Computers cannot yet reliably reproduce the behavior of an individual human – and yet computational models are becoming increasingly good at modeling the complex dynamics of human crowds. In this talk, I will give an overview of the crucial role that computational modeling is beginning to play in broadening our understanding of collective human behavior – from financial markets through to online games and real-world wars. Beyond their simple role as a simulator, computational models – fueled by increasingly precise spatiotemporal empirical data – look set not only to increase our understanding of complex socio-economic systems, but also to help us move toward managing and eventually predicting their evolution.