The rationalist community's location problem — LessWrong