Users' computing behavior is very complex, and capturing it fully is a true Grand Challenge. Yet it seems obvious that if we can predict that behavior more accurately, we could design more energy-efficient systems. We must leverage behavioral information in both directions. On one hand, we need to understand how people interact with computers and design software and hardware that make interaction efficient without wasting energy. On the other, we need to motivate changes in human behavior to make more efficient use of current and future systems.