The increasing availability of laboratory automation tools brings increasing opportunities to explore complex design spaces that would be intractable to with traditional experimental techniques. To take full advantage of these capabilities, however, we must couple them with techniques for experimental design that help us decide *what* to run. Ultimately, the productive combination of hardware and software is what is needed for autonomous discovery. In this talk, I will describe some recent work from the group on the identification of random heteropolymer blends that confer thermal stabilization to enzymes as well as the identification of combinations of building blocks and conditions that lead to new (by at least one definition) chemical reactions.
