I've not seen any evidence that this industry is more or less meritocratic (in the positive sense meant here, that is, based on ability to perform programming tasks) than any other--including academia (whence I come, but from a science background, not CS/math).

On the other hand I've seen plenty of evidence that in certain geographic regions there is a particularly heavy (and, in my opinion, amusingly misplaced) emphasis on categorical errors, equivocation, and conflation of purely academic CS knowledge with practical ability (as you yourself allude to). These are all the CS-world equivalent of the supposedly very un-meritocratic things many who've never been outside of CS (professionally) assume occur in, oh, "banking, consulting, and [the] business world" as a matter of daily routine.

Agree, but then again, isn't there some psych theory that explains that (figures, since the PhDs probably formulated them) -- that once you get so fixated/advanced on a certain object, you lose the ability to do the super basic stuff (or are at least bad at it since you haven't done it in so long?) -- but maybe that's just apologism.

And definitely, tech is arguably the most meritocratic high-paying field out there. Just the ability to build something on your own with your skills without anyone's permissions is not common (i.e. you can't just open a bank).