From the abstract for Neil Thompson & Svenja Spanuth, The Decline of Computers As a General Purpose Technology: Why Deep Learning and the End of Moore’s Law are Fragmenting Computing (Dec. 2018):
It is a triumph of technology and of economics that our computer chips are so universal. Countless applications are only possible because of the staggering variety of calculations that modern chips can compute. But, this was not always the case. Computers used to be specialized, doing only narrow sets of calculations. Their rise as a ‘general purpose technology (GPT)’ only happened because of the technical breakthroughs by computer scientists like von Neumann and Turing, and the mutually-reinforcing economic cycle of general purpose technologies, where product improvement and market growth fuel each other.
This paper argues that technological and economic forces are now pushing computing in the opposite direction, making computer processors less general purpose and more specialized. This process has already begun, driven by the slow down in Moore’s Law and the algorithmic success of Deep Learning. This trend towards specialization threatens to fragment computing into ‘fast lane’ applications that get powerful customized chips and ‘slow lane’ applications that get stuck using general purpose chips whose progress fades.
The rise of general purpose computer chips has been remarkable. So, too, could be their fall. This paper outlines the forces already starting to fragment this general purpose technology.