The tools and concepts of the emerging field of complexity science—like agent-based modeling, network theory, and machine learning—can offer powerful insights to economists and crafters of public policy. Complexity science enables us to explicitly model relationships between individuals and institutions, asymmetric information and influence, the emergence of unplanned emergent social orders, and dynamically adaptive individuals. In the last few decades the tools of complexity science have been applied to the problem of public goods provision, correcting hypothesized behavioral biases, and raising the efficiency of policy implementation. These analyses often lack public choice perspectives, which may complicate and even obviate their findings when the designer becomes entangled with the complex structures in his models. Furthermore, there remains a good deal of work to be done to harmonize traditional public choice work with the tools and insights of complexity science. Uncharted waters must eventually be charted; we hope to begin in such a way that avoids the worst of the dragons.