While I understand the joke and agree that it's oddly funny and somewhat true, I feel this downplays the significant amount of intellect and mathematical genius that goes into engineering theory and practice. If we were to take a relationship of physical accuracy of our models and graph it against the preciseness of material testing, I suspect you would see an asymptotic relationship between how close the calculations get with reality. Dealing with all materials, we have to use statistical averages due to manufacturing defects and inherent fluctuations of material strengths and stiffness, however, the industry balances around a profitable equilibrium in which engineers expect and design to a reasonably small material property standard deviation that manufacturers can consistently reproduce. Of course this all depends on the rigor of your analysis, while Bernoulli will get you close, Timoshenko will get you closer, and finite element analysis would be the gold standard. Maybe there will be a paper published studying FEM elements with material properties varying over expected statistical deviations throughout a system. I would guess the smaller you made your elements, the more accurately you would reflect reality (again, an asymptotic relationship). Unfortunately, my explanation won't fit on a t-shirt and even if it did, I suppose you could expect comments from the layperson along the lines of the movie "Idiocracy."
I'm reminded of when my wife was in college getting her MBA. We were having a conversation with someone about her career, with a blank stare the individual looked back at us and said "Is an MBA a nurse or something?"