Logan Riggs comments on Power Laws in NNs: A Possible Mechanism for Inductive Bias towards Sparse Representations