Which song will Smith listen to next? Which restaurant will Alice go to tomorrow? Which product will John click next? These applications have in common the prediction of user trajectories that are in a constant state of flux but subject to hidden constraints (e.g. geographical location, the links of a website). What users are doing now may be unrelated to what they will be doing in an hour from now. In this talk I introduce the difficulties associated with predicting user trajectories, more specifically how the concepts of non-stationary, transiency, and time-heterogeneity make this task challenging. Mindful of these difficulties I introduce Tribeflow, a general method that can perform next product recommendation, next song recommendation, next location prediction, and general arbitrary-length user trajectory prediction without domain-specific knowledge. Extensive simulations on large and small datasets show TribeFlow to be more accurate and up to 413x faster than top state-of-the-art competitors.