AI Just Beat Prediction Markets | MOONSHOTS

In this brief discussion, Speaker A announces that GPT-5.5 has outperformed traditional prediction markets, specifically beating the Polly Market crowd forecasts for the Super Bowl. He frames this achievement as the start of AI providing "wisdom," likening it to an elder council that guides decisions based on experience, but with AI capable of running billions of high‑resolution simulations to identify the most probable successful outcomes. Speaker B draws a parallel to Isaac Asimov’s concept of psychohistory from the Foundation series, noting that even the current state‑of‑the‑art models achieve only about 25% accuracy, which he dubs the worst possible psychohistorical model. The exchange highlights both the practical breakthrough in forecasting and the broader philosophical implications of AI as a predictive, wisdom‑providing tool.










