How Can AI Exist When Human Writers Vanish?

Is Blogging Actually Dying? The Uncomfortable Truth Nobody Wants to Talk About

Somewhere between the rise of AI chatbots and the endless scroll of TikTok, the internet quietly started eating itself.

 

Remember the early 2000s internet? Not the polished, algorithm driven thing we have now I mean the scrappy, passionate version where some random person in their basement had written the exact tutorial you needed, and Google somehow found it for you.

That was genuinely magical. And most of us never stopped to think about why it worked.

 

It Wasn't Free. It Just Felt That Way.

Here's something people forget: the "free internet" was never actually free. It ran on a business model that was, when you think about it, kind of beautiful in its simplicity.

Someone passionate about Linux, or woodworking, or fixing vintage cameras would pour hours into writing a detailed, genuinely helpful article. Google would surface it to the next person who needed that information. Ads on the page would earn the writer a few dollars. Those few dollars would convince them to write another article. Rinse, repeat, and over two decades you end up with an internet absolutely overflowing with human expertise on every topic imaginable.

Readers got knowledge for free. Writers got paid. Google played matchmaker. Everyone won.

The whole thing depended on that loop staying intact. And for a long time, it did.

 

Then Came the Videos

Somewhere around the mid 2010s, people started watching instead of reading. YouTube grew up. TikTok arrived. Instagram Reels happened. Suddenly, "how to fix a leaky faucet" wasn't a blog post you'd find through Google it was a five minute video with 2 million views.

And look, I get it. Video is great for certain things. Watching someone's hands as they fold origami is genuinely better than reading instructions. Nobody's debating that.

But here's the thing that always bugs me: try debugging a Python error from a YouTube video. Try following a 47 step server configuration from a TikTok. At some point, you need words on a screen you can actually read, copy, and search through. Video has real limits for technical content but that didn't stop audiences from drifting toward it anyway.

Blog traffic fell. Ad revenue dried up. Some writers pivoted to video themselves. A lot more just quietly stopped.

That was the first crack.

 

Then AI Showed Up and Finished the Job

I don't think most people appreciate how completely AI chatbots broke the old model. Not damaged it broke it.

Think about the last time you Googled a technical question and actually clicked through to a website. Really think about it. Because I'd bet it's been a while. Why would you? You type your question into ChatGPT, you get a clear answer in ten seconds, and you're done. No tabs, no scrolling past ads, no figuring out which Stack Overflow answer from 2014 still applies.

It's genuinely convenient. I use it constantly. But here's what's happening behind the scenes that nobody talks about at the dinner table:

When you get that answer from an AI and don't visit any website, the blogger who spent three hours writing the article that trained the AI on that topic gets zero. No traffic. No ad impression. Nothing. The AI consumed their work, gave you the answer, and cut them out of the transaction entirely.

So the blogger looks at their analytics, sees the cliff edge drop in visitors, does the math on their hosting fees, and thinks: why am I still doing this?

And they stop.

 

"But Can't AI Just Learn from Videos Now?"

Every time I bring this up, someone says something like "okay but who cares about text blogs when AI can just watch YouTube?"

I used to half believe this. Then I actually thought about how AI processes video, and it kind of broke my brain a little.

When an AI "watches" a video, what it's actually doing is reading the transcript of what was said out loud, and scanning the title and description. That's it. It's reading text that happens to accompany a video. It cannot look at someone's hands soldering a circuit board and extract the knowledge of how much pressure they're applying, or notice that they slightly modified the technique from the diagram. A human watching that video learns something. The AI reads the narration and moves on.

The tech companies will tell you their models are "multimodal" and can understand images and video natively. And sure, there's some truth to that for basic visual recognition. But deep, contextual, practical knowledge? That still comes from text, written by humans who sat down and actually thought something through.

Text is, and has always been, what AI actually eats. And we're burning down the farms.

 

Here's the Part That Should Actually Scare You

Okay, so bloggers are struggling. That's bad for bloggers. But why should anyone else care?

Because AI doesn't generate knowledge. It aggregates it. There's a massive difference.

An AI chatbot has never debugged code at 2am, never disassembled a broken appliance and figured out what went wrong, never spent three years mastering sourdough and finally cracked why the crust kept going wrong. It's read about all those things because humans wrote about them. Feed it less human writing, and it knows less. Simple as that.

And here's where the loop gets genuinely terrifying: as human written content disappears, AI systems will increasingly train on each other's outputs. One model generates an article. Another model trains on that article. A third model reads the second model's training data. Researchers call this Model Collapse and it's basically what happens when you photocopy a photocopy of a photocopy. Each generation gets blurrier, more distorted, further from the original truth.

A new programming framework comes out. Nobody writes a blog post about it because there's no money in blogging anymore. So AI never learns it exists. Someone asks about it; the AI either confabulates something plausible sounding or just draws a blank. Six months later, the same thing happens with the next framework. And the one after that.

Eventually you get an AI that's confidently stuck in the past, answering questions about a technological landscape that no longer exists.

 

So What's Anyone Actually Doing About This?

The big AI labs have started throwing money at the problem licensing deals with Reddit, Stack Overflow, major news publishers. Paying millions for access to human written content. Which is honestly an admission that they know exactly what's happening.

But those deals flow to large platforms. The person writing niche technical tutorials from their apartment the person who represents the real depth and diversity of internet knowledge sees none of that money. There's no mechanism for it. Nobody's built one yet.

Will someone figure it out before the independent creators all give up and leave? Or will they have already left by the time the solution exists?

I genuinely don't know. And I'm not sure anyone does.

 

The Thing Nobody Wants to Say Out Loud

Here's the uncomfortable truth at the center of all this:

Google and the AI companies got extraordinarily wealthy by building products on top of freely available human knowledge. The deal implicit, never written down was that creators would keep producing that knowledge because the traffic it generated paid their bills. That deal is now broken. Not renegotiated. Not restructured. Just... quietly broken, while everyone was looking at their phone.

When the writers finally walk away for good, the AI will lose the very thing that makes it useful. The systems that replaced the bloggers will starve without them.

There's a phrase floating around in AI research circles: the "Digital Dark Age." An internet so saturated with low quality, AI generated noise that finding something actually true, written by someone who actually knew what they were talking about, becomes genuinely difficult. Like searching for a specific book in a library where someone replaced half the collection with convincing looking fakes.

We're not there yet. But we're moving in that direction faster than most people realize.

Whether we course correct in the next few years or just keep sleepwalking forward is, I think, one of the more important questions of this decade. It just doesn't trend on Twitter, so nobody's really talking about it.

Maybe we should be.

 

If you read this far, thanks genuinely. And if you have a favorite blog or independent website you still love, maybe go visit it today. Click around. Disable the ad blocker for a minute. The people keeping the real internet alive could use it.