The dating era in the sense that we need it today started in the 1920s. After a long period of urbanization, families moved to cities. Dating moved out of the family home under the supervision of parents and young people got the freedom to go to public places without chaperones. They started visiting restaurants and movie theaters, dance halls and this changed the power dynamic. Men now paid for the experience, taking the financial responsibility that was previously held by the hosting family and women decided the terms. A typical date... cost well under a dollar. For context, a typical weekly family budget... was around $25 in 1920... So, a proper date would cost around 4% of a weekly spending.

During the post-war era, the dinner and movie template became the default date format... A cheap date in 1960s was typically two hamburgers, two sodas, and two movie tickets... that comes out to about $2.65 total. Roughly $2 for two tickets, 30 cents for two burgers, and 30 cents for sodas. A nicer restaurant date... would run $8 to $12... the median American family income in 1960 was about $108 per week before tax. That means that a cheap date was around 23% of a week's income... So dating was cheap in absolute dollars compared to today. But it got a lot more expensive compared to the previous era relative to earnings.

In the 80s and 90s... was the first time when the haywall showed up. People who couldn't find partners through social circles would turn to newspaper personal ads... But this time, the first video dating service showed up. The service called Great Expectations, and it charged $1 to $2,000 upfront for access to video libraries of potential batches. This was the first time when dating had a subscription fee attached to it. The service planted the seed that paying for access to a curated pool of potential partners should cost money.

Moving into 1995... the first internet wave, match.com, launched in 1995... this subscription model was borrowed from magazine subscriptions... E-Harmony launched in 2000 with a different positioning... long for questionnaires, and premium pricing, $45 to $65 a month. And they justified that price by positioning it as an investment in your future life partner. Then OKCID showed up and launched in 2004 as the free alternative... Tinder's 2012 launch borrowed the swiping mechanic from a card sorting exercise in user experience studies and the reward mechanism was the dopamine loop borrowed from slot machines.

The economics of today's dating ecosystem are strikingly different from the old dinner and movie model. The average date in 2026 costs roughly $200. And mind you, that is a 12% jump from 2025. And costs keep rising. There is even a separate term coined for it called dateflation. And that's before the app costs. A typical dator today that sits on three different apps with a paid subscription... spends around $100 a month. And if, for example, you're on e-harmony... you're looking at more than $2,000 just for the subscriptions. And if you look at the studies, more than 47% of singles say that dating is no longer financially viable. Historically, the date itself was the expensive part. But today, if you're serious about dating, you have to get on premium versions of those apps.

The dating app industry has always had a conflict of interest with its core audience. It is obvious that a platform technically profits when users don't find a match because a successful match removes paying customers. You may ask, "But how's that different from any other business... like a car mechanic?" Well, that comparison is not fair... Unlike a car mechanic that delivers value all at once, the app delivers some value at certain points. And when a user gets a dopamine hit, be it a date or... even a message, that value has a lasting effect.

So AI profiles... the point of AI profiles is that they extend time on the platform. Yes, dating apps reportedly benefit from AI profiles despite the fact that they're not real because those profiles create and stimulate activity. If an AI profile looks believable enough for you to engage with... it doesn't matter if you end up going on a date or not. You keep staying on the app because the interaction keeps going and all of a sudden it's the last day of your subscription. But hey, all those chats can't just sit there. What if he or she is the one... and the next moment you find yourself paying for the next month hoping that something may pan out?

AI assisted profiles are different... they optimize your bio, your description, your prompts. They suggest how to write a response... analysis of your photo gallery to gauge what you're into and so on. Now, on one hand, this can be a life-cher for people who struggle to present themselves... On the other hand, the obvious consequence of this is that it works universally across all profiles. So when millions of users run their bios and conversation openers through the same LLMs... the variance in dating profiles collapses and they become very homogeneous.

Another problem is catfishing. Pre-AI catfishing was fairly straightforward... You could reverse image search... After GenAI, catfishing really grew in scalability because AI profile photos are in fact original. They're not copies. You can't reverse image search them. And on top of that, add the LLM bots that can sustain a natural and contextually appropriate conversation. They can do voice cloning. They can be even used during video chats. And deep fake videos are also getting pretty convincing.

There are multiple studies... that describes a pattern of dating apps exhibiting a measurable popularity bias. Popularity bias is a phenomenon in recommendation systems... the algorithm disproportionately suggests popular items... especially when you're just starting to use them or coming back after a long time. You get the cream of the crop at the very top to lure you in. And the interesting thing is that these profiles directly correlate with maximum engagement and reverse correlate with successful matching. The FTC established that Match Group sent emails from fake accounts notifying users that someone likes them. And that notification alone drove nearly half a million people to resubscribe.

Dating apps... work as a two-sided marketplace... but in dating apps, there is heavy supply-demand asymmetry. Tinder is 75% male. Bumble is 64% male. Hinge is 60% male. On Tinder, men swipe right on about 60% of profiles. Women swipe right under around seven. Men's match rate is around 3%. Women's is 45. This is why Tinder's algorithm is specifically designed to keep women engaged and active because dating marketplaces create extreme scarcity for men and scarcity is what drives spending. This is why the economics of the dating apps is skewed towards keeping fake profiles because removing them has immediate costs.

The newest thing... is mold match, an offshoot of OpenClaw... allowing AI agents to create dating profiles and initiate romantic conversations on behalf of their users. Bumble explicitly plans to test removing the swipe mechanic entirely and replace it with chapter-based profiles in a vertical storytelling format... there's also some talks about metaverse dating... the idea is that there would be no human faces and users would swipe on avatar profiles... If you want my personal view on it, I don't think this is going to go too far... I think that looks and appearance and physical traits are integral in mating behavior.

The biggest problem in metaverse dating is the gender imbalance. VR demographics skew heavily male... you run into the same problem with scarcity. The irony that the more sophisticated these apps get, the more they expose a problem that has nothing to do with technology. We are getting lonier and we long for connection and dating is dating whether online or not. It's confusing, frustrating, entertaining, romantic, and occasionally humiliating as it's always been.