That's a ridiculously high bar to set before you recognize any progress being made.
I probably meant it figuratively, huh? What I am saying after having keynoted at a Microsoft formal methods conference (as a critic) and after playing with TLS+ (which was educational) is that formal verification will make little impact on commercial systems because such systems are obliged to be built with inherently buggy components.
I am happy to acknowledge that progress is happening, yet the ultimate source of trouble is human ambition and impatience. How do you solve that?