On Superintelligence: Paths, Dangers, and Strategies (2014) with the author. What can we predict about, and how can we control in advance, the motivations of the entity likely to result from eventual advances in machine learning? Also with guest Luke Muehlhauser.