Scott Alexander investigates why self-reported IQ scores in online surveys are implausibly high, identifying issues with conversion tools, reporting bias, and unreliable tests. Longer summary

Scott Alexander investigates why self-reported IQ scores in online surveys seem implausibly high. He analyzes data from two surveys (Less Wrong 2014 and Clearer Thinking 2023) and identifies three main issues: 1) A popular SAT to IQ conversion tool was inaccurate, inflating scores. 2) Only the smartest people tend to report their SAT scores, skewing the sample. 3) Self-reported IQ test scores are often inflated or based on unreliable tests. After accounting for these factors, he estimates the average IQ of the Clearer Thinking sample at 111 and the Less Wrong sample at 128, which he considers more plausible. Shorter summary

Scott Alexander reviews recent developments in prediction markets, including Google's internal market and Metaculus' 'fortified essays', while discussing challenges and applications in various fields. Longer summary

Scott Alexander discusses recent developments in prediction markets, including Google's internal prediction market Gleangen, challenges in long-term predictions, conditional markets for policy decisions, and new weight loss drugs. He also covers Metaculus' 'fortified essays' concept, which combines expert analysis with crowd forecasts, and shares recent predictions on topics like COVID-19 hospitalizations and SAT requirements in college admissions. Shorter summary