Organisms have to predict the future to best choose actions.  How do they do it, especially given the resource constraints that govern their ability to process information?  We evaluate the ability of machine learners, bio-inspired neural networks, neurons, and humans to predict and memorize, with surprising preliminary findings.