In the 1960s, if you wanted to understand the future of computing, you studied IBM. Mainframes. Centralized infrastructure. Enormous capital requirements. A small number of institutions controlling access to computation on behalf of everyone else. The assumption was simple: scale implies concentration, but then ...