This is a well written piece, and I have to say, I agree with a lot of what you propose.
“Metrics may be derived to prove most any biased claim in a manner sufficiently compelling to sell a product yet prohibitively expensive to disprove”- This I believe has always been true in a sense, and I believe is the reason why computational methods have always been popular or sought after.
“Increasingly granular analyses and computations may allow increasingly granular findings, but their inherent complexity allows for the the masking of errors and bias” – Well said. You state this in the context of computational work, but I believe this is very much true in the experimental sciences as well.
Overall, I fail to understand your example of the pot in pot refrigerator in relation to the issues you raise, and I would be glad to discuss why this particular example is representative of what you describe as “the challenge in communication”.