Social machines have long been a topic of theoretical and empirical interest for the Web and Internet science community. Although there is no widespread agreement as to the precise meaning of the term “social machine,” the term has served as something of a conceptual anchor for research efforts into a number of online systems, such as Facebook, Wikipedia, and Galaxy Zoo. Recently, the term “social machine” has surfaced in a somewhat different disciplinary context. In particular, it has been applied to systems that are commonly referred to as cyber-physical systems. Here, the residual ambiguity surrounding the meaning of the term “social machine” has stymied efforts to understand the nature of the relationship between social machines and cyber-physical systems. This situation is further complicated by the absence of a precise conceptual understanding of cyber-physical systems. The present paper outlines an approach to understanding both social machines and cyber-physical systems that is intended to resolve this impasse. Inspired by recent work in the philosophy of science, we present a mechanistic (or neo-mechanical) account of social machines that draws attention to the features of socio-technical mechanisms. Such an account, we suggest, is able to accommodate the intuitions of Web scientists regarding the nature of social machines, while simultaneously illuminating the nature of the relationship between social machines and cyber-physical systems. As an added bonus, the account speaks to the interdisciplinary concerns of the Web science community by highlighting a link between the science of social machines and mechanism-based approaches in the social sciences.