In recent years, many physicists from different fields, but mostly from statistical physics, have devoted a lot of effort to the study of complex socio-technological systems. Why is this the case? Does physics have a say on such systems? In this talk I will discuss these and other issues starting from my research on the problem of cooperation. I will quickly introduce the topic and the background, and the describe several experiments that shed light on how people facing a social dilemma interact, having in mind that interaction is strategic, i.e., individuals anticipate their partner's actions. Subsequently I will show how on the basis of the experimental results one can build stylized models of human behavior that are in good agreement with the observations. In conclusion, what I hope to convey is that the paradigms and methods of physics can be very useful to bring new perspectives into the social sciences.