The autism spectrum disorders (ASD) are a heterogeneous set of neurodevelopmental syndromes defined by impairments in verbal and non-verbal communication, restricted social interaction, and the presence of stereotyped patterns of behavior. The prevalence of ASD is rising, and the diagnostic criteria and clinical perspectives on the disorder continue to evolve in parallel. Although the majority of individuals with ASD will not have an identifiable genetic cause, almost 25% of cases have identifiable causative DNA variants. The rapidly improving ability to identify genetic mutations because of advances in next generation sequencing, coupled with previous epidemiological studies demonstrating high heritability of ASD, have led to many recent attempts to identify causative genetic mutations underlying the ASD phenotype. However, although hundreds of mutations have been identified to date, they are either rare variants affecting only a handful of ASD patients, or are common variants in the general population conferring only a small risk for ASD. Furthermore, the genes implicated thus far are heterogeneous in their structure and function, hampering attempts to understand shared molecular mechanisms among all ASD patients; an understanding that is crucial for the development of targeted diagnostics and therapies. However, new work is beginning to suggest that the heterogeneous set of genes implicated in ASD may ultimately converge on a few common pathways. In this review, we discuss the parallel evolution of our diagnostic and genetic understanding of autism spectrum disorders, and highlight recent attempts to infer common biology underlying this complicated syndrome.