What can neuroscience learn from math and machine learning?  Are multiple theories valid in parallel or can we rule out competing theories?