As next generation sequencing has lowered the price of DNA sequencing orders of magnitude below what it cost to create the first consensus human genome, the sheer diversity of human sequence between individuals has become more and more apparent. This diversity is not only seen in single nucleotide variants but also deletions/duplications, insertions/deletions, and copy number variants. Sorting the wheat from the chaff when performing molecular diagnostics of single-gene (Mendelian) disorders such as Hereditary Breast and Ovarian Cancer, Hypertrophic Cardiomyopathy, Hereditary Spastic Paraplegia, Cystic Fibrosis, is an enormous challenge because very serious medical decision making rests on a correct interpretation of the clinical meaning of variants, both in affected individuals and when sequencing is used for screening purposes. The American College of Medical Genetics and the Association of Molecukar Pathologists took a first major step when they proposed that variant interpretations fall into one of five classes: Benign, Likely Benign, Likely Pathogenic, Pathogenic, and Variant of Uncertain Significance (VUS). The also proposed a set of criteria to apply to guide the placement of any variant into on of these buckets. This proposal has improved the rigor and reproducibility of variant interpretation but much remains to be done to make variant interpretation more accurate and scalable if we are to deploy genomics more widely in medicine.