The unwillingness to make models that produce predictions with error bars 12 OOMs wide is one of the worst things about social science
a Schelling point for those who seek one