Should radical uncertainty about the distant future dissuade us from judging options by referring to their consequences? I argue no. Some short-run benefits are sufficiently high that we should pursue them, even if our long-run estimates possess a very high variance. I discuss the relationship between the epistemic argument and ‘fuzzy’ rankings and also ‘arguments from infinity’. Furthermore, extant versions of the epistemic argument require the assumption that we have no idea about the major consequences of our acts. Even a slight idea about some major consequences will render the epistemic argument less plausible. In most applications of the epistemic argument, long-run uncertainty is not the relevant confounding variable; on close examination the epistemic argument tends to trade on other principles altogether.