The social world is inherently uncertain. We present a computational framework for thinking about how increasingly popular online environments modulate the social uncertainty we experience, depending on the type of social inferences we make. This framework draws on Bayesian inference, which involves combining multiple informational sources to update our beliefs.