“Travis Kalanick observes that even the best current AI agents are still unintelligent and prone to making basic errors, such as taking both sides of the same bet in an investment context.”