It's more like: don't invest $100B in capital when there's still orders of magnitude improvements in software to be made.