Interesting short post on the significance of confirmative outliers in validating our preconceptions. We
want to believe what we believe, are sensitive to potential sources of validation and take singular examples as representative of the whole. More evidence, if you needed any, that we need more robots making decisions, not less.