Social media platforms are structurally engineered to reward outrage, and rational actors are responding accordingly

Anne Applebaum, the historian and political writer, puts a number on something most media observers treat as diffuse intuition. A negative political article, she says, will get “10 times the reach, 10 times the engagement, 10 times the subscribers” of an article taking the opposite position on the same subject. That is not a complaint about reader taste. It is a description of the economic structure that shapes what gets written and published.

Francis Fukuyama, the political scientist, locates the same mechanism at the individual creator level. The platform reward for unreasonable or insulting content is not gradual. It arrives immediately, as a spike in followers. What follows from that spike is predictable: to maintain the growth rate, the next statement has to be more extreme than the one that produced it. The escalation is not irrational given the incentives. It is the logical response to them.

That logic is no longer just an emergent property of how audiences behave. Zach Yadegari, an entrepreneur, has described how he and collaborators made the choice deliberately during a course launch. The strategy, as he put it, was to go “the super controversial route,” modeling it on Andrew Tate-style tactics: “sell the dream, sell the lifestyle instead of sell the product.” Yadegari acknowledged the approach was “definitely not a real representation of any of our characters.” They “very intentionally wanted to rage bait people.” The incentive structure had been so thoroughly internalized that the gap between a creator’s actual character and the persona deployed for engagement was not a source of friction. It was a feature of the plan.

The downstream consequences reach beyond political opinion or personal branding. Ed Zitron, a media critic, points to an Oxford Economics study that circulated widely as evidence that artificial intelligence is eliminating jobs for young people. Zitron says that when the study was actually read, it contained a single line noting some correlation, with no quantification and no supporting data. Journalists who covered it, he argues, did not read that far. The outrage framing spread; the methodological thinness did not. The study’s reach was decoupled entirely from its substance.

The same dynamic operates in corporate communication. Zitron points to how Microsoft reported a figure of around 37 to 38 billion dollars as an annualized run rate in artificial intelligence. Audiences, he argues, hear “they made 37 billion dollars” when the metric is undefined and does not represent realized revenue. The framing is calibrated to produce a reaction. The absence of a definition is not an oversight.

Taken together, these observations describe a self-reinforcing loop with no obvious internal corrective. Platforms reward negativity because engagement metrics treat all attention as equivalent. Creators respond by producing more negative content because the reward is real and measurable. Media organizations face the same incentive at institutional scale: the negative version of any story outperforms the positive one by a margin that makes editorial restraint economically costly. And false or outrageous claims benefit from all of this disproportionately, because they face no accuracy tax in the engagement calculus.

What Applebaum and Fukuyama are describing, from their different vantage points, is not a media pathology that better editorial standards could fix in isolation. The ratio is baked into the measurement system. A publication that consistently chose the less enraging story would, by the logic of the platform, reach fewer people. A creator who declined to escalate would grow more slowly than one who did not. The individuals operating inside this structure are not, for the most part, malicious. Some, like Yadegari, are simply treating the incentive as a design parameter. The structural bias toward outrage does not require bad actors to sustain itself. It requires only rational ones.