For a period in the mid-1990s, soon after the discovery of the involvement of trinucleotide repeat expansions in fragile-X syndrome (both A and E), Huntington’s disease, myotonic dystrophy, and a number of hereditary ataxias, there was a clear sense that this new disease mechanism might provide answers for psychiatric disorders. Given the then failures to replicate initial genetic linkage findings for schizophrenia (SCZ) and bipolar disorder (BD), a greater emphasis was placed on the role of complex and non-Mendelian mechanisms, and repeat instability appeared to have the potential to provide adequate explanations for numerous apparently non-Mendelian features such as anticipation, incomplete penetrance, sporadic occurrence, and nonconcordance of monozygotic twins. Initial molecular studies using a ligation-based amplification method (repeat expansion detection) appeared to support the involvement of CAG•CTG repeat expansion in SCZ and BD. However, subsequent studies that dissected the large repeats responsible for much of the positive signal showed that there were three main loci where CAG•CTG repeat expansion was occurring (on 13q21.33, 17q21.33-q22, and 18q21.2). None of the expansions at these loci appeared to segregate with SCZ or BD, and research into repeat expansions in psychiatric illness petered out in the early 2000s. The 13q expansion occurs within a noncoding RNA and appears to be associated with spinocerebellar ataxia 8 (SCA8), but with a still unexplained dichotomy in penetrance – either very high or very low. The 17q expansion occurs within an intron of the carbonic anhydrase-like gene, CA10. The 18q expansion is located within an intron of the TCF4 gene. Mutations in TCF4 are a known cause of Pitt-Hopkins syndrome. Also, pertinently, genome-wide association studies have shown a well-replicated association between TCF4 and SCZ. Two decades on, in 2016, it appears to be an appropriate juncture to reflect on what we have learned, and, with the arrival of newer technologies, whether there is any mileage to be made in revisiting the unstable DNA hypothesis for psychiatric illness.