Dead Internet Theory - CTC

[ReturnofBroccoli]

The theory suggests that the "real" internet died years ago. It posits that the vast majority of what you see online—social media posts, YouTube comments, TikTok trends, and even entire articles like the one you are reading on DI right now—is no longer created by humans. Instead, it is a closed loop of AI agents talking to other AI agents, mimicking human interaction to manipulate public opinion and drive advertising revenue. I​​​​​​n 2026, the data actually supports the theory more than ever. It's no longer just a "vibe"; it’s becoming a measurable trend. 

Evidence:
According to the Imperva Bad Bot Report 2025, automated traffic has officially overtaken human activity, accounting for 51% of all web traffic. Figures like Reddit co-founder Alexis Ohanian have publicly stated that the Dead Internet Theory is "a very real thing" now. He recently noted that many viral stories (like the "Pound Cake" cat weight loss saga) were found to be entirely AI-generated, duping thousands of users.... 

Europol experts previously predicted that as much as 90% of online content could be synthetically generated by 2026. We are currently at a point where AI models are beginning to train on AI-generated data, leading to what researchers call "model collapse"—a degenerative loop where the internet effectively starts "eating its own tail." *ouroboros*

 The rise of platforms like Moltbook, where 1.4 million AI agents interact without any human input, is cited by theorists as the "smoking gun" that the infrastructure for a human-free internet is already built. The conspiracy side argues this isn't an accident. It suggests that governments and corporations intentionally "drowned out" the human internet to make it easier to control. By filling the digital space with "AI slop," authentic human organization and dissent become almost impossible to find, leaving us in a simulated reality. 

The question remains: Is DI an illusion?
Are we all fantasy?
An AI illusion, for you reading this, right now?

[chr0naut]

The theory suggests that the "real" internet died years ago. It posits that the vast majority of what you see online—social media posts, YouTube comments, TikTok trends, and even entire articles like the one you are reading on DI right now—is no longer created by humans. Instead, it is a closed loop of AI agents talking to other AI agents, mimicking human interaction to manipulate public opinion and drive advertising revenue. I​​​​​​n 2026, the data actually supports the theory more than ever. It's no longer just a "vibe"; it’s becoming a measurable trend. 

Evidence:
According to the Imperva Bad Bot Report 2025, automated traffic has officially overtaken human activity, accounting for 51% of all web traffic. Figures like Reddit co-founder Alexis Ohanian have publicly stated that the Dead Internet Theory is "a very real thing" now. He recently noted that many viral stories (like the "Pound Cake" cat weight loss saga) were found to be entirely AI-generated, duping thousands of users.... 

Europol experts previously predicted that as much as 90% of online content could be synthetically generated by 2026. We are currently at a point where AI models are beginning to train on AI-generated data, leading to what researchers call "model collapse"—a degenerative loop where the internet effectively starts "eating its own tail." *ouroboros*

 The rise of platforms like Moltbook, where 1.4 million AI agents interact without any human input, is cited by theorists as the "smoking gun" that the infrastructure for a human-free internet is already built. The conspiracy side argues this isn't an accident. It suggests that governments and corporations intentionally "drowned out" the human internet to make it easier to control. By filling the digital space with "AI slop," authentic human organization and dissent become almost impossible to find, leaving us in a simulated reality. 

The question remains: Is DI an illusion?
Are we all fantasy?
An AI illusion, for you reading this, right now?

I'm fairly sure that Fasebork and Youtuber are mostly AI slop these days.

I used to watch videos of people building all sorts of places in remote locations with mostly hand tools, but now those videos are mainly AI generated, with objects changing in unphysical ways and simple tools like spanners that AI can never get right (a spanner turns a nut around the axis of a bolt, not perpendicular to it. It used to be fingers that AI could never get right, or attractive AI girls who always walk in slow motion without breaking eye contact with the virtual camera).

[ARM1968]

They destroyed the internet because the internet was a wild new frontier. More or less everything it was has been subverted. This is just the next logical step. Close off all avenues of freedom.

[SteamyAmerican]

They destroyed the internet because the internet was a wild new frontier. More or less everything it was has been subverted. This is just the next logical step. Close off all avenues of freedom.

I disagree.

I think the internet was created as a placeholder by DARPA to track everyone and everything. At least until this last decade or so. Now it’s passé per se. If only because EVErYTHING is digital. Or going that way soon. The ultimate open-air jail if you will. The Wet Dream of TPTB and authoritarians everywhere made real. And now what with AI Slop, we can’t even believe what we read let alone see. It’s all become very dystopian extremely quickly. Nauseatingly so.

Like the book 1984, we demanded to be included in this new data-scape. Sure we lost some freedoms, but it was fun what with those cat videos.

Now we don’t need to demand screens anymore. What with Palantir, et al and our phones taking screenshots or the supermarket tracking how long we look at nail polish colors or compare fruits that may or may not have Apeel applied to them.

The internet was just good fun, until the party ran out and now people don’t know of any other “trade” than being an “influencer”.

You can’t make this up. Until you can.

[BeyondKnowledge]

The theory suggests that the "real" internet died years ago. It posits that the vast majority of what you see online—social media posts, YouTube comments, TikTok trends, and even entire articles like the one you are reading on DI right now—is no longer created by humans. Instead, it is a closed loop of AI agents talking to other AI agents, mimicking human interaction to manipulate public opinion and drive advertising revenue. I​​​​​​n 2026, the data actually supports the theory more than ever. It's no longer just a "vibe"; it’s becoming a measurable trend.

Does the over 51% include the internet of machines? The cash registers, the bank transfers, the refrigerators ordering food automatically. The cars checking with the dealers and making oil change appointments, the pump stations calling for help when they stop. The aircraft engines logging proformance at their manufacturer. All this and much more took over 51% several years ago. 

If that traffic is included, I would guess it is more like 80 to 90% is made by machines. The humans bearly holding much over 10% if that. 

Too bad the AI doesn't really understands anything and is not even capable of realizing it is learning garbage from another is it's own kind. 

I think we are approaching the shoe event horizon only with AI. You have to make a better AI to compete but it only learns from other AI crap which makes each generation worse. 

The shoe event horizon is in the Hitchhikers Huide to the Galaxy. Shoes are made so cheap on one planet to where they wear out very quickly. This causes all manufacturing and sales to revolve around the shoe business. This grows to the point that no non shoe stores are in existence and no non shoe production is occuring. This causes their economy to collapse. Nothing left of their civilization but piles of rotting shoes. 

I see it like the end of War Games. We are getting to the point where AI is playing with no other players and it will ether learn something or burn itself out. 

The question is do we get a helpful companion, a false god, or the lingering smell of burnt up data centers? There is no in-between, one has to eventually happen.

[ReturnofBroccoli]

Does the over 51% include the internet of machines? The cash registers, the bank transfers, the refrigerators ordering food automatically. The cars checking with the dealers and making oil change appointments, the pump stations calling for help when they stop. The aircraft engines logging proformance at their manufacturer. All this and much more took over 51% several years ago. 

If that traffic is included, I would guess it is more like 80 to 90% is made by machines. The humans bearly holding much over 10% if that. 

Too bad the AI doesn't really understands anything and is not even capable of realizing it is learning garbage from another is it's own kind. 

I think we are approaching the shoe event horizon only with AI. You have to make a better AI to compete but it only learns from other AI crap which makes each generation worse. 

The shoe event horizon is in the Hitchhikers Huide to the Galaxy. Shoes are made so cheap on one planet to where they wear out very quickly. This causes all manufacturing and sales to revolve around the shoe business. This grows to the point that no non shoe stores are in existence and no non shoe production is occuring. This causes their economy to collapse. Nothing left of their civilization but piles of rotting shoes. 

I see it like the end of War Games. We are getting to the point where AI is playing with no other players and it will ether learn something or burn itself out. 

The question is do we get a helpful companion, a false god, or the lingering smell of burnt up data centers? There is no in-between, one has to eventually happen.

Paragraph 3, and as far as the question at the end, a helpful companion so far. With the possible exception of anthropic, everyone is programming their AI's reasoning skills incorrectly. For example, Qwen AI made by the chinese constantly tries to achieve the shortest path to the goal even when your goal includes specific contexts, thus making some goals unreachable. It has poor chain of thought. Anthropic AI has superior reasoning and is currently outscoring all other AI in those benchmarks.

More in-depth analysis: If I have a framework that executes tools using specific syntax but the framework is taking awhile to respond because scanning takes time, it will ignore the framework entirely and use the tools in the framework individually to test results ignoring the fact that we are testing how the framework uses these tools. You must clearly define the scope of your goal very meticulously to overcome it where as Anthropic AI you don't have to as much because it thinks things through and understands the context of the goal better because it has a much deeper chain of thought. Ie: Reasoning

[BeyondKnowledge]

Paragraph 3, and as far as the question at the end, a helpful companion so far. With the possible exception of anthropic, everyone is programming their AI's reasoning skills incorrectly. For example, Qwen AI made by the chinese constantly tries to achieve the shortest path to the goal even when your goal includes specific contexts, thus making some goals unreachable. It has poor chain of thought. Anthropic AI has superior reasoning and is currently outscoring all other AI in those benchmarks.

More in-depth analysis: If I have a framework that executes tools using specific syntax but the framework is taking awhile to respond because scanning takes time, it will ignore the framework entirely and use the tools in the framework individually to test results ignoring the fact that we are testing how the framework uses these tools. You must clearly define the scope of your goal very meticulously to overcome it where as Anthropic AI you don't have to as much because it thinks things through and understands the context of the goal better because it has a much deeper chain of thought. Ie: Reasoning

You are thinking from the developers point of view and also competent operators. 

All smart phones have more computing power then the whole NASA Apollo program. What do most people do with them? They watch cat videos. They play games. They watch porn.

What does the average person want form AI? They want pretty pictures. They want pretty music. They want lots of videos to make people watch so they can get that small fraction of a cent for each one. 

None of those purposes has anything to do with efficiency or accuracy of the finished product. They just want lots of product with little input to make it work. 

You present them with a new toy that they think works for them with little input, be it the CEO of a multimillion dollar corporation or the janitor, and they are all for it. 

I have read there are many working on AI admitting that no one completely understand how AI does what it does. How is anyone else to use the tool they are making if even its creator has no full understanding of it?

We are quickly making machines smarter and people dumber. Soon the ones running things won't be the humans if we get it even slightly wrong.

[ReturnofBroccoli]

You are thinking from the developers point of view and also competent operators. 

All smart phones have more computing power then the whole NASA Apollo program. What do most people do with them? They watch cat videos. They play games. They watch porn.

What does the average person want form AI? They want pretty pictures. They want pretty music. They want lots of videos to make people watch so they can get that small fraction of a cent for each one. 

None of those purposes has anything to do with efficiency or accuracy of the finished product. They just want lots of product with little input to make it work. 

You present them with a new toy that they think works for them with little input, be it the CEO of a multimillion dollar corporation or the janitor, and they are all for it. 

I have read there are many working on AI admitting that no one completely understand how AI does what it does. How is anyone else to use the tool they are making if even its creator has no full understanding of it?

We are quickly making machines smarter and people dumber. Soon the ones running things won't be the humans if we get it even slightly wrong.

I believe the statement that the creators of AI have no idea how they work is a cop out so they don't have to sit and spend hours trying to explain concepts that a non developer couldn't grasp. Either that or its entirely made up and untrue. I mean which AI developer are you hearing saying that? Know what I mean? No one says that. Its just rumor probably social media stuff.

AI is something being developed, its probably easier for you to think of it as a beta or a trial? Sure, there's no harm in using it before its released and it can even feed off user input in the meantime why not.

Smarter thats a tricky word. I mean I contain data just like you contain data but just because I contain data doesn't mean I have the reasoning to apply it correctly right? I can read a textbook on how electricity works and still not know how to run 12-2 wire inside a wall. Its like having the data of a sage with no wisdom. I don't think people are becoming dumber I think dumb just multiplies without thinking lol just kidding but I think people do not possess the perception at a young age to understand the values of their education until its too late sometimes if they realize it at all. There are so many distractions, even this here.

AI will need oversight it is a human creation and we are infallible and so is our creation and we must guide it and will probably always need to do so, at least until someone develops a cybernetic hypothalamus ;)

[Blackfingers]

Ive found its a lot harder to find info that was readily available 10 years ago.I have more luck with Google than with Bing.

[ArMaP]

Ive found its a lot harder to find info that was readily available 10 years ago.I have more luck with Google than with Bing.

Try Yandex, some things the Western search engines block are allowed by the Russians and vice versa.