I think this is why it is necessary to fund "searching for problems" research. We don't need to solve the 1970s vision of computing, we need to solve the 2020s vision of computing, whatever that may be.