> I think it's not impossibly hard to find metrics we can associate with "overall benefit to humanity" from a quantitative viewpoint. One would hope for the sl

> I think it's not impossibly hard to find metrics we can associate with "overall benefit to humanity" from a quantitative viewpoint. One would hope for the slightly higher ambition of metrics that are causally connected to "overall benefit of humanity", settling for merely correlated has that unpleasant risk of the association breaking as you try to optimise.