Hacker News new | past | comments | ask | show | jobs | submit login

It's the same pipe dream as "AI content creators will take over youtube".

There is no "formula" for success in the creator economy - the winners are largely random. A better way to look at it is there are 4 million humans out there trying every permutation to crack success, and ~400k actually do it.

Unless you have a sufficiently advanced AI agent that is both varying it's content and it's marketing strategy to the tune of maybe ~1000 different iterations it's unlikely we will see a version of OnlyFans that exists that is majority AI generated.

The "parasocial ai girlfriend" sounds like a flawed premise aswell. OF girls are not therapists - Cardi B, Bhad Bhabie, and others aren't raking in millions because they are good girlfriends (although that is part of the upsell). Social status plays a part in the most successful girls, people seem to subscribe because the creator is popular, especially if she's already built a platform elsewhere.

In short, social status does not have an AI substitute.




> There is no "formula" for success in the creator economy - the winners are largely random.

That observation has echoes of the music industry - another extremely top-heavy creator business. There are formulaic ways to make "good enough" and "catchy enough" songs, but the window for "X enough" keeps shifting. Cranking out grunge won't be sustainable in the age of K-pop.

But the massive runaway hits have been predominantly outliers for their age. They have veered far enough from the mainstream to be interesting in new ways, different enough, and surprising enough to break through.

But to predict in advance what kinds of outliers will win the lottery? Largely random, indeed.


That's arguably all entertainment. Fiction writing, art, music, movies, sports, vloggers, influencers, OnlyFans creators, etc. There's a couple brands so established that literally anything they do prints money, then there's a winner-takes-all dynamics that keeps making few randos briefly successful every season, and then there's everyone else who never makes enough to break even.


Yeah but Rick Rubin is always involved and Max Martin has written a hell of a lot of megahits. Makes you think.


>Rick Rubin

Dear God, I've looked into his discography[1] and nearly every album I think of as great from the last 30 years is there. Seasons in the Abyss, The Life of Pablo, 99 Problems, SOAD self-titled + Toxicity, The Geto Boys self-titled, Licensed to Ill... Is this man a hit printer or something? Really shows that Metallica went to him with Death Magnetic after the joke called St. Anger lol

[1] https://en.wikipedia.org/wiki/Rick_Rubin_production_discogra...


His skill is helping creators do their best work. It’s a rare one.


There's a great interview he's got with Anderson Cooper. A fantastic line from it is "I have no technical ability whatsoever". What a guy. Seemed quite likeable.


Jack Antonoff as well. There certainly exist a handful of people who consistently produce hits for decades.


From another angle, a bunch of us in the tech sector made pretty nice salaries. Very few of us were really all-stars in the sense that everyone knew who we we were on YouTube, etc. Which was fine.


> There is no "formula" for success in the creator economy - the winners are largely random.

I think that strongly depends on what you call "the creator economy". For example, on YT it's really mostly skill: https://www.youtube.com/watch?v=Ip2trao6dYw

Not that I believe its easy, nor do I think AI will be super good at it, at least not before everything else also enshittifies into the habsburg-AI-powered dead internet.


> For example, on YT it's really mostly skill

I watched that video from start to finish and disagree with your conclusion. I watched it all so I could make an informed comment but regret spending those 15 minutes on it.

The author essentially made a video about a popular streamer, then went on their stream and baited them with 50$ and a video about themselves. It was literally click bait. It was so transparent that the streamer realised at the end what had happened but still decided to go along with it since it cost them nothing.

That’s just directed spam (which, by the way, is a word the author used themselves). It was one video about drivel. Granted, it’s not dissimilar from the other garbage that populates YouTube, but it also didn’t get views for being good. It’s the equivalent of video junk food. You know it, the creator knows it, yet it’s still hard to stop consuming.


The idea that success is earned through luck rather than merit is a firmly ideological position, regardless of the domain you’re talking about. If you succeeded via luck then that provides a better moral justification for the related ideological position that you should be deprived of the fruits of your labor as much as possible, for redistribution to others who were simply less lucky than you. It’s really just sour grapes.

The formula for success in any field is simply to make a product that other people want to consume. It’s not 0 variance, but if you have some insight into what people want, and you do the work to execute your idea, then you can simply work through the ups and downs and success is almost inevitable.


One of the few domains where this is testable has also demonstrated this. Writing is about as hard to break into as anything, yet Stephen King demonstrated success writing under a completely unknown alias. [1]

No he didn't immediately received the same level of reception and success as Stephen King does, but neither did Stephen King at first! That's why it's skill + dedication. If you look at some of the old videos of people who have succeeded in e.g. social media, they tend to have terrible production quality yet still significantly stand out from the crowd, even their early days. For instance this [2] is one of the first videos Vertasium ever uploaded, 13 years old now! That video, even now still 'only' has 230k views, and certainly had a tiny fraction of that when it was initially released - but he kept at it, clearly putting way more into his videos than he was getting out of them - until that trend reversed.

[1] - https://en.wikipedia.org/wiki/Richard_Bachman

[2] - https://www.youtube.com/watch?v=DBjZz0iQrzI


> One of the few domains where this is testable has also demonstrated this. Writing is about as hard to break into as anything, yet Stephen King demonstrated success writing under a completely unknown alias.

I don't think it actually demonstrates this. As your wording hints, the hard part of writing is getting yourself out of the slush pile and into an editor and publisher's hands, and Stephen King's actions relied on his existing relationship with said editor and publisher to publish under a different name. He never demonstrated pulling the feat of escaping the slush pile again.

In modern content creation, the similar metric is getting to, say, 1k views, or even as prosaically simple as being part of the 50% of streamers to get 1 view. It's not sufficient to have talent to get to even that level of success; there is a lot of luck necessary to get you there.


The mistake you (and a lot of others are making) is that the people who didn’t make it just weren’t skilled enough.

That isn’t true - I think the people who don’t make it are massively skilled. It’s not random in the sense it’s just selecting randomly from the population. It’s random in the sense that there are 100 elite content producers but at any given moment there is only space for 10 of them.

Stephen King has a massive leg up for already having built the inroads for having a successful book. I think if you give any elite, yet unknown writer, the same tools, editor, and publisher they would succeed. But to truly succeed from nothing may just depend on going to school with someone who became an editor, or the editor’s daughter showing them a TikTok. That’s what is meant by it’s largely random.


> The formula for success in any field is simply to make a product that other people want to consume

Well, the formula for success in selling products is this. Most people don't define success in terms of business acumen.

Except, of course, businessmen. If you perceive our society as centered around successful people, of course you'll see it as merit-based. If you perceive our society as poorly run and catering to the rich, of course you'll see success as primarily a product of circumstance outside of your control. Is it so hard to see that "merit" is necessarily defined in subjective terms?


This is just arguing over phrasing. It doesn’t matter what you’re trying to do, if you’re making YouTube videos, or music, or paintings, or cakes, or web apps, or cleaning diveways, your ability to succeed boils down to your ability to provide something other people want. That is the objective source of your merit.

Perhaps your own idea of success in life is something that revolves exclusively around your own satisfaction, like going off and living in the woods. But this is exactly the same situation, you’re just only trying to provide the things that one person wants in that scenario, yourself. Your ability to do this will again come down to your own merit.

Of course if you’re chronically frustrated by being less successful than you would like to be, then looking for alternative explanations such as luck will be an attractive scapegoat that could excuse you from scrutinising your own capabilities. But the human inclination towards doing that is certainly not morally righteous.


I don't think its black and white. I think sometimes success is a matter of luck. For example, in large organizations there can be a lot of roles generated where there isn't always that much direct pressure and people can be hired through luck (e.g. getting on with the boss, some types of diversity hires, being loyal to a company even if you are not that good etc.). If teams of people make products/reports etc. sometimes it can be hard to shine, and 'talkers' who don't contribute much can get promoted into a 'lucky' role. Its not black and white.


You illustrate a perfect example of simply not understanding what people want. Talkers get promoted because talkers have social skills, and companies are social systems, and social skills are required to advance through them. Social skills are probably more desirable than technical skills most of the time. It’s not luck that these people succeed, it’s the fact that they have the qualities that people want.

You can succeed through partially through luck, like if a record executive decides they going to manufacture some massive level of fame for you. But this isn’t a viable long term strategy, only providing what people want is. Over time the variance of luck goes away. The luck outlook relies on the fallacious idea that you only get one opportunity to succeed, but you don’t, you have as long as you’re willing to keep trying. Maybe a failure on one particular day can be explained by luck, but you get to wake up and keep trying every day, and if you have what people want then luck becomes irrelevant and eventually you will succeed. That’s how basically every single successful person you’ve ever heard of has done it.


A good AI girlfriend wouldn't be a therapist but would mimics every aspect of a girlfriend, including arguments and fights and makeups, because that's how bonding occurs. That's going to be how successful AI girlfriend will be made.


Your assumption is that the status quo provides those things. Nowadays, people will break up as soon as they get "the ick" or just have a rotating group of people they see. Lasting relationships are much less common than they used to be because it's easier to switch partners.

People just want to chase a local maximum of constant validation that they're pretty/smart/correct. They don't see or understand the value in working through fights to create something beyond the sum of two people.

AI excels at maintaining that local maximum. It can confidently reassure you better than any human can even if you're wrong. AI partners following this are successful now and people in their teens and early 20s are being hooked en masse.

Historically, superior pieces of technology haven't displaced older incumbents when the learning curve is too steep.

I don't see why a person dating an AI partner that has lovebombed them for several years would switch to another AI (or a person) that starts fights and bickers. Even if it's better in the long-term, that's still a marked decrease in short-term satisfaction.


The whole point of having fights and arguments at just the right level is to maximize engagement, retention and ultimately making money for the corporation.

I was imagining the most diabolical addictive AI girlfriend. That's necessarily going to include 'negative' elements.


Implementing the cycle of abuse in an AI partner could be as impactful as the invention of the cigarette.

I'm now very concerned about hypothetical young men who enter into relationships with AI in university or high school, then graduate and have an algorithm abuse and take their money.


Your AI girlfriend that goes from crisis to crisis but with microtransactions.

"I need $34.99 for storage space or they are going to delete me, please save me white knight!"

"The met a nice guy yesterday and he was able to afford my premium package, the one that lets me feel more emotions, I just don't know if I feel for you like I once did..."


I completely agree with your point, if it is that ai will be twisted to generate a significant other that will essentially become addictive. I get very uncomfortable thinking about that reality.


> people in their teens and early 20s are being hooked en masse.

Any reference for the scale of this? It feels unlikely to me from my bubble but I only know one or two people I think would be likely to try it.


https://www.ark-invest.com/articles/analyst-research/is-ai-c...

The biggest callout is that NSFW AI already has 10% relative market share compared to OnlyFans. And there are no frontier models in that market.


There are many successful relationships that don't involve arguments - and which are about constant peace.

Relationships don't require 'arguments and fights and makeups' to be real. And if AI girlfriends offer 'ideal relationships', how is that not 'good'?

You are conflating what people actually want with the artificial drama of TV shows and Hollywood/the messy scenario of reality. If people can pay to get their fantasy girlfriends/relationships brought to life, they will, and it will be successful especially if all forms of conflict/relationship dissatisfaction can be avoided.


There are many successful relationships that don't involve arguments - and which are about constant peace.

I am not saying things about successful relationship. I am merely pointing out how exploitation of users can occur.

Emotional bonding often occur in orderal and other challenging events. It is one of the tools that companies will use to push users' button and to exploit them for economic value.

And if AI girlfriends offer 'ideal relationships', how is that not 'good'?

Ideal relationships aren't necessarily good for AI companies' pocketbook.


Bonding to a computer program under control of a corporation is like looking for a sociopath as a partner explicitly. You would lose complete control of your life to the other side. Reciprocity is off the table completely.


Nobody is going to pay for an AI girlfriend service for it to breakup with the user and refuse to get back together - because that's how growth happens in reality.

What AI girlfriends will do is mimic perfect Hollywood relationships, complete with hot makeup sex.


Isn’t there a rule on the internet that says “if you can imagine it, there is porn for it” and “if there isn’t porn for it, somebody is making it”?

I’m pretty sure that applies to this scenario too. I’m 100% sure that there exists a set of customers who would pay good money to get dumped by a realistic AI girlfriend. And once dumped they’ll turn around and pay for the next AI model to dump them only in some other fashion. Maybe the AI model thinks the customers anatomy is the wrong dimensions? Maybe they smell? Maybe they are too short or tall? Perhaps the AI “girlfriend” is a triple tentacled sea monster who wants to return to oceans on Titan? Doesn’t matter. Somebody will pay very good money to experance it.

You want a hot quad breasted space babe who cheats on you with bubble wrap covered little people? Done. Want that with extra bondage? Done.

This is the internet after all. Why pay for a boring “normal” AI girlfriend when the sky is the limit? I say, use your imagination.


[flagged]


I would expect them to have a greater desire for bonding than someone who has already been through a dozen relationships or possibly a divorce.


Why? They're common people, not sociopaths.


I think maybe it's my own personal bias, but it does feel like anyone who pays money for an AI girlfriend really won't want it to disagree with them. I believe they'll want an idyllic and fantasy version of a relationship.


AI girlfriend that always agree with you and never contradict anything you say isn't going to be as addicting as an AI girlfriend who on occasion disagree with you.


While I grant that some and perhaps even most people won't want the AI girlfriend to disagree, there are some out there who treat arguments as a necessary and desirable spice in a relationship.

That said, I can't really think of anything that would be worth arguing with an AI over.


you are saying that the ai gf will not be like a real human female, but you are not making any argument that there is a defect in the bf's attachment to the gf.


I dunno how sustainable that scenario is, because it brings to mind the maybe apocryphal Chinese curse "may you be granted everything you wish for".


Agree. The only advantage of an AI girlfriend is that you can shut her off when you're finished with her.


I'd say the set of people who want an AI girlfriend and the set of people whose defining trait is a lack of empathy is probably a bigger overlap than you think.


I don't see how this is true. I would expect a man who accepts female rejection and stick with AI to have more empathy than the women who rejected him, because e.g he is ugly.

Harassing men for their lack of dating success and ascribing negative personality traits simply because they are down on their luck makes them more resentful so you should stop doing that. It's called having empathy.


> I'd say the set of people who want an AI girlfriend and the set of people whose defining trait is a lack of empathy is probably a bigger overlap than you think.

The defining trait of people who want an AI girlfriend is ugly, lonely men.

Is lack of empathy the defining characteristic of ugly men?

Or is it simply that people in general hate ugly people and thus ascribe various ills and character faults to them?

I'd say the set of people that disparage undesirable lonely men that desire AI girlfriend are usually those whose defining trait is severe lack of empathy. This is blatantly obvious here.


>The defining trait of people who want an AI girlfriend is ugly, lonely men.

average men. 20/80


Which, as rated by average women, is almost the same thing....


I know plenty of ugly men in committed, loving relationships. You're the one bringing up looks. I don't think it matters what you look like; if you want to date a computer you have some kind of psychiactric disorder.


Ok, but then there is also the increasing number of women who don't want a boyfriend either because they want a same sex relationship or because they can't find a man on their educational level, since women graduate at a higher rate. The reason is mostly irrelevant, what matters is that supply and demand are imbalanced and one side has to deal with not getting what they want in one way or another.


Education level is less important than equality is.

Marriage is traditionally a terrible bargain for women, but it was the only choice they were allowed to make. Now, they can make their money and buy property and have kids on their own.

A lot of men haven't realized the era of the provider is over and dead, and they're now optional. They have to make women want to be with them, and a lot of women just aren't willing to compromise on equality these days.


Good luck buying a family home as a single mom with today's property prices. Really it's a rich person thing, not available to ordinary women.


It may seem like this makes sense, but in most places it's the opposite. Low income women are single mothers at a higher rate than women with high salaries.

In most western countries, low income women do not become much poorer if they become single mothers. In some places, it increases their living standards. But for upper middle class families, a breakup tends to be costly.

Also, there seems to be shared causal factors that lead to both stable relationships and financial stability. Such as impulse control, mental/physical health and the ability to postpone gratification.


You don't need a house to have kids. Plenty of people don't and they do just fine. And if they're choosing to have a kid on their own, they've planned out finances, too. Fertility treatments aren't cheap.


>lot of men haven't realized the era of the provider is over and dead, and they're now optional.

Cool, child support and alimony optional now. Right? Because it's always easy to be independent with OPM. People forget single moms became far more practiced after the state incentivized breaking up families.


Why does that matter if they don't marry anyone? They control the means of reproduction, after all.


Whether they have a marriage certificate might not matter, but the stats are clear that children raised by single moms are correlated with a lot of bad outcomes including far more likely to end up victims of the prison industrial complex. If you only give a shit about yourself and not your offspring, maybe that doesn't matter


Just because "plenty" of ugly men might end up with something, does not pose a contradiction.

Nobody "wants" to be in a romantic relationship with an ugly person or computers.

Now just because some people end up doing that, does not mean that it is what they want. It might just be the best option available to them.

Given the choice however they would always rather have someone handsome and gorgeous instead.

Looks are supremely crucially important for somebody to seriously crush on you, be genuinely infatuated with you and have a deep burning desire.

An ugly undesirable dude or ugly fat chick is not ever going to experience this, surely they might (or might not) end up with a pleasant relationship of convenience... but it is not what they want. Works of fiction and computers are the only way they can experience a glimpse of desire and infatuation aimed at them.

This really isn't that hard to understand for people who have empathy.


I'd say lack of empathy correlates better with people that ascrible lack of empathy to groups of people they don't like.


lack of empathy does not stand in the way of forming an attachment. Empathy is a good feature for keeping a partner, but the ai doesn't care about that. It feels like people on your side of the argument are making a sort of moral/judgmental argument, "if you can't hold up your side of the bargain, you don't deserve an ai gf"


If you can only bond with computer software and not other humans, there's something pathological going on there. You definitely have some severe issues that should be worked out in therapy.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: