I'm with the commenter (here) who said that real life is more likely to be an initial-value-dependent or path-dependent PDE. Someone else said having a high-IV person on your team will raise the slopes of everyone else which seems also right: the issues are multidimensional, not affine 1-D.

Ousterhout is taking sides (smarter > more experienced) which is fine; he can make that argument. But using y=mx+b as x→∞ doesn't count as an argument; it's rhetorical flair, not rhetorical substance. The substance of his reasoning seems to be "That's my opinion based on my experience in my past jobs".