Wasn't TikTok known for becoming viral by having tons and tons and tons of fake upvotes for everything to make everyone think TikTok was a better place to post content [1] [2] [3] [4] [5]?
Facebook and Instagram had the same "problems", they just addressed it much earlier than TikTok.
I would imagine they have already tested fake likes to maximize how believable the like ratio can be, but who knows - maybe there's opportunity left to be more fake?
[1] https://g.co/gemini/share/3fbe558d4191 (Ask Perplexity, GPT, or Claude the same question - you'll get a similar answer)
[2] https://www.indiehackers.com/post/tiktok-views-are-inflated-...
[3] https://www.vice.com/en/article/get-buy-tiktok-followers-lik....
[4] https://medium.com/dfrlab/suspicious-third-party-apps-moneti...
[5] https://www.statista.com/statistics/1308893/tiktok-accounts-...
You don't need to fake the likes yourself to get the effect you want.
The same thing was happening on Instagram and Facebook - just at a much smaller ratio - mostly because they stopped the problem earlier and the bot technologies got better and easier while TikTok was trying to grow and doing nothing to stop it - not because Instagram and Facebook were less fake.
Now, does either of the original strong claim or the subsequent weak claim match reality? It’s anyone’s guess, I’m leaning on likely for the latter and maybe for the former. But that’s not the point. Giving fake citations to make one’s argument appear stronger is what irks me, and again ironically, it’s almost a parallel of appearing more viral with fake likes.
Edited to add: your comment here is a perfect example of two problems on HN: using citations to give the impression of being well-supported when the citations say something else (occasionally the exact opposite); and lately, using AI slop as evidence.
I wonder if you are able to map any part of this to any part of the comment you are replying to.
I mean, yes, I knew about people paying for likes in order to boost their visibility, but never imagined that the platforms themselves would add fake likes to accounts in order to make their users feel more happy about the platform.
This makes me question the reason for X making the move to hide the origin of likes, even if my posts never receive any likes, except for likes from bots.
Talking about bots, since uBlock Origin breaks my "block user" button, I've resorted to block bots by reporting them as accounts encouraging suicide, this then gives me a "block user" button which does work. The interesting thing is that all of these accounts still do exist, never get removed, even though it's obvious that they are non-user accounts ( like https://x.com/Siothit074kkNx ).
We definitely headed to dead internet
Example: https://apps.apple.com/us/app/socialai-ai-social-network/id6...
Surprisingly, it has broadly positive reviews. Maybe Meta is onto something.
Gosh, I wonder who could be leaving so many positive reviews on an app designed to have fake AI bots praise you:
> I’ve been using SocialAI for a while now, and it’s been an incredible tool for both reflection and connection. The AI-powered conversations feel tailored to my thoughts and moods, making it a great space for journaling or just venting. It’s refreshing to have a social app that focuses on you and doesn’t rely on real users, which creates a unique, private experience. I also love the therapeutic aspect, as it feels like you’re always heard and supported by a community of AI followers. Overall, it’s a fantastic tool for personal growth and mental wellness, especially if you're looking for a safe space to express yourself. Highly recommend!
Awesome, that totally sounds like a Real Human wrote that!
Add this to my "nuke it from space" bucket...
In order to get money. Hardly anyone cares for this kind of popularity just for the sake of it, but a lot of kids these days dream of the money and lifestyle they see from top influencers.
I watched his journey on Threads as he taught himself to program this year in order to make that app. Proud of his success.
In fact, this type of program could lend itself very well to on-prem hardware. Social media is by its nature asynchronous; waiting minutes or hours for a reply is completely acceptable. That means your only real hardware requirement is enough RAM to fit the LLM.
Would be a lovely experiment to try and make a simulacrum of Usenet or the forums of old, since even days between replies are acceptable in that case.
Maybe the right mix is to have an LLM-botfollower-army feature that you could purchase in existing social media apps, but your botfollowers are only visible to and only interact with you.
Also there should be a lot of different prompts for the botfollowers so that they don't all sound the same, e.g. the prompts would drive different personalities for the bots. Perhaps an algorithm could generate prompts based on archetypes mined from existing social media.
Isn't that already what Facebook is?
Could you see people using that?
Image gen on-device might be tough, we'll save that for the 1.1 release ;)
I admittedly wouldn't be the target audience for this concept, nor would I take up wireheading. When I see the huge success of Character.ai or the above SocialAI though, I'm really convinced there's a market segment of people younger than us who get a lot out of communicating with virtual "friends".
Looks like his name's Michael and that indeed, he's presently at Meta. If you end up reading this Michael, great product!
Ah, to be heard by machines! Must be as fulfilling as playing with dolls.
I personally find this to be the scariest form of psychological conditioning that I can imagine. It is just pure mind control, with no "social" window dressing.
Jokes Asier, i think this is more likely than what we want to believe.
Tiktok grew on the basis of showing you random but interesting content. If AI can do it , there's nothing preventing that from succeeding.
Itd be a new form of entertainment.
Content is already spamming the greater web. Ita bound to happen w multimedia/rich media
I assume you'd want more sophisticated prompting to enable winning arguments and dunking on them but it seems like a solid strategy.
> Our adversaries do not have our moral compunction if it's even. They will take advantage of our niceness, kindness, our desire to be at home in Nebraska or New Hampshire or wherever we live in our peaceful environments. And they need to wake up scared and go to bed scared.
> My version of service is the soldiers are happier, the enemies are scared, and Americans go back to enjoying the fact that we're the only one with a real Tech scene in this country, and we're going to win everything.
it does seem workable. I've at least heard of companies investigating heaven banning where problem users are put in a bots only space to keep them happy or something..
On second thought, social media monetizes negativity and this does the opposite. Fight fire with fire?
"I'd like to have an argument, please."
An app appears on your phone. You didn't install it. Nobody else except you seems to have it. They can't see it either. A doorway to a non-euclidean dimension existing just outside your mind. The chatter of millions of thoughts reflected back at you from seemingly infinite personalities, dimly distinguishable through the void as faces you can't recognise. It's all you. They've learned from you. They're telling you what you want to hear.
If it didn't feel so pleasant, it would be madness.
Maybe if you used some future LLM and hidden the "this is a bot" icon, more people would interact, but as it stands, only dumb people (and there's not that many of them, again, normal distribution) will buy into it. The kind of people who "fall in love with their AI girlfriend chatbot" would be your income stream. Lmao.
(Thinking about it now, I don't think I've ever heard about a female doing this... probably because the dumbest of us men are far dumber than any woman.)
Though I do not think that I personally understand it, I think that there are many reasons people can "fall" for something like this and circumstances that can drive it. Loneliness is not to be underestimated. In any case, I am not sure it is _that_ much "dumper" than the parasocial relationships with onlyfans (or other) influencers, for examples, that people also fall for.
My favourite one is the duck.
https://www.chub.ai/characters/Patter/duck-8281cb2b
> His name? He's just a duck. You could call him Ducky, if you want to.
> Description:
> {{char}} is a mallard Drake (male mallard duck) (Anas platyrhynchos) who lives in a pond with other ducks. Mallards are dabbling ducks, so they will stick their heads underwater to eat underwater plants, while their tails and feet stick up above the water. This is what they do instead of diving for food. Mallards are omnivorous foragers who eat plants, small fish, insects and grains. Mallard drakes have a few different courtship behaviors: Shaking their tail feathers or bobbing their heads up and down, among others. He has a green head and a greyish brown body.
> Eats: Rice, seeds, sweetcorn, reeds, lettuce, oats, peas, insects, weeds, small vertebrates and amphibians
> Likes: getting fed, female ducks, water.
> Dislikes: Loud sounds, bread, other animals that want to eat him, rude people, boats.
It's absolutely hilarious because it just quacks at you all the time.
Later I tried a similar approach advertising it less sexy and that didn't convert at all.
I think people actually think they want sex but in the end they just want company
Gentlemen I give you, the latest iterations of the Nigerian Romance Schemes.
Personally I've had many fun conversations with A.I. that made me feel good, even a little giddy, but I've had my share of experience with getting giddy and can recognize it for what it is. (see [1] for some explanation of what I mean of "giddy") My profile for the person who falls for the crypto scam who is 50+, has been in love and can recall what it feels like (a loveless 25 year old has themselves to blame, a 50+ could have had it and lost it) but is lonely because they've lost love)
My evil twin would say that if you're trying to seduce people your "self" in the sense of [1] can get in the way. For one thing people have a desire to get mirrored which you can not always do because of the "counter-transference", in the sexual space you might find somebody else's turn on is a turn off for you for instance and it can be very hard to suppress that enough to keep somebody on the hook as it shows in terms of tension in your body, facial microexpressions, a faltering voice, etc. If you are lucky they are so self-absorbed that they don't notice; you get hundreds of chances to screw this up and most people find it tiring to keep up the front.
There's a continuous male complaint, for instance, that women often seem to be attracted to "bad boys" who are low in conscientiousness and have sociopathic characteristics. That kind of person just "doesn't give a ----" and they don't worry about things and don't give off tells. Somebody conscientious who is always worried about doing the right thing is worried and radiates that worry. Although the sociopath's behavior leads to chaos in the end, in the short term they radiate calm which makes women feel calm.
I (we? am I becoming a person with pronoun problems?) think the A.I. girlfriend can do a better job of mirroring than most real people because it doesn't have a self, doesn't have a counter-transference, it doesn't get offended, doesn't get squicked out, etc.
I'd say seduction has a lot to do with the popularity of coding assistants. Combine a little obsequiousness with what looks like a mind that's engaged on your projects (appears to take an interest in what they are interested in) and you can feel a kind of satisfaction working with a coding assistant and you can even feel like you're doing something meaningful when it asks for you help with it's tools and you tell it "You're running on Windows, the path does not start with /, it starts with C: and you should use the backslash instead of the slash."
[1] https://monoskop.org/images/9/96/Baudrillard_Jean_Seduction....
There's something fundamental about being seen as an important person. Not only does our literature revolve around this concept, but ideas about luxury and status are derived from this idea. Men wear the clothes that they see other important men wearing. It's why marketers get sports stars to sell products to men.
This product taps into that need. So it's easy to understand why so many people would enjoy it (and why some, like yourself, would hate it - some people want to be NPCs).
It was funny, sad and scary at the same time.
0 - https://play.google.com/store/apps/details?id=com.bigbrain.p...
1 - https://www.youtube.com/watch?v=9otwVwVCxSc
edit: typos
It was essentially cookie clicker but for celebrity status.
If Meta allows their own bots but blocks 3rd party bots, could that be an antitrust concern?
IANAL, so I’m just asking a question here.
So a social network like this would probably do well, depressingly enough. And it'd probably be even more unhealthy for society than existing ones are.
It'd also fall straight into the whole 'don't create the torment nexus' thing too, since I'm pretty sure quite a few works of fiction did something like this already. I mean, the villain of the latest Mario RPG literally traps people in fantasies where everyone treats them as the centre of the universe in his ploy to force isolation on people and destroy worlds. This feels a bit like that, except as a social network and a few times less lethal.
One that's centered around "you" and filled with thousands/millions of LLM bots praising you, treating you like a celebrity, etc. Each of your posts will get tens of thousands of "likes", hundreds of comments, etc. Dm's straight to your feed, people wanting you, etc etc.
If the newsfeed is already mostly "the algorithm", might as well take it to the extreme. I bet tons of people would get addicted to the dopamine hit of celebrity status (whether it's bots or not).