Why is it so difficult to predict events and consequences too far into the future? Scientists and philosophers have studied the process of how knowledge is expanded. And what they have learned has lessons for understanding where generative AI models may go.