Not that I have specific issues with the recommendations there, but from a conceptual standpoint, why should I trust recommendations on every subjects from a centralized non-specialized site when I can search for sites specialized in each subjects and take their recommendations on their own subjects instead.
To me, this article strongly recommends reading textbooks and softly recommends this list. They even go into detail about what methods should be used to validate textbooks generally!
If you input the term "discrete mathematics" in the search field "Books" on Amazon, you get about 15 books (not counting the advertised ones) per page and 101 pages of those. If you slightly change criteria (change the search field by subject or tweak the date or add another term to "discrete mathematics" or ...), you get even more hits (many more). This is essentially a bottomless well. Most books found in this way (many of them incredibly good) will never be found in the lists like that offered by OP for many reasons: lists like that of OP favor 'classics", every month there's a new book on the market that tries to outdo the existing ones, etc Many writers of the newer books are aware of the existence of the classics (in fact, many had to learn from them), so in their books they point out the pitfalls, fill in the holes, add more details and make your learning experience all around more pleasant.
The books in the OP's list might not work for everyone, so you have to do your own searching, sorting and analysis. This is especially true if you are a self-learner and much more so if the subject you want to learn is math because in math you are the one who has to prove every one of your assertions and be as independent as possible.
tl; dr:
don't worry about the "expert opinion". Instead collect data, clean it, analyze it and the consume the parts you need. That way you avoid 'expertise' that comes bundled with traditionalisms, bureaucracy, calcified thinking and other undesirable bullshit.
I share your doubts. For instance, I have a URL saved on my home computer where the (admittedly non-universal) HN consensus was that the CLRS algorithms book recommended here is not a great one to learn from.
However, I think your faith that it's easy to find good recommendations by subject matter experts underestimates the difficulty of finding them in the mass of things on the internet. To take one example, I remember reading a post about two approaches to denotational semantics, along with recommendations about the best textbooks, but have never subsequently been able to find that link.
Because you've spent some time on the site and think that the people on it are likely to come up with good recommendations. And if you haven't, maybe you shouldn't.
