Lay theories suggest that people who are overconfident in their knowledge are less likely to revise that knowledge when someone else offers an alternative belief. Similarly, one might assume that people who are willing to revise their beliefs might not be very confident in their knowledge to begin with. Two studies with children ages 4–11 years old and college students call these lay theories into question. We found that young children were simultaneously more overconfident in their knowledge (e.g., believing they knew what chartreuse meant) and more likely to revise their initial beliefs (e.g., choosing another color after seeing a peer choose a different color) than older children and adults. These results bridge the metacognitive and epistemic trust literatures, which have largely progressed independently from each other. We discuss the potential causes and functions of the dissociation between the confidence with which beliefs are held and the revision of those beliefs across development.