The sub-seasonal to seasonal prediction project (S2S) is a 5-year project, established in 2013 by the World Weather Research Program (WWRP) and the World Climate Research Program (WCRP). This paper briefly describes the S2S project in the context of extended range prediction of extreme events. We provide evidence that S2S forecasts have the potential to predict the onset, evolution and decay of some large-scale extreme events several weeks ahead. For instance, S2S models displayed skill to predict high probabilities of extreme 2-m temperature anomalies over Russia during the worst week of the 2010 Russian heat wave up to 3 weeks in advance. In other cases, like for tropical cyclone prediction, S2S models can produce useful information on the probability of the occurrence of tropical storms within sufficiently large areas through the prediction of large-scale predictors, such as the Madden-Julian Oscillation (MJO). In future, S2S forecasts of extreme events could be integrated into a "ready-set-go" framework between seasonal and medium range forecasts, by providing an early warning of an extreme event a few weeks in advance. Finally, S2S forecasts can also be used to investigate the causality of some extreme events and we show evidence that the cold March 2013 over western Europe and North Asia was linked to a MJO event propagating over the western Pacific.