Leo wrote a lengthy critique of the idea proposed by a16z's research partner, a professor of political economy at Stanford Graduate School of Business, to use large language models to judge and predict markets. In short, the idea seems appealing, but it's not very reliable in practice. Last year, I saw many prediction market projects mentioning similar approaches, but I've always felt this method has problems (can we really replace oracles with large models?). Of course, this theory is too complex; it could probably be discussed in ten podcast episodes... Feel free to raise any questions you're interested in learning more about and discussing, and we'll see how to arrange it. @jack_xiong137