The supposedly unreasonable effectiveness of mathematics is just a reflection of the fact that systems near equilibrium are well approximated with only first and second derivatives.
-
I have this idea that most (all?) subfields of industry/math/science benefit significantly from "linearization", where linear approximations enable some breakthrough: CAPM, bundle adjustment, Markov chains, inverse kinematics, etc. If your field doesn't have one yet, look for it.
Yes, but those are old hat. The problem now is that the fixation on linear models blinds researchers to nonlinear phenomena, which are almost everything.
