The increasing levels of multi-drug resistance in human pathogenic bacteria are compromising our ability to treat infectious disease. Since antibiotic resistance determinants are readily exchanged between bacteria through lateral gene transfer, there is an increasing interest in investigating reservoirs of antibiotic resistance accessible to pathogens. Due to the high likelihood of contact and genetic exchange with pathogens during disease progression, the human microflora warrants special attention as perhaps the most accessible reservoir of resistance genes. Indeed, numerous previous studies have demonstrated substantial antibiotic resistance in cultured isolates from the human microflora. By applying metagenomic functional selections, we recently demonstrated that the functional repertoire of resistance genes in the human microbiome is much more diverse than suggested using previous culture-dependent methods. We showed that many resistance genes from cultured proteobacteria from human fecal samples are identical to resistance genes harbored by human pathogens, providing strong support for recent genetic exchange of this resistance machinery. In contrast, most of the resistance genes we identified with culture independent metagenomic sampling from the same samples were novel when compared to all known genes in public databases. While this clearly demonstrates that the antibiotic resistance reservoir of the large fraction of the human microbiome recalcitrant to culturing is severely under sampled, it may also suggest that barriers exist to lateral gene transfer between these bacteria and readily cultured human pathogens. If we hope to turn the tide against multidrug resistant infections, we must urgently commit to quantitatively characterizing the resistance reservoirs encoded by our diverse human microbiomes, with a particular focus on routes of exchange of these reservoirs with other microbial communities.