Concept · part of What the filter rewards

Ask people what they wish would spread online and most give the same answer, whatever their politics: more positive, nuanced and accurate posts, and less outrage. Then look at what spreads. The researchers behind Psychology of Virality call the gap the paradox of virality: widely shared content is not necessarily widely liked.

The gap rules out the simplest explanation of a bad feed, that people get what they ask for. The book reads it as evidence about whose choice is being made. The choice of what spread was not the reader’s. It was the platform’s: the ranking algorithm and the design around it, tuned for time spent. One audit the same paper reports found the Twitter/X algorithm amplifying political content that did not match people’s stated preferences. In the book’s argument about selection, the paradox is the sharpest sign that the dominant filters serve someone other than the people scrolling.

Hugo Mercier offers a second reading that needs no algorithm. People share alarming claims as signals of belonging while holding them only as reflective beliefs they never act on. Vigilance also has to compete with identity and threat at the moment of sharing. The book keeps both readings.

Read the full argument

How it connects


What the filter rewards (1 of 10) · Superspreader dynamics →

Explore all the ideas · Read the book in order · Open the interactive map