There’s a fringe idea that’s been circulating in the darker corners of the internet since around 2021. It goes like this: the internet is mostly dead. The accounts you interact with aren’t people. The content you consume was generated by bots, shaped by algorithms, and seeded by corporate and government actors to manufacture the illusion of organic human activity. Real humans, the theory goes, are a minority presence on a web that has been quietly hollowed out and replaced with synthetic noise.
When I first encountered it, I filed it away as paranoid. Interesting, but paranoid.
I don’t file it away anymore.
What the Theory Actually Says
The Dead Internet Theory isn’t just “there are bots online” … everyone knows that. It’s a more specific and unsettling claim: that the character of the internet has fundamentally changed. That the web used to be driven by genuine human curiosity and connection, and that somewhere in the 2010s it was replaced by something else … a managed, artificial environment designed to optimize for engagement, ad revenue, and ideological influence at a scale that makes authentic signal nearly impossible to find.
Proponents point to the consolidation of traffic around a handful of platforms, the rise of AI-generated content farms, the documented existence of state-sponsored bot networks, and the uncanny feeling … familiar to anyone who spends time online … that you’re often not sure if you’re talking to a person.
Again: I used to think this was paranoid.
What I See From the Publishing Side
I run a publication. I watch our traffic, our referrals, our engagement. I talk to other editors and writers. And here’s what I can tell you from the inside:
The content ecosystem is flooded. In the last two years, the volume of content being published across every vertical has exploded in a way that has no relationship to the number of humans who could plausibly be writing it. SEO-optimized articles, “thought leadership” posts, news summaries, listicles … they appear in quantities that are simply not human. We know this because we’ve seen what human publishing capacity looks like, and this isn’t it.
Engagement metrics have become unreadable. Clicks, shares, comments … they’ve always been gameable, but now they’re aggressively gamed. We’ve published pieces we know resonated deeply with real readers, and watched them get outrun by hollow AI-slop with better distribution. The feedback loop that used to tell you something about what people actually cared about has been corrupted.
Discovery is broken. The pathways that used to connect real writing to real readers … Google Search, social referrals, organic shares … have either collapsed or been taken over by platforms optimizing for something other than quality or relevance. Google’s search results are increasingly a display case for Google’s own products and AI-generated summaries. Social platforms bury links to keep users on-platform. The infrastructure of discovery, which the open web depended on, has been quietly dismantled.
The audience feels smaller, even as the numbers grow. This is the hardest one to explain, but every publisher I talk to feels it. There’s a qualitative difference between the engagement we get now versus five years ago. Comments that feel less human. Email replies that feel templated. A vague suspicion, when you look at your analytics, that a meaningful portion of what registers as “an audience” is not an audience in any meaningful sense.
The Dead Internet Theory in its most literal form … that there’s a coordinated, centralized effort to fake the internet … probably overstates the intentionality involved. What I’m describing doesn’t require a conspiracy. It requires an incentive structure.
Platforms are incentivized to show you content that keeps you on-platform, regardless of whether that content is human or genuine. Advertisers are incentivized to buy impressions without scrutinizing their source. AI companies produce tools that make synthetic content trivially cheap to generate. Bad actors … commercial and political … use those tools at scale. No single mastermind required. Just a market that rewards noise over signal, quantity over quality, and engagement theater over actual human connection.
The result, though, is functionally similar to what the Dead Internet Theory describes: a web where the human presence is increasingly thin, increasingly hard to find, and increasingly irrelevant to how the system actually operates.
What This Means for Publishing
Honestly? I’m still figuring it out.
The optimistic read is that scarcity of authentic human voice becomes a differentiator. When everything is synthetic, genuinely human writing … specific, embodied, wrong sometimes, personal always … becomes more valuable, not less. There’s a version of this future where real publications, with real writers and real editorial judgment, find that readers will pay directly for the thing that can’t be faked.
The pessimistic read is that we’re already past the tipping point. That the discovery mechanisms are too broken, the noise too loud, the audience too conditioned to the frictionless consumption of synthetic content to seek out the real thing. That publishing, as a business model, cannot survive in an environment where the infrastructure of distribution has been captured by entities with no interest in what we’re trying to do.
I hold both possibilities at once. Most days, depending on what I’ve just read in my analytics, one feels more true than the other.
The Question Worth Sitting With
Here’s what I keep coming back to: if a significant portion of what we call “the internet” is now synthetic … content generated by AI, engagement manufactured by bots, audiences that exist only as data points … what does that mean for the idea that the internet is a reflection of human culture?
We built enormous institutions and industries and political movements on the premise that the internet showed us what people actually thought, wanted, and cared about. We made editorial decisions, investment decisions, democratic decisions based on signals that we now have good reason to believe were, in substantial part, fake.
The Dead Internet Theory might be paranoid. But it’s asking a question that everyone who works online should be taking seriously: how much of this is real?
From where I’m sitting, the answer is: less than it used to be, and less than we’d like to admit.
This post was written by a human. I know that’s a strange thing to have to say. (lies)
