Would current use of QALYs factor in some of the features you mentioned? I suppose the focus on lives saved is a criticism of the post not of GiveWells method in general?