The origin of novel complex traits constitutes a central yet largely unresolved challenge in evolutionary biology. Intriguingly, many of the most promising breakthroughs in understanding the genesis of evolutionary novelty in recent years have occurred not in evolutionary biology itself, but through the comparative study of development and, more recently, the interface of developmental biology and ecology. Here, I discuss how these insights are changing our understanding of what matters in the origin of novel, complex traits in ontogeny and evolution. Specifically, my essay has two major objectives. First, I discuss how the nature of developmental systems biases the production of phenotypic variation in the face of novel or stressful environments toward functional, integrated and, possibly, adaptive variants. This, in turn, allows the production of novel phenotypes to precede (rather than follow) changes in genotype and allows developmental processes that are the product of past evolution to shape evolutionary change that has yet to occur. Second, I explore how this nature of developmental systems has itself evolved over time, increasing the repertoire of ontogenies to pursue a wider range of objectives across an expanding range of conditions, thereby creating an increasingly extensive affordance landscape in development and developmental evolution. Developmental systems and their evolution can thus be viewed as dynamic processes that modify their own means across ontogeny and phylogeny. The study of these dynamics necessitates more than the strict reductionist approach that currently dominates the fields of developmental and evolutionary developmental biology.