Social computing applications are transforming the way we make new social ties, work, learn and play, thus becoming an essential part our social fabric. As a result, people and systems routinely make inferences about people's personal information based on their disclosed personal information. Despite the significance of this phenomenon the opportunity to make social inferences about users and how this process can be managed is poorly understood. In this paper we 1) outline why social inferences are important to study in the context of social computing applications, 2) how we can model, understand and predict social inference opportunities 3) highlight the need for social inference management systems, and 4) discuss the design space and associated research challenges. Collectively, this paper provides the first systematic overview for social inference research in the area of social computing.