Incomplete thoughts on cognitive diversity

I'm currently reading about the ways in which cognitive diversity leads to better problem solving, and methods through which we can make this effect larger.

One of the results that I've found more interesting is by Kevin J. S. Zollman [1]: limiting information can lead to more accurate results. "Bad Information" tends to spread less in networks with more restricted connections, while networks that are highly connected tend to be much faster.

I can the slow-but-accurate model is preferable in long-term foundational projects or strategic decisionmaking, while the highly connected one works when we want to address experiments quickly (even converging multiple times and testing different hypotheses).

I'd like to try some of these networks and see how well it works in practice.

[1] Zollman (2010)