Galit and I discuss the independent roles of prediction and explanation in scientific models, their history and eventual separation in the philosophy of science, how they can inform each other, and how statisticians like Galit view the current deep learning explosion.