> Not to mention social events on any subject when talking to professionals from other domains triggers solution to a long lasting problem.

This is becoming more and more important as science/technology keeps getting compartmentalized. Not to go as far as the medical doctor who rediscovered numerical integration [1], but I have the certainty that in my research I'm doing many things in a horribly inefficient way, out of ignorance. Just the other day I learned about k-d trees and sped up a piece of software in my lab by an order of magnitude...

[1] https://news.ycombinator.com/item?id=1964613