We Built the Dead Internet Ourselves—It was never about the bots.
This is a follow-up to my earlier piece, The internet got what it deserved—and how to fix it. It stands alone, but the two are better together.
Dead Internet Theory, if you haven't gone down that rabbit hole, goes something like this: the internet is mostly bots now. Fake engagement, AI-generated slop, sockpuppet accounts running at scale. Real humans are a dwindling minority, screaming into an automated void.
The conspiracy version is a bit paranoid. But the feeling it describes? That one's real.
Here's what I think actually happened.
A few years ago, I was making videos for social media. Talking to my audience. Being authentic. Genuinely, actually authentic. And the tactic worked the way it's supposed to. People replied like the video was made just for them. Deeper connection, they said. Like I was speaking directly to them.
And it made me feel really weird.
Not because I was lying. It was my real voice, my actual thoughts. But I was broadcasting something designed to feel personal to many people at once. When someone felt truly seen by a video I made for everyone, I felt the machinery underneath it. I eventually stopped making them. Not because the strategy failed. Because it worked.
That's Dead Internet Theory. Not bots replacing humans. Humans optimizing themselves into bots.
We do this. Humans. We build something, we find the exploit, we scale the exploit until the thing we built doesn't work anymore, and then we're genuinely surprised. Content farms, keyword stuffing, 2,000-word guides written to rank and never to help—those aren't aberrations. They're the logical endpoint of a system that decided attention is money, volume drives attention, therefore maximize volume at any cost. The dishonesty was always incentivized. We trained the machines on that web. Then we were surprised when the machines sounded hollow.
That's the passive version of the problem. Nobody meant to do harm. They just chased the wrong metric long enough to hollow everything out.
The active version is darker.
Russian bot networks. Coordinated inauthentic behavior. AI-amplified smear campaigns that spread faster than any correction ever could. Comment sections flooded with accounts that don't exist, pushing narratives that serve someone's agenda. These aren't machines acting autonomously. These are humans, with specific goals, deploying bots as weapons against your ability to know what's real, against your trust in other people, against the basic functioning of shared information.
Dead Internet Theory gets blamed on AI. But AI is just the latest tool in a project that's been running for years. The project is: make you uncertain enough that you stop trusting anything, and then fill that vacuum with whatever the person holding the firehose wants you to believe.
That's not a glitch. That's a feature.
What makes the dead internet feel dead is that you can't tell anymore. Not just human versus bot. You can't tell if anything in front of you was made for you, or made to approximate you closely enough that you'd keep scrolling, keep doubting, keep clicking. The simulation of connection and the simulation of truth have both been optimized past the point where you can find the real thing underneath.
And then the pandemic happened. The in-person communities a lot of us had, or were building toward, didn't come back the same way. Division got state-sanctioned. The desire for gathering, for actually being somewhere with people, got complicated in ways that haven't fully untangled. And the internet was there. It was always there. Offering connection at scale to people who had fewer and fewer places to find it.
That's not a small thing. That's also not a comfort.
Honestly, writing this piece felt weird too.
I've been second-guessing my own vocabulary for years. I think in a certain register and I've learned, slowly and annoyingly, that using it makes people disengage. So you start editing before you even open a document. You swap the precise word for the accessible one. You round off the edges. And at some point you look at what you've written and it's technically yours but it's also been processed for an audience that may or may not exist, on a platform you didn't design, optimized for attention you're not sure you want.
That's before we even get to AI.
There's something uncomfortable about showing up online personally, using my real experience and my actual ambivalence, knowing it also functions as content. Knowing it might perform. I wrote this because I meant it. And I'm aware that "I wrote this because I meant it" is exactly what optimized content sounds like.
I don't know how to fully escape that loop. I'm not sure anyone does right now.
What I do believe is that building honestly, content actually for humans and not for the performance of them, is the only thing that doesn't eventually rot. I believe that structurally, not just ethically. It's what builds trust that doesn't need to be re-earned every algorithm update.
And I'm also aware of the world that belief has to exist inside. The data being harvested. Platforms consolidated into fewer and fewer hands with clearer and clearer agendas. Infrastructure increasingly owned by people who are not neutral about what it's for. We built systems we didn't fully understand, and we keep being surprised when someone finds the exploit. Humans are very good at building things. We're not as good at anticipating the consequences of what we've actually built.
Being offline has become a privilege for some and a death sentence for others. So we're here. Showing up in a system we didn't fully choose, trying to do something real inside it.
For me, right now, that means building things that are honest. Not because it fixes everything. Because it's the only part I control.
This piece was copy edited with AI assistance. The thinking, the discomfort, and the social media videos I quit making are mine.
