Nate Silver discusses the benefits and challenges of using agent based modeling as a computationally intensive method to analyze limited data and non-linear phenomena. He highlights the importance of considering edge cases and the correct functional form of a model when dealing with complex systems.
