Thinking of the world in mechanistic terms—how things work—is both cognitively natural and motivating for humans from the preschool years onward. Mechanisms have distinct structural properties that go far beyond mere causal facts. They typically contain layers of causal clusters and the systematic interactions between those clusters that give rise to the next level up. Following developments in the philosophy of science and studies on children’s questioning behaviors, recent research shows that, from an early age, people appreciate the informational and inductive potential of mechanistic information. People selectively notice and choose mechanistic explanations as especially useful opportunities for learning; but they also soon forget the details of what they encounter. We argue that enduring cognitive abstractions from such details provide powerful ways of accessing and evaluating expertise in other people.