AI interpretability has the same problems as philosophy of mind