history-society

Most expert predictions are barely better than chance. But Tetlock found rare individuals who consistently outperform — even beating intelligence analysts with classified data. The traits separating them from the rest have nothing to do with IQ.

Polls ask what people think. Prediction markets ask what people are willing to bet. That difference in skin-in-the-game consistently produces better forecasts — for elections, policy outcomes, and corporate events. Here's the mechanism, the track record, and the limits.

Ask a crowd to guess the weight of an ox and their average will be almost exactly right. But ask that same crowd about a complex policy question and you may get catastrophe. James Surowiecki's framework tells you precisely when collective intelligence works — and when it fails.

The wisdom of crowds requires diversity, independence, and good aggregation. Remove any one ingredient and you get a mob. This is what the science of collective intelligence actually says about when groups are worth trusting.

Groups can be smarter than their smartest member — or dumber than their dumbest. The difference comes down to information aggregation, diversity of views, and independence. Understanding the conditions that unlock collective intelligence changes how you design teams and decisions.

The standard story casts human institutions as stubbornly static. The evidence says otherwise. From non-hierarchical megacities in prehistory to democratic experiments in unlikely places, humans have been more inventive at building social structures than our pessimism gives them credit for.