GiveWell: A case study in effective altruism, part 1 — LessWrong