I address from the perspective of construal level theory two fundamental questions about human learning: How broadly we generalize what we learn and how large is the scope of experience that we choose to learn from. A series of studies on predictive learning suggests that introducing longer temporal distance and lower probability between the predictor and the outcome, as well as between the learning experience and its anticipated application increases generalization breadth. I also show some evidence suggesting that people extract broader, more general lessons from narratives that are more spatio-temporally distal and hypothetical. A series of studies on decision making suggests that people seek to obtain more and broader experience when they face a more abstract question, as well as when they make decisions that are more distal in time and when they advise others rather than make decisions for themselves. I discuss these lines of research within the broader context of questions about how people regulate themselves with respect to distal and proximal targets.