Investigation

AI agents are replacing human workers at massive scale—85% of customer service, 30% of HR roles eliminated. But GPT-4 scores just 33% above random guessing on strategic economic reasoning. Both numbers are real. That’s the paradox.

Prediction markets are treated as truth oracles, but they fail systematically under specific conditions. I started this essay believing they were fundamentally flawed. The research changed my mind.