In many strategic environments, information acquisition is a central component of the game that is played. Being uncertain about a payoff-relevant state, a player in a game has a two-fold incentive to acquire information: learning the state and learning what others know. We develop a model of information acquisition in games that accounts for players' incentive to learn what others know. In applications to rational inattention and global games, we prove the power of this incentive. When information acquisition is ``independent''---that is, players can acquire information only about the state---severe coordination problems emerge among rationally inattentive players. When information acquisition is ``unrestricted''---that is, players can acquire information about the state and each other's information in a flexible way---we show that rational inattention admits a sharp logit characterization, and we provide a new rationale. for selecting risk dominant equilibria in coordination games.