Found myself wondering about an interesting philosophical problem with some of the assumptions of machine learning, which led me to an interesting philosophy stackexchange thread: https://philosophy.stackexchange.com/questions/40581/on-wittgensteins-family-resemblance-and-machine-learning