Microorganisms are Nature's little engineers of a remarkable array of bioactive small molecules that represent most of our new drugs. The wealth of genomic and metagenomic sequence data generated in the last decade has shown that the majority of novel biosynthetic gene clusters (BGCs) is identified from cultivation-independent studies, which has led to a strong expansion of the number of microbial taxa known to harbour BGCs. The large size and repeat sequences of BGCs remain a bioinformatic challenge, but newly developed software tools have been created to overcome these issues and are paramount to identify and select the most promising BGCs for further research and exploitation. Although heterologous expression of BGCs has been the greatest challenge until now, a growing number of polyketide synthase (PKS) and non-ribosomal peptide synthetase (NRPS)-encoding gene clusters have been cloned and expressed in bacteria and fungi based on techniques that mostly rely on homologous recombination. Finally, combining ecological insights with state-of-the-art computation and molecular methodologies will allow for further comprehension and exploitation of microbial specialized metabolites.