I’ve always thought ‘absence of evidence is not evidence of absence’ is flawed for material truths because of the second law: presence of X, X being at different entropy from background, would necessarily produce evidence. No fire without smoke basically https://en.m.wikipedia.org/wiki/Evidence_of_absence …
Wait that example may be backwards 
I’ll think of a better one when I’m in a smarter mood... general point being I think I have a non-Bayesian reason to be skeptical of absence of evidence principle