Comment Sections Were the Internet's Town Square. Now They're a Landfill.

There was a brief, shining window — somewhere between 2005 and 2012 — when the comment section beneath an article felt like something real. People disagreed, sure. They were occasionally rude. But there were also regulars, running jokes, genuine corrections, and the occasional insight that was better than the article itself. It felt, against all odds, like a community.

That window is closed. Boarded up. Someone spray-painted something incomprehensible on the boards.

Today, scrolling through a comment section is less like visiting a town square and more like wandering into a gas station at 2 a.m. — you might find what you need, but the odds are not in your favor, and someone is definitely going to say something that ruins your evening.

How did we get here? Predictably, as it turns out. Embarrassingly predictably.

Step One: The Platform Decides Engagement Is the Same as Quality

The rot starts at the top, and it starts with a metric. Somewhere in a product meeting — probably in a room with exposed brick and a kombucha tap — someone decided that the number of comments was a useful proxy for the health of a community. More comments meant more engagement. More engagement meant more ad impressions. More ad impressions meant the quarterly numbers looked good.

What nobody asked was: what kind of comments?

Algorithmic sorting made it worse. When platforms switched from chronological feeds to engagement-ranked ones, they handed the keys to whoever could generate the most reactions. Nuanced takes don't generate reactions. Measured disagreement doesn't generate reactions. A guy comparing the article's author to a historical war criminal? That generates reactions. So that's what floats to the top.

You didn't design a comment section. You designed a rage amplifier and called it community.

Step Two: Moderation Gets Expensive, So It Gets Abandoned

Human moderation is slow, costly, and — here's the kicker — it actually works. Which is precisely why most platforms treat it like a luxury item they can't quite justify.

For a while, the solution was volunteer moderators. Reddit built an entire empire on the backs of unpaid users willing to spend their evenings deleting slurs for free. It worked, more or less, until it didn't — until the community grew too large, the abuse too relentless, and the volunteers too burned out to continue. Then the spam crept in. Then the bad actors. Then the really bad actors.

Facebook tried AI moderation and famously, repeatedly, catastrophically failed — flagging breastfeeding photos while missing actual incitement to violence, because the algorithm was trained on engagement, not ethics. YouTube's comment section on anything involving children became so thoroughly colonized by predatory behavior that major advertisers pulled out entirely in what became known as the Adpocalypse.

The pattern is always the same: invest minimally in moderation, watch the section degrade, respond to the PR crisis with a press release about community standards, quietly do very little.

Step Three: The Bots Arrive, and Nobody Stops Them

If you've spent any time in a YouTube comment section in the last three years, you've met them. "Great video! I just made $4,700 this week working from home — ask me how!" Followed by fifty identical variations, posted by accounts created six days ago with profile pictures stolen from stock photo sites.

Bots have always existed on the internet. What changed is the sophistication — and the economics. Running a bot farm used to require some technical know-how. Now you can rent one. And with large language models capable of generating passable human-sounding text at scale, the next generation of comment spam doesn't just look like spam. It looks like your neighbor wrote it.

The platforms know this. They have the data. They have the detection tools. What they don't have is a strong enough financial incentive to act, because bot comments still count as engagement, and engagement still pays the bills.

Step Four: The Real Users Leave

This is the part that should keep platform executives up at night, but probably doesn't.

When a comment section becomes visibly toxic — when every post about a new product is buried under political arguments, when every news article devolves into a screaming match between accounts that may or may not be human — the people who were there for genuine conversation leave. Quietly. Without filing a complaint.

They don't delete their accounts. They just stop scrolling down. They stop contributing. And the community that remains is, by definition, the community that stays in toxic environments — which tells you something about who's left running the conversation.

This is why the comment sections on major news sites have become so uniformly unhinged. The New York Times eventually moved to a curated, limited comment system because the alternative was a daily reminder that civilization might be on borrowed time. ESPN's comment sections were so reliably awful that they became a punchline. Popular Science shut theirs down entirely in 2013, citing research showing that hostile comments literally changed how readers perceived the articles above them.

The readers who cared most about quality left first. That's not a coincidence. That's a design outcome.

Step Five: AI Completes the Circle

We have now arrived at the logical endpoint: comment sections populated, in significant part, by AI-generated text responding to AI-generated articles, moderated by AI tools, on platforms optimized by algorithms, viewed by humans who are increasingly unsure whether any of it is real.

The AI spam in comments isn't coming in the future. It's here. It's the weirdly enthusiastic five-paragraph response that agrees with the article in general terms without referencing anything specific. It's the "As an AI language model" slip that occasionally makes it through. It's the product review that reads like a brochure.

The humans are still in there, somewhere. But they're getting harder to find.

So What's Left?

The interesting thing is that good comment communities still exist — they're just smaller, more deliberate, and usually built by people who actively decided to prioritize quality over scale. Certain subreddits. Niche forums. Substacks with engaged subscriber bases. Newsletter reply threads.

The common thread is that someone, somewhere, is making active choices about what the community values and enforcing those choices with actual effort. It's not glamorous. It doesn't scale. It definitely doesn't optimize for quarterly earnings.

But it works, which is more than you can say for the alternative.

The comment section didn't have to become a dumpster fire. It became one because the people running the platforms decided, repeatedly and deliberately, that the fire was good for business.

They were right, until they weren't. And now we're all standing around the ashes wondering where everybody went.