Insights from Dath Ilan — LessWrong