> The prevailing framing—in the essay and in the market—makes all economic disruption from AI look like a capability story: AI gets smarter, therefore jobs disappear, therefore the companies that serve those jobs lose revenue. The more accurate framing is: AI gets smarter, but institutional language games have their own logic, and the speed of disruption depends on whether AI can enter those games or must wait for organizations to rebuild them. I believe that the answer, for most enterprise software, is the latter—and rebuilding institutional games is a process measured in years and decades, not quarters.