This lecture explores Bayesian updating—how rational agents should revise beliefs (e.g., assessing a biased coin’s probability after repeated heads). It contrasts this with confirmation bias, where people overvalue evidence supporting preexisting beliefs (e.g., political views or stereotypes).