ChatGPT Can’t Pick the Stocks
Also private credit pricing, a fake margin loan, shipping charter benchmarks and Paul Anka.
I often write around here that there are two fundamental ways to use artificial intelligence to trade stocks:
The first model is, approximately, a “quant hedge fund.” People have been doing some form of this for a long time, and some of them (Renaissance Technologies comes to mind) have had a great deal of sustained success. This model works. Not always, not for everyone, and you might reasonably worry about risk or crowding or regime changes or whatever. But hiring really good machine learning engineers to build models to predict stock prices makes sense both theoretically and empirically.