The paper explores the implications of rational inattention, where actions can only depend on observations through a communication channel with finite Shannon capacity. This constraint is shown to play a role similar to signal extraction problems or adjustment costs in standard control problems, leading to a theory that resembles familiar dynamic rational expectations theories. The author argues that this approach is useful and practical, with implications for policy that differ from traditional models. The introduction reviews Keynes's idea of price stickiness and subsequent theories that account for deviations from the seamless model of continuously optimizing agents. The paper then introduces the concept of limited information-processing capacity, which is appealing because it accounts for a wide range of observations with a simple mechanism and avoids the need for detailed psychological detail. The paper formally derives the implications of adding information-processing constraints to dynamic programming problems used in macroeconomic modeling. It shows that these constraints alter the behavior implied by models in ways that align with observed macroeconomic behavior. The author discusses how this approach affects the modeling of policy changes and welfare criteria. The paper also explores dynamic optimization problems with information-flow constraints, showing that solutions mimic signal-extraction problems but with endogenous noise. Examples demonstrate that the tendency of information constraints to induce delay is not limited to discontinuous data paths and can be observed in oscillatory data as well. Finally, the paper applies the rational inattention framework to a permanent income model, showing that it reproduces results of standard permanent income theory for a special case. The model predicts smooth and delayed responses to income shocks, which align with empirical observations. The paper concludes by suggesting that the rational inattention approach can provide a basis for critical thinking about the results of other macroeconomic models.The paper explores the implications of rational inattention, where actions can only depend on observations through a communication channel with finite Shannon capacity. This constraint is shown to play a role similar to signal extraction problems or adjustment costs in standard control problems, leading to a theory that resembles familiar dynamic rational expectations theories. The author argues that this approach is useful and practical, with implications for policy that differ from traditional models. The introduction reviews Keynes's idea of price stickiness and subsequent theories that account for deviations from the seamless model of continuously optimizing agents. The paper then introduces the concept of limited information-processing capacity, which is appealing because it accounts for a wide range of observations with a simple mechanism and avoids the need for detailed psychological detail. The paper formally derives the implications of adding information-processing constraints to dynamic programming problems used in macroeconomic modeling. It shows that these constraints alter the behavior implied by models in ways that align with observed macroeconomic behavior. The author discusses how this approach affects the modeling of policy changes and welfare criteria. The paper also explores dynamic optimization problems with information-flow constraints, showing that solutions mimic signal-extraction problems but with endogenous noise. Examples demonstrate that the tendency of information constraints to induce delay is not limited to discontinuous data paths and can be observed in oscillatory data as well. Finally, the paper applies the rational inattention framework to a permanent income model, showing that it reproduces results of standard permanent income theory for a special case. The model predicts smooth and delayed responses to income shocks, which align with empirical observations. The paper concludes by suggesting that the rational inattention approach can provide a basis for critical thinking about the results of other macroeconomic models.