<p>Virtue-first approaches to regulative epistemology describe an ideal that individual inquirers should attempt to approximate, and then specify some of the measures that we can take collectively to produce more virtuous inquirers. These approaches ignore the fact that we are cognitively limited in ways that prevent us from approaching the virtue-theoretic ideal, and that we can transcend our limitations only by working together. Consequently, virtue-first approaches get things backwards: we should focus first on the conditions of collective intellectual flourishing, and then suggest ways that individuals can function as virtuous team members. This is the team-first approach to regulative epistemology. Among the conditions of collective success is an efficient division of epistemic labor that harnesses a diversity of competences, traits, and beliefs. In Sects.&#xa0;<InternalRef RefID="Sec2">1</InternalRef> and <InternalRef RefID="Sec4">2</InternalRef>, I argue that virtue-first approaches reduce these forms of diversity by having individuals attempt to approximate a common epistemic ideal. In Sect.&#xa0;<InternalRef RefID="Sec7">3</InternalRef>, I argue that virtue-first approaches should be de-emphasized in favor of a team-first approach that targets efficient forms of coordination and cooperation through incentive structures, role assignments, feedback mechanisms, and the like. I reinforce this position throughout the paper by making the case that efforts at science reform are more likely to succeed when they take this approach, rather than a virtue-first approach. Finally, in Sect. 4, I contend that epistemic virtues have a role to play in the team-first approach, but that it is a subsidiary role.</p>