The recent modeling of infectious disease spread for COVID-19 has been quite successful, and has received an unprecedented media prominence. However, the vast majority of these models are premised on a highly unrealistic assumption, namely that contacts between individuals that can spread the infection are entirely random. In contrast, there is widespread evidence that contact patterns have important structure, such that some individuals have many more contacts than others. It has been known for decades in the network theory literature that such contact patterns, characterized by “heavy tailed” skewed “degree distributions”, give rise to qualitatively different epidemic curves than the exponential curves emerging from the standard differential equation models which typically embody an assumption of random mixing. In this talk, we review some of the relevant literature with two basic objectives: (1) How might an appreciation of ubiquitous real world non-random mixing affect projections of the epidemic curves of newly emerging infectious diseases and in particular to what extent might they diverge from the usual resulting exponential growth; and (2) how might an appreciation of the reality of non-random mixing affect the kinds of data collection systems needed for future epidemic preparedness?