This is a reply to Chow, Halperin, and Mazlish’s paper which argued that we can infer that AGI isn’t coming, because real interest rates haven’t risen. Implicit in that paper is an assumption that the marginal utility of a dollar of consumption will fall. We get more and more things, and care less about each additional thing. This need not hold if there are new goods, however. We could develop capabilities which are not available now at any price. This also implies that the right way to hedge your risks with regard to AI depends on precise predictions about AI’s capabilities.