Scott Alexander, curated
← Back to curation

Why Not More Excitement About Prediction Aggregation?

Quality
58
Solid
Claude Shift
42
Moderate
RWI
2
of 10

Summary

Prompted by a Metaculus sidebar ad, Scott asks why crowd-sourced prediction aggregators haven't caught on, given they dodge the regulatory and 'profiting off terror' objections to money markets and Tetlock's superforecasting proved they work. The substantive core is a quoted reply from Metaculus's Prof. Aguirre: there's no 'secret sauce'; the power needs *lots* of people/questions/data plus recalibration (so a single firm gains little); and most decision-makers want to know what *will* happen, distrust 'squishy' 70-30 numbers, and don't know what to do with a probability. Scott ends puzzled at the 'stupid bottleneck.'

Why this score

Quality 58 · Solid. A minor, partly promotional musing whose best content (why probabilistic forecasts resist adoption) is borrowed from the quoted Aguirre. Genuine question, thin original contribution. Solid.

Claude’s paradigm shift 42 · Moderate. Largely recapping Hanson/Tetlock/Cowen on why prediction markets don't spread; little new. Moderate-low.

Real-world impact 2 · Minor. A small signal-boost for Metaculus (a platform that did grow); the post's own causal contribution is slight. Within-blog influence.