In his insightful essay on the potential impact of automation on future mathematical research, Akshay Venkatesh calls attention to the current mechanisms through which mathematicians negotiate the value of results and reach consensus. My talk will expand on Venkatesh’s discussion by focusing on three topics. First, I will illustrate how the valuation of mathematical work has shifted over time and across different subsets of the mathematical community. I will provide historical and contemporary cases, with an emphasis on debates over whether certain kinds of proof are better than others. Second, I will consider the role of so-called “mathematical folklore” (i.e., unpublished results which circulate informally among specialists) in research, as well as the challenge of formally verifying such folklore. Third, I will consider the role of tacit knowledge (i.e., knowledge which is difficult to formalize or codify explicitly) in mathematics, for example in what some researchers might informally call “intuition” or “deep understanding.” Altogether, I will argue that although tacit and informal modes of knowledge have always been constitutive and inextricable parts of mathematics, the partial automation of research continually transforms both their scope and their role.