SummaryBot comments on Sensitive assumptions in longtermist modeling