Teams and Coaches
Harness the power of outlier individuals
One way to design an outlier organization is to focus on attracting and cultivating outlier individuals – humanity has a meme of great scientists being outliers that goes back at least as far as Aristophanes’ 423 BC depiction of Socrates in the play The Clouds. Sports is another domain driven by outlier individuals: the best athletes are almost always extreme – tall or short, broad or thin, rarely in between. Driven by incredible competitive pressure, sports has honed the “team and coach” model as a way to harness the power of outlier individuals. Science could do the same.
In sports, teams work together over multiple years, supervised by a coach who gives detailed feedback with an outside perspective. This structure couples with competition to create steady improvement; across professional sports, today’s worst teams outclass the best teams from 30 years ago.
Today, scientific collaborations are temporary, ad-hoc, and part-time unless they are focused on a single big problem. Instead, you could build teams of individuals who move together from problem to problem and engage in regular group practice with their coach. These teams would learn to work as an elite unit, not just as elite individuals (all-star games are rarely peak performances). This structure requires fair credit sharing, confidence that the team will be funded across multiple projects, and broad recognition of the value of deliberate practice. This is not what Kelly Johnson claimed was important about Lockheed’s famous Skunk Works, but the Skunk Works appears to have come close to having this structure and these properties. We are not aware of any contemporary examples of this approach, but they might be hidden inside corporate R&D labs.
> confidence that the team will be funded across multiple projects
I totally subscribe to this vision. The point highlighted above is in my opinion the main barrier to such teams, at least in traditional academia. My impression is that you have such structures in French academia though, but possibly underfunded.
I also think it is valuable for teams to be more ad-hoc in science. What would science lose if teams stuck together? What questions maybe couldn't be answered that way? How would you build a team like this? The science of team science has a lot to say about building teams.