"Likewise, if you can't know _when_ the scientists are wrong with >90% accuracy, you should just shut up and trust the scientists _always_. But, of course, this is not what people do."

The problem is that it's trivially easy to find MAJOR errors in the methodology or interpretation of many or even most journal articles, and whatever passes for the scientific consensus is usually based directly off a very shallow reading of the abstracts and conclusions of these articles without taking any of their flaws into account, let alone questioning the epistemology behind them. There are probably fields where this isn't true, but they're definitely the minority. When the scientific consensus changes it's rarely because of brilliant new research, but rather because the flaws in old research gradually get publicized and so the old conclusions get replaced by the conclusions of new research that's also flawed, but where the flaws aren't yet widely known. But there's no reason why you can't actually figure out what these flaws are, it's not difficult at all, it's just that most people (including scientists) are too lazy to read the actual papers and think critically about them.

The point is that people who buy into scientific consensus wholesale aren't actually the rational skeptics they market themselves as being, but rather are just as dumb and ignorant as the worst of the religious fundamentalists.