Keynote Lecture at the 2014 BayesiaLab User Conference

The development of Bayesian Networks, so people tell me, marked a turning point in the way uncertainty is handled in computer systems. For me, this development was a stepping stone towards a more profound transition, from reasoning about beliefs to reasoning about causal and counterfactual relationships. In this talk, I will survey the milestones of this journey, and  summarize the practical and conceptual problems that we can solve today and could not address two decades ago.