In their new book, Religion Evolving: Cultural, Cognitive, and Ecological Dynamics, B. G. Purzycki and R. Sosis (2022) argue that religions should be viewed as adaptive systems not just by-products of cognitive capacities that evolved for other purposes. In this essay, I argue that they are correct that the types of religious systems that are widespread across cultures and eras (i.e., “world religions”) have been selected for because they help individuals to solve adaptive problems. However, I disagree with their claim that this fact proves the by-product camp wrong. Rather, I argue, religious systems that are successful are both cognitively catchy and provide adaptive benefits. It is not either/or, it is both/and. Further, I argue that to be truly adaptive, religious systems must accomplish the ultimate evolutionary goal of genetic replication by sexual reproduction and not just the proximate goal of pro-sociality or group cohesion.