The main problem as I see it is that we like to fall back on intuitive ways of knowing despite having an increasingly large number of areas in which those intuitive ways are unreliable. Scientism is a symptom of this where we substitute "trust science" for "do science". "Trust God", "Trust me,", "Does this cold winter feel like climate change to you?", and so on, are all unreliable guides to how things are.

If there were more pressure from society for science to be sciency (i.e. use good epistemological techniques) rather than trusted--which means more open data, more discussion, more evidence--then science would be better and human knowledge would advance faster and we would more clearly understand when we ought to be using some other framework to gain understanding (e.g. direct perception, historical custom, etc.) because it would be plain that we simply cannot, as a practical matter, get our epistemological ducks in a row here. You're not going to decide whether to buy a puppy by buying multiple puppies, and not, and surveying your kids' happiness (or measuring their oxytocin levels), and then weighing that against the expected happiness (or oxytocin) from the default outcome from using that much time and money. But if you say, "We're gonna buy a puppy cause science says puppies improve emotional well-being," that could be a problem if everyone involved has personalities that don't mesh well with the typical behavior of dogs. Making that sort of error is less likely when people embrace science more.

Nobody says, "I don't trust this math because I believe in science!" As an epistemological matter the two are distinct ways of knowing, but on a personal level they're both "use your head".

But people do say, in effect, "My tribal wisdom is as good as your science because they're just different ways of knowing." And they say, "There is no scientific evidence that X does Y so you shouldn't do it" as if that's final even when nobody bothered testing it and there are other reasons to suspect it might work.

We really don't need any help to accept authority, to act based on direct perception and hunches, to accept long-standing custom, to follow straightforward logic, and so on. We've got that down. We do need help approaching problems scientifically when it is practical and efficacious to do so, and the biggest problem by scientism by far is not that it gets in the way of other ways of knowing but that it gets in the way of proper scientific epistemology and turns "science" back into the old ways of knowing with all the limitations they have.

So while I agree point-by-point about e.g. flaws in Mizrahiis's reasoning, I still think the sociology should be "Yay, science!!! It's the best! (If you can use it!)" not "Oh, there are lots of ways to know things; science is not that special, and that's why you shouldn't worship it."

(Admittedly, I have come to this conclusion only moderately informed by science, and would feel much more confident if we had more targeted psychological studies that would indicate to what extent my suppositions are accurate.)