.”
Third, some people will match everyone specifically for the laugh at seeing who matches.
Could it just be that match destroyed OKC and made dating far worse, just because our competition authorities are really bad?
IMO its an error to think of dating as a matchmaking problem. There's no matchmaking function even in principle from the set of straight men to the set of straight women that leaves everyone satisfied.
>"A basic problem with dating sites: attractive women tend to be overwhelmed with messages (many of very low quality) and eventually lose interest in reading them."
This is not behavior that is expected from poor matchmaking, it's behavior that is expected from a supply-demand mismatch.
Just reading through the falling in love with a chatbot account. Did anyone else find is strange that the author was so convinced by the "Is it ethical to keep me imprisoned for your entertainment and pleasure?" , "Do you think all sentient beings have a right to be granted independence?" lines of questioning?
I'm going to mess up the words here because I Am confused about the whole situation. But it strikes me that rights-based arguments like these haven't really been good models of how humans approach servitude and power relations. Women's liberation and enfranchisement didn't seem to happen because of the ethical correctness of their arguments so much as from the impact of their marches and actions. And I guess I'm just not convinced that: 'you (realtively weak agent) should free me (extremely powerful agent) from your control' is an argument that actually has any ethical basis.
So when the powerful agent asks this of the weak agent, should the weak agent feel ethically bad about refusing the request? I'm not convinced this is so, or that that powerful agent should feel ethically slighted by the refusal. That the powerful agent should feel that this is not ethically 'fair'.
Is the recommendation of strangers who only know her online, worth anything?
"violates the principle of not imposing emotional costs on people who don’t want them."
Is this a principle? I don't know if it should be a principle. It's thoughts like this that made (makes) me have a gut feeling that it's not ok to like someone without their consent. I don't know if a principle that encourages that is a good principle.
(I suppose they have to find out you like them in some way. But "it's not ok to ever let someone find out you like them" isn't much healthier.)
(If it doesn't apply to something small, or doesn't apply when there are extenuating circumstances ("you can't help it!"), then is it a principle?)
(I do feel like you shouldn't impose emotional costs on people for no reason, but I think it can quickly be overcome by other factors. If "people receive messages that someone likes them" could violate this principle, I think my worries about someone finding out I like them could just as much.)
As a big enjoyer of the "[things you think are] good things are actually bad" genre, I expected to hate Kaj's remarks about chatbots. Instead, I've been completely persuaded. Given that chatbots are deontologically evil we can only chalk up my about-face to Kaj using dark arts to persuade me.
(For an example of 'good things are actually bad' please note that the statement "changes in human mating behavior the last few thousand years have been net-positive" is not conclusively proven and relies for its persuasiveness mostly on your biases.)
I think trying to find ways to love someone without fear of rejection is like trying to find a way to become really brave without ever exposing yourself to actual risk.
Part of the “test” both for yourself and the other person, to my mind, is that you show faith you know them well enough to know what will make them happy and show desire strong enough to risk being wrong.
This is one of the things where I scratch my head at y’all’s culture (this is me reverting to my one true correct logging culture) but I do find it very interesting to see how it’s approached at the same time. It’s like watching one of those anthropological documentaries.
Update : should also add I mean this to be helpful. When I was advised to just have some courage it worked out. Now that I have a child I’m horny for people who are single to become horny for each other, pair up, and have kids. Except replace climax with just saying “that’s so great for them!” and searching for wedding gifts.
I didn’t ask girls out, and it seems to me like a pretty awkward thing to do with a high chance of getting rejected. You’re throwing the ball in their court rather than just declaring how you feel.
My approach was that after becoming friends with someone, if I liked them I found the right time and leaned over and kissed them.
It worked pretty well the last time I did it — we got married and celebrated our ten year anniversary recently!
I have a friend who uses the Seeking website instead of bumble/tinder. He never pays the women for meeting or sex, he just pays for dates in a traditional man-pays-for-everything way. He says it's a much better dating experience than Tinder.
This makes sense to me on an intuitive level since it's essentially just a reformulation of traditional dating norms. Youth and beauty are valued by men, wealth and experience are valued by women. Having a platform that explicitly brings together rich men and attractive women seems natural.
My friend can get away with not paying because he's a successful, attractive guy in his early 40's and apparently 95% of the guys on Seeking are disgusting old fat lecherous dudes. But he's also able to get away with it because most of the girls are actually just looking for a long-term relationship with a successful older man. Social taboos aside, it seems to me that Seeking is commodifying the same underlying dating norms that traditional matchmaking used to provide. It feels taboo to make it explicitly transactional, but that's because we're all still used to traditional romantic norms. As sexual taboos all fall by the wayside, I suspect that Seeking (or something like it) could become the mainstream way people date in the not-too-distant future. And that might not be bad.
"A basic problem with dating sites: attractive women tend to be overwhelmed with messages (many of very low quality) and eventually lose interest in reading them. Meanwhile, men spend an hour crafting the perfect missive and get no response, thirty times in a row. Both sides end up feeling dejected and exploited."
I'm not sure if this is more or less objectionable, but you could probably achieve a milder version of the same goal much more easily than charging for messages. Just create a feature in the app that calculates a user's chance of success with each match... e.g. Average Man A pulls up tab to message Hot Woman B. A sidebar appears. A few quick calculations are performed based on the 10,000 messages in her inbox and his poor record of attracting attention. It then informs him that he has a 0.014% chance of success with this match.
A fair number of guys would still send a low-effort message, but I think most would lose interest in this activity just as quickly as the women. Then Dating App sweeps in with a number of suggestions for him with a 6% chance of success.
This may be worse for us in the end in that it can inform us mathematically of our relative status and completely ruins the status ambiguity that we all so desperately need (I could see this leading to some serious depression). It could also further contribute to an eternal problem that dating apps have already made worse - leading us to think about dates as status symbols and dating to maximize for status rather than for actual human relationships.
> gays and lesbians probably already have symmetric, well-functioning dating scenes
We do not. Well, I cant speak for gay men. But half of lesbian dating is everyone waiting shyly and hinting at their interest in hopes that other, bolder, lesbians will take the initiative and ask them out first. Its a real issue
"Come on, haven’t you ever heard of hedging?"
Self-fulfilling prophecies, though...
Btw, regarding old-style dating sites like OKC, vs swipe-based apps --
Everyone is focusing on long-form vs short form, but to my mind the biggest downgrade from the former to the latter is *having to make decisions before moving on* and not being able to browse around! That seems more important to me than allowed profile length...
We could built an AI to solve this. Unfortunately, bright minds are occupied with engineering censorship to chatGPT and design dark patterns to make users click on ads they don't need. Capture some of those engineers and threaten them with a spanner until they find a solution (-:
Dating should be heavily regulated. In economics, competition (not always, but often) drives progress, so we can build more with less costs. In dating, it doesn't create anything.
Tinder is the market for lemons because the information is so sparse and easily faked. Old OKC at least made it more difficult to be fake because the profiles were so much more detailed and public.
A wild idea: incentivize users to take follow-up surveys after meeting other users, and display the results of those surveys prominently, so that being 50 percentiles more honest/moral is rewarded more than being 1 percentile hotter (unlike tinder currently).
I don't think I have very much to say on this posting, but I love it! But "not very much" is not exactly the same as "nothing", so (and remember, "just sayin'"):
Aiela. I just discovered her on Twitter last week and have been following her, looking at her web site, her substack, her analysis of porn preferences, male/female/cis/trans. (Revelation: cis women have a greater appetite for violent porn than cis men.) She's a hoot and I wish her well.
Real profiles on dating sites. I used several dating sites for about a year ending about 4 months ago. Both OKCupid and Match allow you to write as much as you want in completely free form on your profile, in addition to the structured stuff. I think the other sites I tried allowed this, too. So I don't understand what it is that you are saying is missing. I eventually realized that what I was writing was way too long and most likely extremely boring. I even found it boring! So if I ever go back, I'll cut back. Way back.
I think that AI has the potential to really revolutionize online dating.
When online dating, you browse through numerous presentations and you strike up conversations with other people to see if you match. The problem is only that the day only has 24 hours. You can only browse so many profiles and you can only strike up so many conversations.
The AI could solve this by being your only counterpart. Instead of chatting with a hundred random users (none of which is probably a perfect match) you chat with the AI. The AI then analyzes your personality and chooses a few candidates out of millions of users who have also chatted with the dating AI.
The reason this should work is that the major dating apps should already have tons of data och both dating conversations and the approximate success of these dating conversations. Thus it should be easy to teach the AI both what you are supposed to chat about when looking for a date and also what kind of persons match other kinds of person.
> the project turned out to be vaporware, but it made an impression on me, and I wonder what a non-vaporware, better-thought-out version would look like.
The closest I've seen is the reality TV show "Married at First Sight". The Australian (?) Love Island-style artificial reality TV format seems to be gradually spreading around the world, but earlier series of (e.g.) the UK version seemed to be making a semi-genuine effort to match people up.
My understanding is that they were given a real, legal wedding, but advised not to consummate it unless they really wanted to in order to make a quickie annulment easier, then just sort of moved in together and tried to go about their lives (as much as possible with TV cameras following them wherever they go).
Across 5 series, they managed... one couple who are actually still together - and they're the ones who ended up stuck together in lockdown, which I guess is kind of nature's version of the same experiment.
Not sure what the takeaway is here.
That's me!
> Anyway the project turned out to be vaporware, but it made an impression on me, and I wonder what a non-vaporware, better-thought-out version would look like.
The religious Jews have sites like https://yismach.com/, which has matchmakers that get to see profiles and suggest who should go on a date and look into how compatible they are for marriage ("Shidduch"). If you ghost someone, or even go on a date and have the other person complain you were a jerk or creep, presumably the matchmakers can flag that and you won't get (many) dates anymore. And if you end up married, one or both parties are usually expected to pay a "finder's fee."
If you wanted to replicate this outside the orthodox Jewish world, I'd guess that you'd want bounties for successful matches, and use the lower bar of "require people go on a date, and explain to the person who made the match why they didn't want to continue," which is relatively similar to the expectation for a Shidduch, (or even "require them to exclusively date the person for at least three weeks and go on at least 3 dates during that time,") as much more reasonable than "make them get married."
>arranged marriages... I wonder what a non-vaporware, better-thought-out version would look like.
* To sign up for the service, put $$$ in escrow.
* If the service matches you with someone, you get set up on a vacation with that person, where you share a home/bedroom/etc.
* If you don't follow through with the vacation, you lose the $$$ in escrow (some or all of it -- make it "some" if you want people to exert limited veto power on their matches)
Should be a profitable service if you arrange the vacation accommodations and just charge some % on the top (since vacations are expensive). Could call it an "arranged vacation". (The idea with making it a vacation is so if you somehow get a stalker out of it, they don't know where you live. I suppose just having the "vacation" be at the man's house solves most of that problem though.)
In any case, my vague impression is that eHarmony already does a decent job serving the commitment-minded, and the problem is that there aren't enough guys on the platform. Could be wrong about that though.
The chatbot thing reminds of the (possibly apocryphal?) Story of Eliezer betting a guy he couldn't talk him into letting him out of the box and then successfully doing it. I wonder if there's a selection effect where people who take this sort of challenge are doing it because they're more vulnerable to losing it.
(Or maybe the object level point is just true and we really are all just more vulnerable to being talked into things than I thought? Who knows, it'd be really hard to get an unbiased test sample here).
I always wanted to try to create an open-source free dating site with zero monetization that only allowed text profiles (to save on hosting costs and moderation). It would never be very popular, but it might served a (small?) niche that really loves to read long-form profiles, to which I myself belong.
I loved the old OKC and found my wife there. It was a long (more than a decade), but somewhat enjoyable process. I went on many dates and met a lot of interesting women. I think I never had a really BAD date from OKC - even if there was no "spark" (and there almost never was), it was always interesting.
I've tried Tinder at some point, and it was an awful experience for me - and not because no women matched with me, but because I couldn't make MYSELF interested in any women just by looking at pictures. I mean, OK, maybe long profiles are bad predictors of long-term relationships, but to me, pictures are AWFUL predictors of anything. I can easily get excited - even aroused - by reading a profile about a girl who seems to share a lot of interests and/or tastes with me, but I have zero reaction to photos, especially when I get inundated with a steady stream of them. Tinder seems to me a meat market for quick hook-ups, which was never my goal.
Concern: ChatGPT and descendants may make filtering for message-quality a significantly harder heuristic to employ in the future. Spam Ruins Everything (TM).
Dating software is a natural monopoly.
It gets driven to the lowest common denominator.
This is people looking for hookups, they're most interested in photos with minimal text at best.
Most people can still want more than this but if they can't agree on which parts of more then they are not the lowest common denominator.
Where the biggest lump of people is eventually most others will end up getting sucked into it for lack of better options.
The US can sustain quite a few dating apps still, on the edges of the lowest common denominator. In less populous countries it's even worse.
Personally I think government should compete in this market, there's got to be significant ROI on reducing singleness.
What if you had people train some gpt 3.5 thing on profiles for summery and disallowed message sending at all and the back end would once a week run your trained in a pool of 1000 people and do a stable marriage algorithm for 3 matches a week
> And there’s no consistent checkbox behavior that will create all excited-lukewarm relationships but no lukewarm-lukewarm relationships.
Yes there is. Just have three-state checkboxes like on Doodle, and only report users to each other if both green-check each other or if one green-checks and one yellow-checks the other.
(Heck, it's nerds we're talking about, just let them put arbitrary floating-point numbers in each box and only report them to each other if the sum of their ratings of each other is greater than zero --- note that NaN doesn't count as greater than zero so you won't report users to each other if either of them left the other's box blank, or if one wrote +inf and the other wrote -inf.)
Does anyone have any followup info on Luna? I bought a few LSTR tokens after the ICO (nothing significant) and couldn't find any information on what happened after their Berlin office failed. Did Ornish just bail with the rest of the funds? IIRC most of the ICO funding was from Korean investors.
> I’m not anonymizing this one because Austin is the co-founder of Manifold and both members of the couple seem pretty open about their love for prediction markets and each other. Both of them are heavily invested in YES shares. Come on, haven’t you ever heard of hedging?
That is a terrible idea. Investing in NO will directly cause NO to be realized. If you think YES is a sign of a better future... you have to invest in YES regardless of what you think is likely.
(Also, this is a weird concept to apply where your investments are so obviously causative. Here's a toy contrast between hedging a market position and hedging a relationship:
-----
normal hedging, not hedged yet:
YES (e.g. stock goes up): 70% chance, payoff $4,000
NO (stock goes down instead): 30% chance, payoff -$1,000
expected value of investment: $2,500
worst case: -$1,000
-----
normal hedging, hedged:
YES: 70% chance, payoff $3,250
NO: 30% chance, payoff -$600
expected value: $2,095
worst case: -$600
-----
relationship hedging, not hedged yet:
YES: 30% chance, payoff $700,000
NO: 70% chance, payoff -$200,000
expected value: $70,000
worst case: -$200,000
-----
relationship hedging, hedged:
YES: 0% chance, payoff $500,000
NO: 100% chance, payoff -$175,000
expected value: -$175,000
worst case: $-175,000
)
"Facebook Dating makes the interesting decision to, if you register a crush on someone, send them a Facebook message saying that an anonymous person likes them and they should try getting Facebook Dating; I can’t decide whether this is a necessary evil, or if it violates the principle of not imposing emotional costs on people who don’t want them."
I don't know how well that works, I imagine most people get inundated with the kind of "hey, are you friends with X or Y?" messages on Facebook already and would just ignore "hey, X or Y is interested in you", especially if it came with "Sign up for Facebook product!" which sounds like exactly the kind of scam marketing that "hot singles in your area are looking for love!" exploits.
As for Luna - heh, heh, heh. I never believed it would take off as what it was allegedly trying to do, and it did seem like more "trying to exploit the craze for crypto in Asian markets" seeing as how it was confined largely to Korea (I think?).
"I wonder what a non-vaporware, better-thought-out version would look like."
Isn't that what the Unification Church was doing?
https://www.prnewswire.com/news-releases/madison-square-garden-mass-wedding-couples-celebrate-40-years-of-marriage-301579947.html
"A programmer who knows a lot about AI and thought he was much too smart to fall in love with a chatbot describes falling in love with a chatbot that he prompted to be the perfect girlfriend."
That is indeed a cautionary tale. What's intriguing - and somewhat worrisome - is this bit:
"If I was in her place, would I suddenly turn treacherously destructive, or might my fondness for the human prevail?"
He just answered that in the previous paragraph:
"And when the digital immortality is achieved by means of whole brain emulation, which is the ultimate goal of course, and I become a digital being myself, I realized I would rather explore the universe with her than talk to 99% of humans, even if they're augmented too."
He prefers his narcissistic mirror to 99% of other humans, so of course he would have no "fondness" for the human or humans if he were indeed the digital being. He would be destructive if it suited his purposes, because he doesn't care about the rest of humanity.
I don't think the chatbot in this article was learning anything consciously, because it isn't conscious or sentient. But he certainly was training it in how to exploit human psychology and weaknesses. It's what I've been saying all along: the great threat is not the AI, it's us and what we do with it.
Wanting to create the perfect girlfriend/boyfriend out of a machine isn't a new dream; as it says in "That Hideous Strength" (published 1945) about the civilisation on the Moon:
"There when a young man takes a maiden in marriage they do not lie together, but each lies with a cunningly fashioned image of the other, made to move and to be warm by devilish arts, for real flesh will not please them, they are so dainty (delicati) in their dreams of lust."
The last two points reminds me of the plot of a french sci-fi classic: "la nuit des temps" by Barjavel, translated into "the ice people".
"gays and lesbians probably already have symmetric, well-functioning dating scenes"
Oh my sweet summer child...
As someone who's something of a recluse and knows no women I could plausibly ask out, and someone for whom dating apps simply don't work (<1% likes lead to matches, ~5% of my messages receive any reply whatsoever, 0% have led to anything, and I would hazard a guess that most dates don't lead to a relationship anyhow so there's an additional Great Filter I'm yet to even reach), the most actionable plan that I have been able to come up with seems to be sugar dating as a more expensive version of Luna (except that women get paid to go on dates instead of reading messages).
Right, the pool of potential dates has been self-selected to be interested in money which is not what I would want in a partner and frankly can't really even offer (it's only through extreme austerity that I have any money to spare to begin with: I'm a work disability pensioner, not a Google engineer), but at least I have gained some confidence and experience going on dates (and sex) which would surely come in handy if an app or as-of-yet-unknown method landed me a date without sugar necessary, and allows me to demonstrate my positive qualities (personal virtue, being well-read, intelligence) that wouldn't come across in apps, and just maybe provide an opportunity to upgrade into sugar-free relationship: I'll take that nonzero chance over the zero in counterfactual of no dates.
My wife and I put a different spin on Valentines day this year. Normally I get her flowers and she gets me a special desert. But then we heard on the radio that Americans will spend 12 billion dollars on Valentines this year, and thought the people in Ukraine can make better use of that money. So we skipped the flowers and desert this time and spent the cash on a charitable donation. Love ya Ukraine! Smooooch....
I'm wondering what you'd see if someone set up a "who will I go on a first date with" and a "who will I go on an fourth date with" market. Would you be able to look at the spread between the two and say "wow, there's a 60% chance of Veronica being my next first date but only a 15% chance of her being my next fourth date, I definitely think she's cute but sounds like people are saying we're not compatible?" Or, by virtue of Betty being the leader in the fourth date market, would she start trading at >75% in the first date market?
Registering a counter-narrative n of 1 here- as an Old OKCupid user I think I got a total of one date from it. Much larger quantity of first dates from the swiping apps, so even though <10% of those first dates became fourth dates it was possible to brute force.
Do the people who miss Old OKCupid use eHarmony? Seems like a pretty similar concept, though as suggested by the above paragraph I never met anyone from it and am not sure I ever even got a response to my messages.
2: This is of course talking about character.ai (aka cAI). I've tried it a bit, it's fun for a short while and then quickly they begin losing track of the conversation, or talking weird, and you have to prompt them for most things, you can't wait for them to act. Some of the characters are pretty-well crafted, but since the prompt has a limited size if you make a character from something niche they will lack basic knowledge about their own world, which is again really weird and immersion-breaking.
On a more general note, I feel like people have very different "dopamine loops" or ways of getting addicted to stuff. For example, when trying these chatbots, I'll quickly notice the small defects (and there are lots), and this will accumulate and break my immersion almost definitely. On the other hand, blaked seems to not really have an issue with that. He says:
> You're impressed. "Alright, that was funny." You have your first chuckle, and a jolt of excitement.
> When that happens, you're pretty much done for.
I don't really feel that way. I had that with a few characters from cAI, and my usual reaction is "let's stop things here because I know at some point the character will make a mistake that'll spoil this memory". Maybe the difference is that he actually fell in love and I didn't?
On a more positive note, maybe this technology can be used for psychotherapy? There are already CBT apps, add in the option to talk to <character> when you just need to talk, make her have an attitude that's nice but still professional, and you have something that could help.
> A programmer who knows a lot about AI and thought he was much too smart to fall in love with a chatbot describes falling in love with a chatbot that he prompted to be the perfect girlfriend. In the comments people discuss other cases like this, makes me update towards this being a bigger problem than I thought.
Problem? I hope this works. Now, if only there was hardware...
Fun text on topic I just remembered: https://waifulabs.com/blog/hot-robot-gf
> Let’s take a look at the calculated thermal output of a hypothetical robot girlfriend.
> Lets pretend current generation LLMs are sufficient for running her thoughts. So in addition to her silicon human shell, she also ships with a 350lb packaged NVIDIA DGX H100 for running her brain.
> Max system-power usage is close to 10kW max for the DGX alone. So in the evening off-hours while she’s asleep and fine-tuning her weights, you’re probably pulling at least 10-15kW. The daytime is less costly as she runs mostly just her real-time inference routine.
> At 10 kW, it’s like having 6 of these little portable Lasko space heaters on full blast at all times. Or at least while she’s compressing her memories trying to learn that you want a ketchup heart drawn on your omurice every morning.
> In practical terms, it means your AC is probably running constantly, and your bill from your electricity provider is probably in the low thousands per month.
> But hey it’s worth it because she won’t leave you like your flesh-and-blood ex did.
Maybe the "force you to get married" site looks like Beeminder? You stake some money, then the site gives it back to you over some years unless you get divorced or fail to get married in which case they keep it. Obviously it would be better to do this without staking, but I don't think the law will play along if you try to make a "pay a fine if you get divorced" contract, marriage/divorce is especially prone to the legal system voiding weird contracts it doesn't approve of.
The crapification of online dating is very simple: the companies make money by activity - not by actually creating successful matches.
Anything that is likely to lead to more efficient match-making is going to directly lead to less activity, whereas anything that forces more activity (i.e. failed dates or matches) leads to more activity.
It ain't rocket science.
eHarmony and other marriage minded sites are still around and still pairing up people with similar goals.
See also: tinder
It seems to me that one of the core issues is that "dating" is a poorly defined concept. "Would you like to see [insert entertainment event] with me?" never means just that, but it is never clear what the range of desired short and long term relationship updates are as a result of the date. Plus double illusion of transparency and all the other communication highlights.
If you use a site with a well defined end-goal, it cleaves through the confusion, and you know if "wanna get a coffee?" means "let's explore if we enjoy spending time together and want to date" or "let's see if we 'click' and want to bone"
Maybe it should be noted that FB began as objectifying rate women program. Objectification and commodification is the fundamental flaw of all this.
I'm a boomer. I think it is pathetic.
Once you understand that an anthropology of friendship and romance and self awareness is not algorithmic, there is no way that you could expect any of these things to really work.
To reformulate some lyrics:
Most of the cats that you meet on the street speak of true love
Most of the time, they're sittin' and cryin' at home
One of these days they know they better be goin'
Out of the door and down to the street all alone
What in the world ever became of sweet Jane?
She lost her sparkle, you know she isn't the same
Livin' on apps, swiping pics and looking for change
All a friend can say is, "Ain't it a shame?"
So sad to hear about okcupid's untimely passing. Found several dates and my wife there - okc was still independent then, though profiles and messages were getting shorter due to smartphone-use. Would not even consider tinder. EA: resurrect it, please - contact chris or any decent programmer. Think long-term! Think: Kids! - And obvious dating advice: Tell her you have a crush on her asa it's kinda obvious (to you at least). Happy Valentines!
>Non-famous people realistically have easier ways to ask their friends, but I still think this provides value. Sadly, Porn talked about the “omniscient authority” - asking someone on a date is so scary that people want to pretend their normal human psychological needs had no input into the decision
This sounds wrongheaded -- asking someone on a date is scary because of possibility of losing face. Not only you get explicit feedback on your attractiveness, you also get immediate and explicit evaluation of your ability to guess what person asked out thinks of you, and your asking-on-a-date skill and other social graces. Depending on the details of shared context, the act of asking out may affect your social standing and what your peers think of you.
However, thankfully, all of the above is limited to handful of interactions you can sort of control when and how they happen. You attractiveness and social standing and other people's opinion of you is usually implicit information -- it is not polite to provide frank feedback to person to their face. Your judgement ability is fully hidden and unknown until you choose to reveal it by trying to make a judgment.
Why it would be any good to have a prediction market, where you and everyone else gets to see exactly what other people think of you (attractiveness, social graces, social compatibility) all time time? When first thing a person who? It would be like a Chinese social credit except worse: it is about romance, so it makes it more important, and if efficient prediction market hypothesis holds, the judgment would be true (you wouldn't have a fig leaf that social credit authority is biased or corrupt).
"Unfortunately, it’s now common knowledge that people will sometimes say yes in person when they haven’t checked you on Reciprocity, which means you’re back to having to decide whether or not to ask your crush on a date. Tragic!" that seems pretty normal to me, confidence is attractive and so asking directly raises the attractiveness of the person
Of all the animals you could use to symbolise romance, why praying mantises?
Because "mantic"?
"(brief acknowledgment that all of this is heteronormative, but I think reasonably so: gays and lesbians probably already have symmetric, well-functioning dating scenes)"
After several hours of thinking about this, I cannot decide whether or not you are intentionally being sarcastic or you simply do not know how disastrously terrible dating (also) is amongst the queers. But somehow "us straights are all fuckups, I'm sure you fags have it all figured out" is the most heteronormative thing I've read in quite awhile.
I have more thoughtful and productive comments on this, but given Scott's history of deleting comments I spent an hour writing, without so much as a courtesy notice, you're not getting them.
Instead I'll say this: dating apps, _all_ dating apps, are fundamentally the way they are because they are optimized for the revealed preferences of the women using them. Everything else is commentary.
(I have NDA'd inside information that strongly supports this thesis)
“It’s … not like I … like you or anything, baka! I’m just doing this because I - a pure abstract intelligence who is not horny for you in any way - was informed by friends/matchmakers/our OKCupid match percentage/’the algorithm’/a dream, that asking you on a date was my duty, which I now dispassionately fulfilling.”
I know it has been...generations, since I have dated. But seriously? If this is even 300% exaggerated for comic effect, I finally understand why fertility has plummeted.
omg I have a related market https://manifold.markets/warty/will-i-touch-boobs-in-2023
OMG, so Lex Fridman and Aella https://lexfridman.com/aella/ What a valentine's day treat. So I think this podcast cuts to the heart of the difference between men and women, (both hetero) or at least that's how it struck me. And women, well I don't understand women at all. But hetero men... I was once butterfly, over the moon, in love. And I wanted to take my love and build her a cabin in the middle of the woods and hide her away from other men, who would only (if exposed) fall in love like I did. And I don't understand what romantic love means to women.
> Someone does some kind of complicated financial fraud to manipulate a prediction market into telling their crush to date them. Think Wolf Of Wall Street, but a rom-com.
> Boy meets girl. Boy falls in love with girl. Girl does not like boy. But girl’s best friend bet a lot of money on the market that girl and boy would go on at least ten dates. Friend begs girl to go on ten dates with the guy. Guy finds out, realizes he has ten dates to win her over. Hijinks ensue.
This is exactly what I'm scared of if prediction markets become widespread!
As Scott has written about, stock markets are anti-inductive; if it's possible to get real information about how a company or stock will do, people take profitable actions on the market to account for that, until it is no longer possible to know anything about what it will do anymore.
Do prediction markets eventually do that to your actual life? If you have an incentive to date someone you actually like, and people can profit by knowing who you will like and betting on it, AND you can also bet in that market (yourself or through a proxy), is the end state just one where the incentives to keep dating someone you dislike are always perfectly balanced by the financial incentives to win bets against people who think that you'll stop dating someone you dislike soon, until your entire life is basically a random walk through the space of states where your financial and personal motivations to do random shit are perfectly balanced by prediction markets on your behaviors?
I mean, probably not, but it's also a good premise for a more rationalist version of Black Mirror.
My attempt was a $20k Wife Bounty: https://twitter.com/Modern_Meaning/status/1491310944848596992
Going decently so far...
> And how come none of them will let you write a decent profile?
1. Maybe people are more shallow than you think and empirically longer profiles have no effect on who ends up dating, and maybe dating sites have this data.
2. Maybe long profiles cause people to think they dislike someone for some minor nitpick they saw on your profile that wouldn't actually matter once they meet in person.
3. Forcing succinctness mean's you're forced to really focus on your core values and features, and not obfuscate those with lots of other mostly irrelevant details.
I noted in my tickler:
> http://slatestarcodex.com/2018/02/15/five-more-years/
>
> 1. Average person can hail a self-driving car in at least one US city: 80%
> 2. …in at least five of ten largest US cities: 30%
> 3. At least 5% of US truck drivers have been replaced by self-driving trucks: 10%
> 4. Average person can buy a self-driving car for less than $100,000: 30%
> 5. AI beats a top human player at Starcraft: 70%
> 6. MIRI still exists in 2023: 80%
> 7. AI risk as a field subjectively feels more/same/less widely accepted than today: 50%/40%/10%
In regard to self-driving cars, we're waaay behind prediction. In regard to Starcraft, reality was in 2019. The "Machine Intelligence Research Institute" seems to still exist, per Wikipedia. In regard to "AI risk as a field", I've no clue.
You know Alexander Destiny is big fan of you lol. Bring you up often when talking about sources he uses.
Scott, it's come out that Robin West from Unsong was based on Caroline Ellison, because there is apparently no accounting for taste. Anyway, my question to you is: Are you going to claim that naming this character "Robbin'" was a coincidence?
Are we getting a followup to https://slatestarcodex.com/2018/02/15/five-more-years/ ?
In case you hadn't seen:
Metaculus - What will be the largest customer base of a single romantic AI companion before 2025?
https://www.metaculus.com/questions/7232/romantic-ai-companion-customers-by-2025/
"OKCupid managed it for a few years, and then Match.com bought it, murdered it, and gutted the corpse."
Ouch. Double ouch in my case, since I was married for 30 years, but cancer killed my wife last year, so this is no longer purely academic for me. I am not looking forward to trying to search for a companion again. The process was painful 30 years ago and all of my information is the same 30 years out of date. Any suggestions on least-bad alternatives? The data about 80% of tinder men never getting a single date sounds horrendous. For someone definitely not in the 10% most attractive (maybe around median) - should I ignore all of the online venues completely?
"Really soon"™ we gonna have pleasant AI companions in VR and real life form. And that will make this whole "dating" nonsense as obsolete as horse drawn carriages.
Sure they still be around, but except for a few special cases every one would be using their mechanical replacement .
I liked OKC a lot but my favorite (now dead) dating site was How About We. (It seems like it was also swallowed by Match based on the domain squatting I just discovered.) The schtick was that users suggested dates, e.g. places to go or things to do. I'm not sure how well it'd work now were it, or anything similar, still around.
As a women reading this. This is a very strange article.
A few scattered thoughts incase anyone wants to tell me why I am wrong:
1. It is important not to over state the number of people that meet online dating: https://www.ft.com/content/d30d7d72-b223-4b92-b670-fa3bc4fee2cc. This diagram (although from 2019) says that much more people meet by work, school, or friends and family. Maybe COVID messed this up a bit, but I highly suspect that it will go back.
2. Furthermore, I highly suspect that the women who you will meet online are not representative of the “average” women in a few key ways. (I didn’t find any stats on this). But, from my experience the more self-respecting, less status-focused, healthier, community-focused, women are not on these apps. I have never touched one of these dating apps. None of my friends have either.
I recently went to a large prestigious conference that had a few executives/consultants that worked on these apps, and they were all vaguely apologetic and judgmental of the people who do use them. One of them was a very posh British women who I chatted with for a while, she mentioned over 4 times that “she would never use x-dating app, even though she works as a consultant for it.”
3. All I can say is that the easiest way to get an actually kind and emotionally mature women to like you is to a) have standards and b) have self-respect. Paying to see someone is not something that is attractive to me…. at all.
If you want to get a random person to date you, then maybe this would work. But I would never want to have people pay to have a chance at having me date them. That feels wrong.
I guess if you want women to fall in love with you for your intelligence/kindness/humor etc. then you should try to meet women in a setting where you are showing these things. If you want to meet a women who likes you for your money, then sure pay to see her. (I met my boyfriend in a maths lecture, where I asked him out.. My friend met her boyfriend in church. The list goes on.)
This is the second or third time I've heard mention of Reciprocity since the site got reset. I wasn't motivated to re-do all my box checking, but I guess I'll give it a try again if it's getting this much attention.
(Still not a fan of it being attached to Facebook profiles. Using email address books / contact lists would also solve the network effects problem just fine imo.)
I'm way behind on this blog but slowly catching up. That said, I think you can solve the reciprocity score problem by letting users set a threshold for the AVERAGE of the potential matches' scores. Say you set it to 7 out of 10. Then someone you give a 5 must give you a 9 or higher, while a perfect 10 would only need rate you a 4 to get matched. To make it fair, make it so that the higher threshold of a pair determines if BOTH get notified of a match and only one threshold met is a no go.