What If Your Comment Section Had an Editor Instead of a Bouncer?

TL;DR: Traditional moderation works like a bouncer — it removes bad comments without explanation, teaching nobody anything. An editor model coaches commenters to revise before posting, preserving their opinion while removing the hostility. AI makes this coaching possible at scale for the first time.

Every comment section on the internet is run by a bouncer.

Think about it. The bouncer stands at the door, arms crossed, and decides who gets in and who gets tossed out on the sidewalk. He doesn’t explain himself. He doesn’t tell you that maybe you’d have better luck if you weren’t wearing flip-flops. He just says “not tonight” and turns his attention to the next guy in line.

That’s content moderation in a nutshell. Post something over the line, and it gets deleted. Do it again, and you get banned. Nobody tells you what the line was, or why you crossed it, or how you might have said the thing you wanted to say without crossing it. You’re just out on the sidewalk, angrier than when you arrived.

And here’s the thing — we’ve been running comment sections this way for twenty-five years, and by every measure I can find, comment sections have gotten worse. Maybe the bouncer isn’t working.

Nobody ever learned anything from a bouncer

I’ve spent a lot of years as a manager, and if I learned one thing, it’s that people don’t improve when you punish them. They improve when someone shows them a better way.

When I was a young naval officer, I wrote plenty of terrible reports. (I mean genuinely terrible. I apologize to anyone who had to read them.) But my commanding officers didn’t delete my reports and ban me from writing. They marked them up. They said, “This paragraph buries your point. Lead with it.” And slowly, painfully, I learned to write.

That’s what an editor does. An editor doesn’t reject your work — an editor makes your work better, and in the process makes you better. Every professional writer will tell you the same thing: the editor is not the enemy. The editor is the reason the piece is good.

So why does nobody edit comments?

What editing a comment actually looks like

Let’s make this concrete. Suppose someone reads an article about a new city budget and fires off this comment:

“Only a complete idiot would think this budget makes sense. Typical clueless politicians wasting our money AGAIN. Anyone defending this garbage is part of the problem.”

The bouncer’s answer is easy: delete it, maybe ban the guy. And sure, the comment section is marginally more pleasant for the next ten minutes. But notice what got lost. Buried under all that hostility is an actual opinion — this person thinks the budget wastes money. That’s a legitimate position! It might even be right. The bouncer threw out the argument along with the attitude.

Now imagine an editor got there first, before the comment ever posted:

“You clearly feel strongly about this budget. Right now, though, your comment attacks the people rather than the plan — which means readers who disagree will dismiss it instantly. What specifically do you think is wasteful? Naming one or two line items would make this a much stronger argument.”

And maybe, thirty seconds later, the comment that actually gets posted reads:

“I think this budget is a serious mistake. We’re spending $4 million on a parking study while the roads are falling apart. Can anyone defend that priority?”

Same person. Same opinion. Same passion, even. But now it’s a comment that starts a conversation instead of a food fight. Nobody got censored. Nobody got banned. Somebody got coached — the way an editor coaches a writer.

But people don’t want to be edited

I can hear the objection already: people will hate this. Nobody likes being told to revise.

Except we already know that’s not true, because we’ve all worked with editors, teachers, and code reviewers, and the good ones are the people we remember most fondly. What people hate is being rejected without explanation. The bouncer offers no path forward. The editor’s entire job is the path forward.

And notice something else: the editor scales in a way the bouncer never did. Human moderators can’t possibly coach every comment — that’s why we settled for deletion in the first place. Deletion was cheap and coaching was expensive. But that constraint just evaporated. We now have software that can read a comment, understand what the person is actually trying to say, and suggest how to say it better — instantly, at any volume, before the comment ever goes live. The economics that forced us to hire bouncers no longer apply.

The whole industry, though, is still building better bouncers. Every moderation tool on the market — and I’ve looked at most of them — is a variation on the same theme: detect the bad thing, remove the bad thing, remove the person who posted the bad thing. Faster bouncers. Smarter bouncers. AI-powered bouncers. Nobody stopped to ask whether the bouncer was the right model in the first place.

The comment section you actually want

Here’s my claim, and I’ll plant a flag on it: bad comment sections are not a filtering problem. They’re a communication skills problem. Most people who write hostile comments aren’t trolls — they’re ordinary people with genuine opinions and no training in how to disagree productively. Trolls exist, sure, and for them, fine, keep a bouncer on staff. But the vast majority of “bad” comments are just good opinions badly expressed.

Delete those comments and you’ve taught nobody anything. Coach them, and something remarkable happens: your comment section slowly fills up with people who have learned — comment by comment, nudge by nudge — how to argue well. That’s not a moderated community. That’s a good community.

This is, as you might have guessed, the idea behind Respectify. We’re building the editor. Everyone else is building bouncers, and honestly, they’re welcome to that business.

Because here’s what I keep coming back to: nobody ever walked away from a bouncer a better writer. But every good writer I know will tell you about the editor who made them one.