Our current choices regarding the fight against deforestation and climatic disorder, against the overexploitation of natural resources or the erosion of ecologic heritage, have effects which will last long after we disappear. Must we make immediate and major sacrifices for the well-being of our descendants? Speaking only about developed countries, they will anyway be better off than we are if the trend in growth observed for the last two centuries continues. Unless, as scientists and GIEC Experts who work on biodiversity and ecosystem services fear, we are approaching the natural limits which commit here and now our responsibility towards future generations. In a context where current choices have uneven distributed effects in time, how do we arbitrate between the present and future, between the interests of the various generations? One way of considering the problem is to incorporate some normative requirements into a criterion of inter-temporal social choice, and estimate the legitimacy of alternative futures according to their classification by this criterion. This approach results in a general message as important as it is frustrating: looking for the trajectories of management which avoid waste cannot usually be achieved without favouring certain generations. Therefore, the concern for our descendants can be summed up to one question: what are the desirable compromises between efficiency and impartiality? Discounting the utilities of every generation before adding them together is a possible answer. But such a criterion, for a long time applied by default, presents an all the more contestable discrimination against future generations, all the more so as alternatives exist. This note briefly presents the works of the scientists of the SAE2 department on two alternatives to discounting. Chichilnisky’s criterion and the mixed Bentham-Rawls (MBR) criterion are two possible answers to the efficiency-impartiality dilemma.