In my linear-algebra and numerics courses, I frequently heard an anecdote about some professor betting – literally, with money – that there would never be any application where computing the actual inverse of a matrix was the best option. (Obviously, some fineprint needs to be added to that bet, but that’s not pertinent here.)
Who was this and what was the actual bet? Or is it just a myth? Given that there is no shortage of myths, I only care about anything coming with a trustworthy reference.