It would be especially nice if we could build a collection of canonical explanations of recurring points, sort of like the sequences. People have been talking about a LW for existential risks for a long time; I think that topic fits naturally with transhumanism, the singularity, and predicting the future in general.