And the clearest way they serve themselves is by keeping you divided, because a divided public is easier to hold, sell to, and mobilize than a public that thinks for itself.

So here’s the question worth sitting with: what happens if that same public is given balance instead? Not told what to think, but actually handed the strongest version of every side, and trusted to reason it out? That single shift — balance instead of a feed built to keep you agreeing with yourself — is a real threat to how these systems currently make their money and hold their power. This article is about why.

Your Attention Is the Product

Think about the last time you picked up your phone “just to check something” and put it down forty minutes later. That wasn’t an accident, and it wasn’t a lack of willpower. It was the intended outcome of a system engineered by people whose job is to keep you scrolling. Every extra minute you spend is worth money to someone — usually an advertiser. The platform isn’t paid by making you smarter, calmer, or more understanding of your neighbor. It’s paid by keeping your eyes on the screen.

And here’s the part that matters most: not all content is equally good at doing that. Researchers at Yale who study this — the “attention economy,” as it’s called — found that content triggering outrage and moral anger holds people far more effectively than calm or nuanced content does. Posts attacking “the other side” get dramatically more shares than posts simply supporting your own. Anger isn’t a side effect of the algorithm. It’s the fuel.

Why the Platform Can’t Afford to Challenge You

Now flip it around. Imagine a platform that regularly showed you well-reasoned arguments against what you believe. Sounds healthy, right? Here’s the problem, from the platform’s chair: the moment you feel challenged, you don’t necessarily reconsider your view — you’re just as likely to decide the platform itself is biased, wrong, or “up to something,” and you use it less. From a business standpoint, that’s a losing move. Confirming you is safer than confronting you. It keeps you comfortable, keeps you clicking, keeps advertisers happy.

This isn’t new, and it isn’t unique to algorithms. Decades before social media existed, researcher Cass Sunstein studied what happens when like-minded people only talk to each other. The finding: they don’t just stay put in their views, they drift further toward the extreme — because they keep hearing new reasons to agree and almost never hear a real reason to doubt. Social media didn’t invent that pattern. It just built a machine that does it to millions of people at once, automatically, because doing so happens to be profitable.

Put those two mechanisms side by side and you can see the trap: anger at people outside your group drives sharing. Comfort inside your group drives loyalty to the platform. Genuine, well-sourced disagreement drives neither. It’s the one thing that would actually serve your judgment — and it’s exactly the thing the system has the least reason to ever show you.

The Role, Not the Person

This same pattern shows up beyond social media, in any role that holds concentrated power. A platform owner’s job, structurally, benefits from an engaged and divided audience. The role of a head of state controlling a major economy or military benefits from a loyal, clearly sorted base that shows up reliably — a public split into firm camps is simply easier to mobilize than a public weighing each issue on its own. None of this requires anyone in these roles to be plotting anything. It only requires that the role rewards division and doesn’t reward unity — and whoever holds it has little personal reason to fight the very incentive that’s making them successful.

That’s what makes this a structural problem, not a “bad person” problem. Swap out whoever holds the role, and the incentive is still sitting there, pointed the same direction. The next person walks into the same machine.

Everyone Pays, Almost No One Profits From Fixing It

Here’s the part that should actually bother you: tolerance, mutual understanding, and the ability to solve shared problems together are things everyone benefits from — but nobody profits from producing them directly. There’s no ad revenue in “citizens understood each other a little better today.” Division gets monetized. It gets mobilized into votes and headlines and shares. Unity mostly just sits there, doing quiet public good that nobody’s business model rewards — which is exactly why, left alone, division keeps winning.

That’s not an argument for despair. It’s an argument for design. If division is what the current system rewards by default, then a system deliberately built to reward the opposite — one that shows people the real, strongest case on every side before asking them to decide anything — isn’t just “nice to have.” It’s the only kind of system that has a chance of producing a public that can actually govern itself, instead of one that’s simply easier to sell to and easier to move.