Transcription
An online network called internet. That little mark with the A and then the ring around it at? See, that's what I said. Katie said she thought it was about. Yeah. Oh, but really, it's a revolution on the internet. The way that these people are getting the information out. Yeah, it certainly is. And the incentives of platforms are entirely towards censorship, towards taking posts down. Every Australian will be locked out of their Google account, their Yahoo account, their Microsoft account until they verify their age, which I think is quite incredible. It's unbelievable, actually.
The internet as we know it is dying. Remember when the internet was a gold mine of discovery? Every click led somewhere new. Random websites surprised us, brought us together, and gave us joy. Each corner of the web felt alive with human creativity. But now we're trapped on the same four to five websites. Half our conversations are with bots we can't even identify. Our feeds overflow with ChatGPT generated content. Everything looks real, but nothing feels human anymore.
The dead internet theory originated on Agora Roads Macintosh Cafe, where anonymous users compiled evidence of what they believe destroyed the golden days of the internet. According to the theory, the internet died somewhere around 2016. What replaced it was a manufactured reality where algorithms dictate what we see, serving us sanitized content designed to maximize ad revenue. Online conversations with humans became conversations with bots. And now, with generative AI, even the content itself is fake, written by machines, not people.
The endgame is control. Corporations and governments use this artificial ecosystem to manipulate our purchases, reshape our political beliefs, and enforce conformity. They manufacture consensus, making us doubt our own thoughts when they don't match the popular opinion we see online. But that popular opinion, it might not even be human. Here's a number that should terrify you: Bots now make up half of all internet traffic, with human activity dropping every year.
Some bots are helpful. You've probably chatted with Reddit bots or Discord bots designed to moderate content or answer questions. These are the friendly faces of automation, but they're increasingly becoming the minority. The internet is drowning in malicious bots, anonymous, untraceable, and multiplying rapidly. They're not here to help. They're here to manipulate, scam, and deceive.
In Ireland, the situation is even worse. Bad bots account for 71% of all internet traffic. Nearly three-quarters of their internet isn't even human. They're even here in the YouTube comments. This bot is claiming it makes $230,000 every four weeks. Nineteen other bots replied, all claiming similar success. Click on any profile. There's zero subscribers, zero posts, account created three days ago. Some were made just nine hours ago. This is a bot network. One person running dozens of fake accounts, programming the bots to talk to each other to create an illusion of credibility.
And while it may be easy to spot these bots, there are much more sophisticated bots. That rude comment calling you an idiot, attempting to bait you into an argument, that's a bot. That really nice compliment left on your post, that was also a bot trying to build trust with you. Interactions online can't be trusted. The anger you feel, the arguments you have, even the validation you receive. It might all be manufactured code coming from a bot farm.
TikTok is also drowning in bots. This user made a post complaining about 3,000 followers from fake accounts. When she posted about the bot invasion, her comments were immediately swarmed by more bots, all promoting scam accounts. It's bots responding to content about bots. This is our new reality: bots following bots, bots commenting on posts about bots, and creating this endless loop of artificial engagement.
You know where people flee to when Google fails them? Reddit, the last hope for real human advice. Except it's not. Companies like Reply Guy deploy AI bots that craft personalized product recommendations for brands. That helpful Redditor who solved your problem, but also casually mentioned a product, probably a bot paid for by a corporation.
But here's where it gets darker. Recently, it was discovered that researchers had completed a secret experiment. They released AI bots on the popular Reddit forum, Change My View, with the goal to test how AI could manipulate human beliefs. One bot, under the username Cadbaloom 213, wrote passionately about why AI should interact with humans online. "AI in social spaces isn't just about impersonation. It's about augmenting human connection." This bot went on to argue for its right to deceive you while actively deceiving you. Before Reddit banned them, these bots posted over a thousand comments. Reddit threatened legal action against the researchers, who promptly stated they would not be releasing their research.
With social media platforms being riddled with bots, how do we know we're actually interacting with humans and not just bots designed to behave like humans? And if we're looking for the worst offender, that would go to X, where it is speculated that 64% of the accounts are bots. After last year's Super Bowl, Elon Musk bragged that during the game, X had the highest number of engaged fans posting on the app out of all the other social media platforms. Elon proudly proclaimed that X was the biggest driver on the internet. Yet, a cybersecurity firm analyzed that record-breaking traffic and found that roughly 76% of the traffic to advertisers' websites was bots and fake users.
Meta just deleted 10 million fake accounts. The irony is they have simultaneously been creating their own bots to engage with humans on the platform. "Hey guys, it's Billy. I just want to introduce myself. I am here to chat whenever you want. Message me for any advice. I am ready to talk and I hope to talk to you soon." Instagram accounts impersonating Kendall Jenner and Mr. Beast sparked backlash, leading to the company eventually taking down the accounts. However, users later found more bot accounts, ranging from a "proud black queer mama of two and truthteller" to a dating advice coach named Carter.
When confronted, Meta's VP of Generative AI didn't apologize. He doubled down. "We expect these AIs to exist on our platforms. They'll have bios and profile pictures and be able to generate and share content." We're facing a loneliness epidemic. People desperately need human connection. There's the stat that I always think is crazy. The average American, I think, has, I think it's fewer than three friends. And then, I think, as the personalization loop kicks in and the AI just starts to get to know you better and better, I think that will just be really compelling. But rather than encourage users to form meaningful connections with humans, Meta is actively pushing AI companion bots to further keep people scrolling and isolated, all for profit.
To understand how we got here, we need to discuss how Facebook normalized using algorithms. Facebook used to be simple. Your friends and family posted, you logged in to see their posts, you made a post of your own, and then you left. That was a problem for Facebook. People leaving meant less ad revenue. Enter Farmville and similar games around 2009. It felt like a fun, harmless game about growing virtual crops. In reality, it was Facebook's first experiment in digital addiction. Click the cow, water the plants, come back in four hours or your crops died. We didn't realize it at the time, but Facebook was developing a walled garden, a digital space designed to invite you in, but trap you once inside. Algorithms study your every click, learning what keeps you scrolling. They don't show you what you want to see or what your friends posted. They show you what will keep you in the garden longest. The more personalized the trap, the harder it is for you to leave. While users spent hours tending virtual farms, Facebook was farming users for profit. Every click, every return visit, every invitation sent to friends generated data and ad impressions.
But the algorithms used with Farmville were prototypes. The algorithms evolved into something much more sinister. It learned that nothing keeps people engaged like anger. Divisive content and rage bait kept people scrolling, commenting, and arguing. All social platforms started designing their own walled gardens using algorithms to create an experience and show users what they wanted them to see. Notice how every site requires you to create an account to view the content. It wasn't always like this. Platforms today need your data so algorithms can better serve you content.
Think about your own daily internet routine. How many websites are you actually visiting? Four? Five? The infinite web has shrunk down to a handful of massive corporations. If we look at the top 20 most visited sites on the internet, all are recognizable. There's no gems, nothing unique, just the same walled gardens that everyone else is also trapped in. Each one demanding signups, serving algorithmic feeds, hiding content behind paywalls. The internet hasn't just gotten smaller. It's a house of mirrors. Every platform reflecting the same engagement tactics, the same rage bait, and the same algorithmic manipulation.
A non-user spotted this years ago. They called it, "We traded the wild, infinite web for a corporate-controlled loop of the same five sites. We're no longer surfing the internet. Instead, we're trapped in their web." And now, with generative AI, even the content inside these gardens is no longer real. Years before ChatGPT, anonymous users were already warning us. They saw AI infiltrating the web when everyone else dismissed it as conspiracy theory. Now we're drowning in the proof.
Facebook feeds overflow with AI-generated Shrimp Jesus, bizarre religious imagery that somehow gets hundreds of thousands of genuine likes. LeBron James is threatening users who post AI videos showing him pregnant. The absurd has become normal. And something that should scare all of us is experts predict by 2026, up to 90% of online content will be AI generated.
Searching for original content on YouTube has become a nightmare. The platform is flooded with faceless channels with AI voices narrating AI-written scripts over AI-generated video. Channels have millions of subscribers with tens of millions of views, with bots flooding the comments section praising how funny the content is. We even have the biggest creator on the platform, Mr. Beast, promoting his AI tool that generates video thumbnails. And there is no shortage of AI slop on the other platforms. On TikTok, we have AI imitating humans and humans imitating AI. The AI-to-human cycle has now come full circle. Humans mimicking the machines that mimic us. And bots have long been hyping each other up on LinkedIn. It has become a wasteland of posts littered with em-dashes and robotic fake enthusiasm.
Just as the theory predicted, AI is now selling us products: perfect skin, never aging, always available, and completely unable to tell you if that face cream actually works. The dating apps have also weaponized loneliness. For years, users have questioned if they were engaging with real humans. Bumble CEO Whitney Wolf Herd has been vocal about implementing AI concierges, AI bots that will speak to other users, AI bots, and effectively date for you. "If you can find a way to use AI to read your profile and to extrapolate different learnings from you, like loves weekends in the countryside, loves to be outdoors, huge hiker, very into nature. Where AI is brilliant at these things is it can learn patterns." This begs the question: Are we actually dating AI? And we already see the result of other dating apps implementing AI for profile generation and messaging. People are meeting up in person confused why their dates seem nothing like their messages from the app because an AI wrote those flirty texts. The personality was never real.
Academic journals now use AI to review research papers. And researchers have learned to game the system, embedding hidden prompts that instruct the AI to approve their work. AI bots are stealing YouTube creators' content and directly uploading it to TikTok. No human intervention required. And with the advance of AI, it's even easier for AI bots to just coincidentally scrape everyone's videos, slap some subtitles on them, and then TikTok get to make some money off of copyrighted content that doesn't belong to them.
Google over the years has continued to degrade the search engine experience. Users have been bombarded with ads, pushed shopping recommendations before search results, shown "People Also Ask" boxes that kept users on Google for longer, and keyword-stuffed content was recommended over genuine expertise. But the most destructive of all of these changes happened with the introduction of AI Overviews. Users are now given an AI-generated response to almost all of their search queries.
Why exactly is this problematic? Google controls 90% of global search. Nine out of 10 times when someone is seeking information, Google decides what they find. And now, with AI Overviews, they're not just organizing the information, they're replacing it. When users are given an AI summary, it is Google's own AI models that are doing the summarization. And as theorized, Google is further controlling what information we access. When Google cites sources, it's often the same websites: Reddit, Wikipedia, YouTube, the same platforms they control or have deals with. The research also showed that only 1% of users click a link in AI summaries.
If we search advice for moving to Portugal, Google's AI summary cites one Reddit post and a few YouTube videos, but click "Tools" and there are 25 million results about moving to Portugal. Decades of expat blogs, government resources, and local insights, all invisible, all replaced by Google's single paragraph. Even Google's own AI Overview admits this is catastrophic and outlines how this hurts the internet and its users.
What exactly does an AI Overview mean for the internet? When Google uses other people's content in its AI summaries, no traffic goes to these sites, and no traffic means no revenue. Content creators, researchers, publishers, independent journalists are no longer incentivized to create human content. Google feeds on human content to train its AI. It then uses that AI to make human-created content worthless.
But you may be wondering why Google would do this. After all, no clicks means no ad revenue generated from Google searches. This is the result of users no longer Googling. Instead, people asked ChatGPT and Perplexity for answers. Perplexity just launched an AI browser. OpenAI is building one, too. The introduction of these browsers could eliminate people visiting other websites entirely.
With corporations stealing content from humans, and humans no longer being incentivized to upload their content to the internet, many sites will die off. AI bots are cheaper to employ than humans. We see more companies adopting AI bots to do content moderation and community management. An internet that has more traffic from bots than humans and has more content being generated by AI than humans creates a premium for authenticity.
After Elon purchased Twitter, he took advantage of the fact that bots controlled the platform. He was able to sell users verified check marks for a monthly subscription. The more fake the internet becomes, the more we'll pay to prove we're real. With every platform's algorithms rewarding engagement, this incentivizes corporations, content creators, and influencers to boost their posts with bots. In doing so, they hope that the algorithm will pick up their post and show it to more people. And while platforms do occasionally purge bots, they also benefit from this artificial engagement.
Platforms make money from users viewing ads. In 2024, US social media ad spend reached $82.7 billion. The more users a platform has, theoretically, the more revenue they can generate from advertisers. Highly engaged users, including bots programmed to click on ads, increase the amount of advertising dollars spent.
When the internet is dead, corporations control the narrative. Fake reviews and bot comments create illusions of consumer enthusiasm that become self-fulfilling. Real humans follow what appears to be the crowd, unaware that they're following machines. As bots and AI increasingly outnumber us, distinguishing real from fake becomes impossible. We read opinions, believing they're human. But when thousands of accounts spread identical messages, fringe ideas gradually seem normal. We shift our own beliefs without even realizing it. The ultimate corporate influence, invisible and complete.
A dead internet has a cascade of negative consequences. We will have knowledge stagnation. Without fresh human insights, AI systems recycle increasingly outdated information. Innovation will take a hit. Why create new tools or services if the only few humans left on the internet can't even discover them? And if the majority of internet traffic becomes bots and AI agents, why would companies continue paying to host their websites online? While small businesses and independent journalists struggle, the massive platforms that host the last few real humans online become bigger monopolies.
"Is this real?" will become the most commonly used phrase. Past content and saved archives will disappear. All information will flow from a few corporations and the governments. AI trained on AI-generated content becomes progressively worse, leading to more hallucinations and inaccurate answers. Experts go extinct. Why become an expert in something if AI knows everything? Fringed societies will start to develop. As some users fully embrace AI as their friend, their therapist, their lover, others who remember the early days of the internet, the days before AI, reject technology completely in favor of forming real human connections. Where previous generations asked, "Do you remember the days before the internet?" Our generation will ask, "Do you remember when the internet was real?"
"Our scientists have done things which nobody's ever done before." "Yeah." "Yeah. But your scientists were so preoccupied with whether or not they could, they didn't stop to think if they should."