Discussion
The dead Internet is not a theory anymore.
d--b: Well, it's not such terrible news, is it?I get nostalgia for the 90s/00s, but that time was never coming back anyways.The best we can hope now is for people to be less online. And if it comes from people drowning in AI crap, I think it's kind of funny.
righthand: I don’t think people are going to get offline and the best we can probably do is create free and open p2p platforms that don’t require registration necessarily to use. And allow people to control their own databases. A lot of communication is locked behind corporations that run the services that are building these tracking and identification databases.I actually think it’s more about getting people off browsers and other tracking software.
august-: do you think small, invite-only communities will end up being the last holdout for genuine human conversation online? or will bots eventually infiltrate those too?
floatrock: Bots will absolutely infiltrate them eventually, but I think it's the only solution.Internet promised ability to connect with anyone anywhere around the world. It felt limitless and infinite.Turns out in an infinite world, the loudest voices are the ragebaits, the algorithmically-amplified, or the outright scammers.Human social brain doesn't work in an infinite world, it works for a Dunbar's Number world. And we all like our psuedo-anonymous soapboxes (I'm standing on one right now), but trick will be to realize that the glitter of infinite quantity isn't the same as small-scale connection.
bsaul: I only see two outcomes for this problem : an internet of verified identities (start by uploading your ID card). Or a paid internet, where it doesn't matter who you are, but since you're going to pay for that email or that reddit account, the probability that it's AI spam is greatly reduced.And i'm looking forward to none of them.
kubb: I want cool cryptography where I can, e.g. verify where I'm writing from and what my age is without giving away any other information.Or if I want, I can verify that I'm myself, and eschew anonymity, and certain platforms should only accept contributions from people who don't hide their identity.Everyone knows who you are in the town square.
nathancahill: I think this was the premise of Keybase?
daedrdev: It's absolutely why discord is popular.
retrocog: IMHO, this is the exact instinct and there's a way to verify identity, location, and age without even having to share those directly.
jeandejean: Next step is: we get back to speaking to each other in the real world. That would successfully close the loop.
sebastiennight: I think this will be a tiny minority who is already currently not terminally online (so, no big change for them).And the vast majority will just be driven to more AI-mediated interactions.
tines: I think it's a symptom of being terminally online to think that most other people are also terminally online. The internet has a way of convincing you that most of the [interesting] events in the world happen on the internet. But I think this isn't the case; most stuff happens in the real world, most people live in the real world most of the time, and a tiny fraction of trite drama happens online.
ticulatedspline: Dunnow, everyone is in a bubble of some sorts. I'm online a good bit but rarely on my phone, If I'm away from my desk I'm offline. My social circle is similar so I would naturally have bias for what I experience.Year or so ago I took an Uber and was mesmerized by the driver. He had his phone up mounted on the left and was pretty constantly interacting with it. Checking for new rides, watching a video, checking facebook. It was quite impressive how much content he consumed while at a red light and how dexterously he navigated to and through like 10 different apps.I very much got the feeling that this was a person that was terminally online and suspected that he's not alone. A bit alienating really, living in the same country speaking the same language but realizing there's this huge cultural/behavior divide between us.
gr8tyeah: I pay for my ISP and the financial institution the money comes from has age verificationSocial media, HN and the rest of internet first business can go brokeI don't see anyone out there propping me up directly. Why would I give crap if some open source hacker or etsy dealer doesn't have a home next month? Yeah I don't because they're not caring in the same wayThoughts and prayers everyone else but your effort is clear, not going to be 1984'd into caring for people who clearly don't care back.
yakattak: Honestly the $10 barrier to SomethingAwful back in the day (and I guess now since it’s still around) definitely made a huge difference. I hate the idea of subscribing to a site like HN or Reddit… but one time $10 to post? I’d accept that if it meant less bots.
geek_at: reminds me of Bill Gates in the 90s when asked about email spam. He said it would make sense to make an email cost like 1 cent so the spammers can't spam as much but this didn't sit right with the mindset of the people at the time.
selfsimilar: Hashcash was a proof-of-work system that would have put a computational tax on email. I don't know what kept it from getting more traction other than simple chicken-and-egg network effects, but it's a good idea, and worth resurrecting.http://www.hashcash.org
amiantos: Why is it being called dead internet theory when, as far as I can tell, what's really happening is that big centralized systems are being overrun with bots? The internet existed and was pretty great before these large centralized systems came into being.Anyone can still run a blog/website, and/or their own discourse server. There's no need to mourn for these centralized systems that largely existed only to exploit us in some way. Let's celebrate "small internet theory", an internet where exploitation is effectively impossible because every company that tries it is overrun with AI bots. That sounds awesome to me personally, but I was also up late last night watching clips of Conan O'Brien from 1999 and the nostalgia for that era / what the internet was like back then hit me so hard it was almost painful.
ambicapter: Anyone can run a blog/website and be subject to AI bot crawlers using terabytes of your bandwidth for no reason, yeah.
ticulatedspline: Kinda been dead a while, also not dead there's still good stuff out there. Lots of it, but it's in the corners and under the carpets. Things created in the original spirit of the "let me show you my interests" that the older web was built on.While back was toying with the idea of building out a new web on a new protocol (not http based). Thus no existing browser would understand it. Deliberately obscure to force a "Reset" button of sorts.Though would be short lived, over time we've learned to ruin stuff faster and faster. I'm not sure there's any network so alien that it could hold on to that golden era of innocence from the past, it would be found then expediently and expertly exploited.
moritz: https://geminiprotocol.net/docs/specification.gmi
dehsge: Members only comment blogs. Where you need an invite to comment also solves the problem. You need to know a real human to get access.
NoMoreNicksLeft: >Where you need an invite to comment also solves the problem. You need to know a real human to get access.Bittorrent trackers, as absolute retarded as they are, have performed this experiment for us and the lesson we're supposed to learn is that this does not work. Someone, somewhere, has an incentive to invite the wrong sort eventually, which because of the social network graph math stuff, eventually means "soon". Once that happens, that bot will invite 10 trillion other bots.
aerhardt: I have to use LinkedIn to sell. I only occasionally look at the feed but I am ruthlessly muting or blocking anyone who is blatantly foisting their AI drivel on other humans. I’ve had enough of this shit.
robotswantdata: unfollow works too
hinkley: Verified entities defecting by using AI to generate their content for them will break this.
Pigalowda: We can use verification mountain dew cans. No big deal.
hinkley: I can't recall the last time I did the Dew. Should I turn myself in to a reeducation camp?
rpcope1: I recall a WSJ article during the 2024 election that was about the fact that Tim Walz and JD Vance were both big consumers of Diet Mountain Dew, and how basically America ran across the board on various types of Mountain Dew. Can you really call yourself "American" if you're not doing the dew?
pixl97: Dead internet prophecy.
pixl97: >create free and open p2p platforms that don’t require registration necessarily to useAnd how do you create this without it being overran by bots, spam, and people posting gargantuan amounts of porn?>allow people to control their own databasesThere are two types of people that want to control databases. 1: The freedom seeking type who want information sovereignty. 2: The type of people that want to hoover up as much data as possible for money and power.Guess who has more ability to control the world out of those two.Lastly, most people want to use curated websites free of spam and content they don't want. Almost nobody wants to do that curation themselves. Hence curated platforms will attract the most people via network effects.ah shoot, that wasn't lastly...> getting people off browsersand putting them on what exactly? phone apps, that's not better at all. Multimedia attracts people like flies to poop. It's seemingly a natural human response to move to an application that is more visually interesting regardless of it's security safety.
robotswantdata: How would the small internet fight the bots?Aggressive moderation? Disable UGC?
MeetingsBrowser: That might raise the initial barrier, but it assumes every user behaves appropriately.All it takes is one invited user to open the door to bots.
ahofmann: Bandwidth is only expensive in the US, somehow. Here in Germany I didn't bother about bots and their additional traffic since 1998 (there are other annoying things about bots though).
JumpCrisscross: > Why is it being called dead internet theory“A social networking system simulates a user using a language model trained using training data generated from user interactions performed by that user. The language model may be used for simulating the user when the user is absent from the social networking system, for example, when the user takes a long break or if the user is deceased” [1].(More seriously: https://en.wikipedia.org/wiki/Dead_Internet_theory)[1] https://patents.google.com/patent/US12513102B2
amiantos: So why isn't it called "dead social media theory"? The internet is not only social media services, though I understand a lot of people seem to think that without centralized social media services there is no reason to use the internet.
pixl97: Have you been on the internet at large lately? With google you may get one authoritative site on something and 50 bot copies of the site on different domains. Sometimes the stolen site is the number one return. Also, if you ran sites years/decades ago, you realized way back then the any local user posting was getting overran by spammers/bots. Now is so much worse that it's not worth doing in most cases.So, most posts on social media aren't real.Most user posts on non-social media are spam/not real.Most websites in searches are copies/ad spam.So yea, dead internet reality.
amiantos: I spend all day every day on the Internet and I don't share your perspective. I might dislike centralized social media and yearn for a bygone era, but just in the past two days I had a very positive interaction with multiple real humans in the Commodore 64 subreddit that helped solve a problem I was having that isn't documented anywhere else on the internet yet. So then I went on my personal blog and blogged about it, which will get it out there on Google and help others. In this way, I am helping to keep the internet alive, I guess. "Be the change you want to see in the world," and all that.
dylan604: > So then I went on my personal blog and blogged about it, which will get it out there on Google and help others.That's some of the boldest optimism I think I've seen in awhile. Maybe your blog is more popular than I assume, but still
ekropotin: What stops someone from handing over their idendity’s private keys to an agent?
soco: Switzerland just voted recently to officially implement Selective Disclosure JWT, which does exactly all that. Social network registration can ask "are you 18?" and run with that - and only that. Or the club entrance. Or whatever, because it's all controlled by yourself in your app.
TacticalCoder: > Members only comment blogs.There, sadly, needs to be some gatekeeping and then it can work.For example I'm member, since years, of a petrolhead forum where it works like that: a fancy car brand, with lots of "tifosi" (and you don't necessarily want all these would-be owners on the forum). To be part of the forum you must be introduced by some other members who have met you in real-life and who confirm that you did show up with a car of that brand.If you're not a "confirmed owner", you can only access the forum in read-only mode.It's not 100% foolproof but it does greatly raise the bar.It's international too: people do travel and they organize meetups / see each others at cars and coffee, etc.Or take a real extreme, maybe the most expensive social network: the Bloomberg terminal. People/companies paying $30K/year or so per seat each year probably won't be going to let employees hook a LLM to chat for them and risk screwing their reputation. Although I take it you never know.It is the way it is but gatekeeping does exist and it does work.
floathub: Emacs will solve this too:https://github.com/tanrax/org-social:-)
empath75: Reddit in particular is overwhelmed by bots. There are small niche communities where it’s mostly people talking to people, but the vast majority of popular posts are made by bots, voted on by bots and commented on by bots.It’s not even like commercial astroturfing, it’s just karma farming and public sentiment manipulation.
demaga: I think it makes sense, since most people don't post anything or at least don't post much. So someone (something) must fill in this void.
Krei-se: Your welcome to link your proven human page with mine.Id even run a dedicated UT99 server lol
Invictus0: You could have easily said this twenty years ago when photoshopped photos were going viral on the early internet. Turns out people are completely fine with ai content and photoshop.
crab_galaxy: Fine in what way? What people?I have not seen or heard of a single person who is excited about AI generated blog posts, or TikToks, or commercials, or images. In fact it’s the opposite, the internet coined the term AI slop, and my non-internet addicted friends hate the fact that chatGPT is killing the environment.The only people I’ve ever seen champion AI are the few who are excited by the bleeding edge, and the many many peddlers
Invictus0: There are probably more than 10^17 AI model executions occurring per day. I know in ye olde HN there are many Purists that are Too Good For AI, but the majority of the human race is consuming AI at a blinding rate, and if they really didn't like it, they would stop.
girvo: > and if they really didn't like it, they would stop.I can’t really articulate why, but this doesn’t feel true to me. There are plenty of things humans do especially at scale that we don’t like, or we do that we don’t like others doing, and don’t stop
pixl97: >The "Moloch problem" or "Moloch trap" refers to a scenario where individual, rational self-interest leads to a collective outcome that is disastrous for everyone. It describes competitive, zero-sum dynamics—often called a "race to the bottom"—where participants sacrifice long-term sustainability for short-term gains, resulting in a loss for all involved.Hence why we have to keep feeding the orphan crushing machine.
drdaeman: Neither of those solves it, just tries to conserve the status quo.The issue, as I understand it, is literally a new Eternal November, just that instead of “noobs” there are “clankers” this time.Personally, I don’t give a flying fuck about gender, organs (like skin or genitalia) or absence thereof, or anything alike when someone posts something online, unless posted content is strongly related to one of those topics. Ideas matter no matter who or what produces them. Species fit into the same things-I-don’t-care-about list just fine - on the Internet nobody knows^W cares you’re a dog. As long as you behave socially appropriate.The problem with bots is that they’re noobs and unlike meatbags they don’t just do wrong and stupid things but can’t possibly learn to stop (because models are static). Solving that, I think, is the true solution. Anything else is just addressing the correlations to the symptoms.
throwaway94275: Blog admin sees who invited the bots and recursively kicks that account and any invited by it.
gpderetta: signal/noise
starkeeper: I just searched for a video game tip: "Bannerlord II where to sell clay?" and google's top result was an AI generated page FOR THIS GAME that directed me to ebay.It's dead Jim.
mrkramer: Tbh I don't care if I speak to a human or a bot as long as they are "useful"; by useful I mean if they provide me useful information but then again humans can provide unique information that bots can not. But I think identity is not relevant anymore, what's relevant is reputation. People think internet bots are bad per se but we need to build useful bots, just like there are chat bots that are useful on various platforms like Telegram, Discord or whatever other platforms people use.
rustystump: Everyone here is so far from a normie it is almost painful. Dead internet is an outcome of supply and demand.The fundamental issue is that a plurality of humans pref the direction things have gone and are moving in. Is it a good direction? By this crowd’s standards, no.To be clear, i dont like either but when i watch the speed kids swap between 5 insta accounts and 3 reddit accounts, it seems the majority are happy with it.
twoodfin: Let’s not kid ourselves: Every day, multiple “I just asked the LLM to clean up my notes” posts are voted up to the front page here, often with highly engaged, appreciative comment sections.LLM’s for all their faults are well-trained to produce what we want.
talkingtab: It is the corporate internet - the one by the corporation, for the corporation that is dead. Or at least everything in it is dead. The death blow is AI, but it was almost there anyway.The good news is that the community internet - for the community, by the community - is just starting.What is a community internet? The internet is layered protocols. UDP, ICMP, TCP, HTTP, HTTPS etc. The community internet is just a new layer of protocols. Coming soon.
chewbacha: Sure, but just for reference: https://xkcd.com/927/
obsidianbases1: > what's really happening is that big centralized systems are being overrun with bots? The internet existed and was pretty great before these large centralized systems came into being.This is a great point. Suddenly, I'm looking forward to this
keithnz: No, the old internet wasn't that great. There were so many problems. Finding things was hard, buying things was hard, integrating things was hard, compatibility was hard, everything was super fractured. It felt great at the time because you discovered all these random things and it was all novel at the time. Centralized (Or decentralized collaborative services like IRC or Usenet) really unlocked the power of the internet.
okanakyol: It might be too late, but AI companies should have developed an offline-verifiable standard before scaling. This would work on our browsers and operating systems to automatically detect and mark AI content.
RobRivera: Maybe we should get more Jannies. They only cost zero dollars and zero cents
abcde666777: For a while video was a holdout of sorts - e.g. if someone posted video content of themselves or their voice you could trust a real person was behind it.But now convincing fake video generation is easily accessible, so one more holdout stands to fall.It does seem like some kind of ID system is going to be the only way. Sucky but inevitable.I often have the following thought: technological advancement, for all its boons, inevitably leads down destructive roads in the long run. Sooner or later we open a pandora's box.
ramses0: Email2000 is the only answer: https://cr.yp.to/im2000.htmlTLDR: Mail storage is the sender's responsibility. The message isn't copied to the receiver. All the receiver needs is a brief notification that a message is available.
jasondigitized: Bring back BBS. Getting into the good ones was a process back in the day.
glimshe: BBSs have been in theory replaced, but in reality they haven't even been approached by modern social media. Small forums full of dedicated users, often local. So many great memories.
bytehowl: The technology isn't inherently evil. The actual problem is the way our societies are set up, ironically incentivizing sociopathic behaviour even among members of a single nation, nevermind when geopolitics get involved.
hinkley: I drank enough for three people in college. My lifetime average is probably still in the margin of error.
bonesss: A $10 one time not-an-asshole fee is totally reasonable.History also shows you can take a $10 fee and maintain quality on SomethingAwful for quite some time.
pndy: Didn't that fee allowed to change account names of other users or something like that?
yakattak: You could pay another $10 (or maybe $15?) to change someone else's avatar.
tracerbulletx: Parasitic zombie internet.
toddmorey: I see many, many startups that promise to be an automated marketing agent that will do this exact thing: scour sites for conversations and post links to your product.Obviously that burns down the human Internet, but it’s also a business that will have a short lifespan and bring about its own demise.I guess they don’t care about anything enduring as long as they can grab some quick cash on the way out.
girvo: > I guess they don’t care about anything enduring as long as they can grab some quick cash on the way out.As far as I can tell, that is basically all AI-related businesses. Including those non-AI ones jumping on the bandwagon to throw all their employees in the bin and expect 10x productivity somehow: if they are right and these tools do become that good, well the economy as we know it is over as white collar knowledge work disappears.But hey, we made money in those few years right!
pixl97: At least in the US very few industries actually seem to be about making a product.A good example is this, car companies don't make cars for the most part, they make loans. Financial companies first, car companies second.Consolidation, collusion, and rent-seeking behaviors by companies are going out of control too. The fact AI companies can do what they are doing has much to do with the previous brick and mortar businesses weakening any business regulations down to nothing.
D13Fd: > A good example is this, car companies don't make cars for the most part, they make loans. Financial companies first, car companies second.I get that this is true from a certain point of view. But car companies clearly compete in a very healthy way on features and quality.In fact, cars are a great example of a market where the companies clearly care about making the product, and the competition between them has driven that products to incredible heights. Cars these days are vastly better than they were in the past.
mvdtnz: > convincing fake video generation is easily accessible, so one more holdout stands to fall.Is it though? I have absolutely no doubt we'll get there but I haven't seen any evidence of this in the wild. My Youtube feed is becoming overrun with content with clearly generated scripts and often generated narration. But I haven't seen a single instance (that I'm aware of) of generated video being passed off as real.Yes I have seen hundreds of tweets and reddit posts showcasing game-changing video technologies like AI face replacement and yes they look incredible in the 45 second demo reels, but every instance I have seen of real-world usage was comically bad.
pseudalopex: Who wrote what you quoted?
pixl97: "You can do everything right and still lose".At the end of the day there is no real penalty for being a bad actor on the internet. They get unlimited retries on spamming and otherwise causing problems. In many ways this helps Google entrench itself as the search/ad company. No one else has the money or compute resources to continuously update the internet. Furthermore they have told us it's their job to shove unskippable ads in our faces. They'll gladly let the public internet die in the future if they can push out their own version of "SafeInternet by Google/now with more ads!".
amiantos: Every single one of your comments in this thread is some slippery slope stuff where you think corporations and federal government are going to work together to kill off the (public?) internet. It's okay that you feel that way, even if it's just a big ol' fallacy, but you don't need to repeat it in six different places. You made your point, you think the internet is doomed no matter what happens, great, let's move on.
ryandrake: I would probably not pay $10 to post on HN, but many spammers who expect some kind of tangible return would pay that, so the fee just makes the problem worse.
munk-a: The spammers wouldn't pay it once though - the idea is that it's a good way to scale moderation. Each time an admin needs to ban a user there is a 10$ subsidy supporting that action - and if the bots come back then they get to pay 10$ to be banned again.Assuming the money isn't wasted and is actually used to fund moderation 10$ is probably comfortably above the cost to detect and ban most malicious users.
jnettome: That’s basically what Valve does on cheaters with premier accounts on cs:go/cs2. And the revenue still growing up.
AngryData: Yeah, I love HN, but I wouldn't pay and I know many if not the majority of other people wouldn't. It would increase quality for awhile for sure, but what happens a year or two down the road? It would kill the user count and reduce comments and become less valuable over time.
lovich: Ironic that there’s a dead bot comment at the bottom of this article trying to pose as a human
protocolture: I reckon we are going find an inverse metcalfes at some stage where the value of a network is proportional to the square of its connected users, minus the square of the number of connected bots. Heck I would be surprised if meta hadnt figured this out, or wasnt on the way to figuring it out.
