Yeah, I wasn't sure how to address this one. In addition to most software not existing inside a pure abstract context, things like using 3 random points to det

Yeah, I wasn't sure how to address this one. In addition to most software not existing inside a pure abstract context, things like using 3 random points to determine the quick sort pivot just sat there undiscovered for a few decades. Since hypothesis-generation is the big undefined area, it would seem that diverse backgrounds would necessarily help. Plus we're implicitly assuming that diverse backgrounds lead to sub-par work in the pure theory domain, which is kinda shitty on its own. And I think the parallels between a solid understanding of police "kettling" as they would apply to distributed systems and/or queueing are fairly clear. But I wouldn't expect someone who thinks inverting a matrix is tricky to appreciate either.