The theme of this meeting is Modeling Coupled Earth and Human Systems. Since the World3 model and the Limits to Growth report of 1972 there has been a sustained effort of integrated modeling of human activities and the Earth system. Despite the existence of integrated models, there is an increasing recognition that the social science is largely lacking from the modeling efforts. Having worked in both natural science and social science departments, I reflect on the different modeling cultures and the challenges in social science to use simulation models. Building on the work of the CoMSES Net I also provide some promising examples of agent-based models advancing social science.