I'd love to see the distribution of random seeds chosen for ML experiments. I bet 42 is over-represented.