The 'Dead Internet Theory'
theconversation.com
theconversation.com
> A parent asked a question in a private Facebook group in April 2024: Does anyone with a child who is both gifted and disabled have any experience with New York City public schools? The parent received a seemingly helpful answer that laid out some characteristics of a specific school, beginning with the context that “I have a child who is also 2e,” meaning twice exceptional.
> On a Facebook group for swapping unwanted items near Boston, a user looking for specific items received an offer of a “gently used” Canon camera and an “almost-new portable air conditioning unit that I never ended up using.”
> Both of these responses were lies. That child does not exist and neither do the camera or air conditioner. The answers came from an artificial intelligence chatbot.
> According to a Meta help page, Meta AI will respond to a post in a group if someone explicitly tags it or if someone “asks a question in a post and no one responds within an hour.”
"Dead Internet" is not a theory. It's an active goal for some companies.
Firstly their search /discover was pretty terrible. I am sure that local groups exist, but searching for keywords mostly turned up communities with commercial angles.
Secondly even tho I only joined groups related to a very specific European location, every 3rd post in my feed is now bizzarely a scantily clad teen from Africa. I can dismiss these posts individually, but there is no way to tell Facebook's algorithim that it placed me in the wrong bucket, they provide no way to opt out.
If they really must insert random algorithmic content, Facebook already has all the clues they need to make it relevant, based on my searches and groups that I have joined. Instead their AI tried to play 4D chess with my preferences and outsmarted itself, serving me a horrendous swamp of teenbait slop.
Facebook is dead.
I bet that's just one of the ways they track behavior on their platform. So if you slow down to check out the ladies, I bet more show up, even if you don't click.
My friends, now we're all a bit older, don't post as frequently as they once did. I'm also a member of a couple of hyper-local groups which don't necessarily get posts every day, and I have no interest in anything else. I'm fine with this, if nobody's posted for a while then I'll see that and move on.
Facebook is not fine with this. Apparently I must be interested in archaeology, or sexist jokes, or various displays of flesh, or cars, or sports, or following someone that's making misogynist 'jokes' or ancient aliens conspiracies or a not-very-funny comic or more flesh or... I just must be interested in something, surely? It's just a matter of time before they wave the right thing under my nose and I bite and increase my engagement?
Is that the thinking? Because now 90%+ of my feed is unwanted shite and it's hard to find the few things I do go there for. And this is driving people like me and my friends away further, leading to a downward spiral of less content for those that remain, more 'suggestions' etc.
It's also baffling - multiple bullshit 'archaeology' groups will post the same picture (often of something not related to archaeology at all) on the same day, so they're clearly run by the same entity, spinning up millions of crappy 'groups' with low-quality content. Why do they exist? Who is benefitting from this?
AI can only speed up this enshittification, but the platform is already drowning in it.
Not to mention embarrassing that a viewer of my feed would think that I am some sort of knuckle-dragging horndog.
Perhaps the internet is bifurcating into lowbrow and highbrow, and facebook has chosen a side.
Spurious recommendation algos currently have enough engagement, bots et al, that they'll keep rattling away... for now.
Thus I have more than one account to use these as topical pre-filters. And when you loose your cookies, YouTube wants a phone number to "verify" that it's "me". Luckily, as a European living near enough to French borders, I can purchase temporary SIM cards in France and "verify" things during holiday trips and don't need to use my main phone account.
when I think about "the dead internet theory", this is what I imagine. all of this happening over top of our heads as we chat on a forum, out of the way.
the essay mentioned a comic series called BLAME, in which machines programmed to build cities, buildings, infrastructure, got derailed after humanity perished, endlessly creating walkways and stairs and buildings to nowhere, for nobody, forever unempeded, as they were programmed. Much of it makes no sense, like the Winchester House's stairs into solid wall etc, but it doesn't matter because it's all just being done according to their railroaded programming.
wasn't like 70% of the activity on Twitter during the last superbowl all just bot-related traffic?
This is BLAME! (with the bang) by Tsutomu Nihei. There's a one tankobon prequel: NOISE.
Definitely a recommended read. The art is magnificent, the mastery of perspective incredible, managing to give you a sense of the unconceivable scale of the megastructure and by way of consequence, the mind-boggling energy involved in some events. It made me feel vertigo in spite of the small tankobon pages. The storytelling is top notch: few words, but a very complex story and setting connected through a ton of small details.
SNIKT! is a crossover with Wolferine (or rather, Logan), an unexpectedly excellent rendition of the character.
I feel that in search of a few more dollars from engagement, they are slowly killing their main product. People will eventually learn that this is not a useful thing, and avoid it, although it happens slow. But once this enters the mainstream consciousness, it's over.
Maybe other social networks will appear, that will somehow guarantee you are interacting with humans, or maybe something else will take their place. But anyone who thinks they can extract value from humans longterm by drowning them in machine generated nonsense is wrong. People will adapt.
I hate that I find myself thinking, it really was better back then
It won't necessarily be "over top of our heads as we chat on a forum, out of the way." You missed people replying to bots, thinking they're replying to a person: https://www.404media.co/facebooks-ai-spam-isnt-the-dead-inte....
We meat sacks don't stand a chance. We need to evolve.
The theory predates the current era of AI generated content:
> The dead Internet theory's exact origin is difficult to pinpoint, but it most likely emerged from 4chan or Wizardchan as a theoretical concept in the late 2010s or early 2020s.
https://en.wikipedia.org/wiki/Dead_Internet_theory
And just my personal 2cents; I believe this was the result of predatory SEO.
but then again, go outside and holla that fake shit. everyones trained and tutored by of through and for emotional bondage. even, or rather, thanks to the dreaming mind the art itself is disconnected and alienated from the manipulated and corrupt artist. it's not the kids playing music anymore, it's their shadows or their guts if you so will. once the music stops they are all back to being regular bots again. we all have only so much energy for putting up a good show, there's none left for playing our selves, for being honest, which is why it's easier to outsource as much as possible to digital and superego algorithms; just choose a flavor or more and the sim remains stable and fun and the kids will have stuff to digest and remix forever
i shouldn't do this again, though. apologies, if I annoyed someone or got them into a bad mood.
Not content to fuck up the real world, we've done the same in the digital.
Social sites, like systems, consist of inputs and feedback loops. How the feedback inputs are filtered is what we call moderation. Sometimes good, sometimes not.
Google tamed the web and tuned itself for its own benefit. SEO engineers noticed and motivated by culture's dear friend greed, they were some of the first to automate - faking traffic and content to shape google, the shaper. Fixing the fixer.
And now we add LLMs to the mix.
In control systems, negative feedback shapes the outputs in useful ways while positive feedback amplifies. We've set the dial to 11+.
Possibly it was extrapolating from the ruins of MySpace and similar.
> You, a 48-year-old man who dies in a car crash, meet God, the narrator, who says that you have been reincarnated many times before, and that you are next to be reincarnated as a Chinese peasant girl in 540 AD. God then explains that you are, in fact, constantly reincarnated across time, and that all human beings who have ever lived and will ever live are incarnations of you. [1]
[1] https://en.wikipedia.org/wiki/The_Egg_(Weir_short_story)
Kurzgesagt adaptation of the story (31M views):
There is a physical theory that claims that all electrons are the same electron, just travelling endless through time from the beginning to the end of the universe and back (becoming positrons on the reversed timeline).
That explains why all electrons have the same exact charge.
https://en.wikipedia.org/wiki/One-electron_universe
Not sure it's a serious theory capable of proving something, but there you go.
I think one of the responses to the growing phenomenon, especially as LLMs make fake content even easier and more convincing, is that we will see increased uptake on platforms like Discord where you have a curated and more intimate circles. Or maybe those like Musk will get his way and platforms like X will require paid verification!
Bots don't kill internets, people do.
The equation is roughly result = some human's intent * available technology
If some human (or organisation)'s intent is to deceive the elderly at scale and their reach is that much greater now due to new technology, what part of the conversation stops due to there being a human in the loop?
I have written many pieces of software that did something I didn't intend. I have the feeling it's the same for everyone who's written anything more complex than hello world.
Yes, there's a group of people behind those, but it's hard to ignore the difference between a single person controlling a couple of bots to post harmless updates or whatever, and Governments trying to sway the agenda of other countries their way.
I would like for verification to be mandatory on platforms meant for "real people" to interact, but not "paid" on a proprietary platform, but using new standards for internationally valid identities. I am betting on w3C's Verifiable Credentials for that: https://w3c.github.io/vc-overview/
It's still a bit unclear who would have the authority to issue "real person" credentials, but at least in the EU that's currently being worked on: https://commission.europa.eu/strategy-and-policy/priorities-...
In the USA, they seem to be trying to use Driver's Licenses for this sort of thing, see the ISO/IEC specs at https://www.iso.org/standard/63798.html
I think the fundamental disagreement is that, for some, the solution to poor governance is less governance, and for others the solution is better governance.
I am in the latter camp because I think it's apparent than in the absence of better governance, poor governance is replaced by even poorer governance.
Given that the internet is filled with explicitly predatory algorithms and deceptive ‘dark patterns’ both trying to attract attention, thereby extract personal data or cash from humans, letting them know you are a human is merely to invite attack.
There may be advanced civilisations hidden online, but we are not them.
Discuss
Like, resources and energy are fairly abundant in the universe. There's really nothing of material value that we have here in Sol that other species wouldn't be able to get closer to their own systems. So, scarcity isn't a big factor, I think.
Instead, the more scarce thing in the universe is life, with intelligent life presumably more scarce than that. We, ourselves, are the valuable things.
But Dark Forest theory isn't totally an economics argument, it's more of a security argument. With the vast distances and times, other species could become dangerous quickly, so kill them before they are. Again, the vastness and long time scales are just too loose of a thread. Even in the books, humans manage to seep out and hide among the dust and planets, doing things in the dark or in the blinding radiation of stars. So, I don't buy the security argument either. There's just too much space and material out there to really ever be sure that you could ever really sterilize the other species and not have them come back to haunt you. Especially by the time you actually figured out they existed, they'd be everywhere. (The books invent this dimension collapsing weapon idea as a way around this).
So, I think we should have the counterpoint to the Dark Forest theory of hunters with flash lights in a dark forest.
We should have "Used Car Salesman / Lemon Law" theory: Since you can't really beat them 100%, intelligences realize that you should join them; but on favorable terms, of course. So, beware of alien civilizations that appear in your system with lots of flags and balloons and bad ties and big smiles and want you to sign on a dotted line, fast. You can't really know if they are trying to sell you a lemon before the other civilizations show up and negotiations can really start.
The bot content is very successful at displacing the human content, since it's engineered to maximize visibility in a way most humans don't bother with.
Either way, the end-state in Accelerando is looking more and more prescient.
(Fby vf znqr ubfgvyr gb uhzna yvsr ol cncrepyvc znkvzvfref gung unir qrnq-raqrq gurzfryirf vagb n Erq Dhrra'f Enpr-fglyr fgehttyr sbe pbzchgngvbany erfbheprf. Uhznaf hc fgvpxf naq zbir arkg qbbe).
It is a sort of tragedy of the digital commons, where the commons in this instance is the availability of extremely usable digital channel for people to communicate online in a decentralized way.
The early pioneers were really excited about the empowering potential of all this, but alas they did not work out the full dynamics of human nature (seeking short term profit, influence and control) especially when it is combined with the amplifying impact of automation and algorithms.
In short, the channel is being ruined by noise, the signal is swamped, the digital commons is being overgrazed and dying.
Nobody feels responsible for this. The stewardship of this most unique resource that was gifted to us by the digital age is laughable. A cacophony of commercial and state interests that have not produced a single positive development in the last thirty years since the invention of the web.
Has anyone worked out this … in terms of governance or anyone organized grouping of humans?
"The real is produced from miniaturized cells, matrices, and memory banks, models of control - and it can be reproduced an indefinite number of times from these. It no longer needs to be rational, because it no longer measures itself against either an ideal or negative instance. It is no longer anything but operational. In fact, it is no longer really the real, because no imaginary envelops it anymore. It is a hyperreal, produced from a radiating synthesis of combinatory models in a hyperspace without atmosphere. By crossing into a space whose curvature is no longer that of the real, nor that of truth, the era of simulation is inaugurated by a liquidation of all referentials - worse: with their artificial resurrection in the systems of signs, a material more malleable than meaning, in that it lends itself to all systems of equivalences, to all binary oppositions, to all combinatory algebra. It is no longer a question of imitation, nor duplication, nor even parody. It is a question of substituting the signs of the real for the real, that is to say of an operation of deterring every real process via its operational double, a programmatic, metastable, perfectly descriptive machine that offers all the signs of the real and short- circuits all its vicissitudes. Never again will the real have the chance to produce itself - such is the vital function of the model in a system of death, or rather of anticipated resurrection, that no longer even gives the event of death a chance. A hyperreal henceforth sheltered from the imaginary, and from any distinction between the real and the imaginary, leaving room only for the orbital recurrence of models and for the simulated generation of differences."
At this moment I feel it's impossible to believe in any comment on the internet unless you can vouch for the person.
... I'm sorry, but I had to. It's scary, especially considering how many people cope with loneliness. Do you imagine that someone falls in love with AI or keeps thinking it's their friend while unaware it's just an AI run account? I do and it's depressing to think about.
Was sending direct mail, i.e., junk mail, the highest and best use of the postal service. It might have been the most profitable but does that mean it was the best.
The internet is a computer network with an enormous amount of junk travelling across it. In theory, that could change.
That so much of internet is bots, specifically different bots, from multiple players, and some on auto-pilot aka AI.
So now, these AI bots will pick up on themes from other bots and the whole group of bots will go down some rabbit whole. One AI picks up on some idea, makes a meme, the others kind of follow riff on it, then that is the seed for the next AI to take off. Like their own society, like a bunch of teenagers.
Today I recomment using tools like Mastodon and RSS, they are a breath of fresh air. You may encounter AI generated or shallow content, but you can prevent it by unfollowing or blocking. Curate your news feed. Your mental health is worth the effort.
As the major social networks continue to enshittify themselves, interest in indieweb and the fediverse is growing. I've been building an RSS directory based on parsing blogrolls to form a recommendation network, as one example. It's still an early release.
Then we get into staged content. Usually it's harmless; acting out a made-up scenario for a comedy bit. But sometimes it gets more insidious, particularly the ones where people pretend to do good deeds but are just inventing a sob story about strangers who didn't ask for help and didn't actually need it.
And people eat this up. You'll find some skeptical comments but it's absolutely drowned out by the volume of people who think it's real and share. And it's unlikely that all tens of thousands of those people are bots. I'm lead to believe that this kind of low-quality, questionably-real, feel-good content is just legitimately popular. Not hugely popular but enough to make me groan that people are engaging with it at all.
To tie things back to AI, the latest groan-worthy trend is people sharing pictures of things made out of bread or fruit that are obviously made by AI but the comments are "incredible" "wow how did you make this?" which makes me fear for when somebody uses AI image generation for misinformation instead of just stupid stuff.
It is true whether one believes it or not and this needs to be undone which shall never happen.
Just damn too sad!
As well as out of control positive feedback amplifying the fake (expedient, efficient, profitable) there's negative feedback suppressing the authentic (complex, messy, emotional, unique).
We are doing this to ourselves. People make and program bots, and so it's people that want to kill the Internet, but also what it represented about human connectivity.
There are plenty of humans actively attacking what is human. I've heard them called misanthropists, nihilists, cosmists, trans-humanists, tech fanbois, singularity wingnuts, species traitors... all kinds of names and explanations for the same phenomenon - a masochistic, self-loathing or disgust for humanity presented as a "love" of technology.
Maybe it's climate change, overpopulation, first-world guilt, who knows? Perhaps the idea of AI genesis linked to human sacrifice gives some a sense of fascistic "new broom" purpose?
Whatever the explanation, we built the internet, and we (humans) are actively destroying it.
lol