To create government that is neither bigger nor smaller but better at solving problems more effectively and legitimately, agencies need to use big data and the associated technologies of machine learning and predictive analytics. Such data-analytical approaches will help agencies understand the problems they are addressing more empirically and devise more responsive policies and services. Such data-processing tools can also be used to make citizen engagement more efficient, helping agencies to make sense of large quantities of information and invite meaningful participation from more diverse audiences who have never participated in our democracy. To take advantage of the power of new technologies for governing, however, the federal government needs, first and foremost, to invest in training public servants to work differently and prepare them for the future of work in a new technological age.