All of the arguments about the presumed reliability of science are ridiculous and easily shown to be false. Science is no more “self-correcting” than accounting. Peer review is more commonly known as “proofreading” by the rest of the publishing industry and is not even theoretically a means of ensuring accuracy or correctness. And scientists are observably less trustworthy than nearly anyone except lawyers, politicians, and used car salesmen; at least prostitutes are honest about their pursuit of “grants” and “funding.”
These days, the scientific process is mainly honored in the breach by professional, credentialed scientists. And we have a word for testable, reliable science. That word is “engineering.”