I don't think this is something that can be crowdsourced. The arguments here rest on authority (and in this case, it's well justified). You want the opinion of someone who has surveyed the field and has studied further than the respective textbooks take you. There may be some authorities whose replies are worth listening to, but these are difficult to distinguish.
For instance, a user recommending CLRS over TAOCP is an odd piece of advice, since TAOCP is distinctly not a textbook. The user purports to have read TAOCP, an assertion with a low prior-probability.
Unfortunately, I do not think we can rely on the advice of anonymous internet contributors. With no way to enforce the rules, LessWrong fails to do due diligence and is merely spreading the opinions of those that reply.
So, to remedy that, I would add a 4th rule: You must state where your authority on the subject, and on the choice of textbooks comes from.
I agree with your sentiment that recommending textbooks to a beginner is a very hard problem and crowd-sourcing may not be the best idea, but I do have questions for the example you provide to prove your assertion and some of the claims you make.
First of all, why would'nt you consider TAOCP a text book on algorithms? Are you trying to discredit discredit the recommender by supplying your own biases? If TAOCP is not a textbook and CLRS is and you provide no arguments against CLRS, why would you come to the conclusion that the contribution is poor? "LessWrong" has it's faults but the disclaimers that it provides is pretty clear and wouldn't you consider it worse if it interjected with it's own opinions?
If we need to listen to an authority "who has surveyed the field", would Knuth be such an authority and his recommendations contained in TAOCP be something you would recommend?
I haven't read TAOCP; in what sense is it not a textbook? If I wanted to buy a book to learn about algorithms, TAOCP is the first that would come to mind, so saying "you should buy this book instead of TAOCP" is potentially useful advice even if TAOCP is not technically a textbook.
Are any of the criticisms of TAOCP wrong? ("Knuth's TAOCP is wonderful but: very, very long; now rather outdated in the range of algorithms it covers; describes algorithms with wordy descriptions, flowcharts, and assembly language for a computer of Knuth's own invention. When you need Knuth, you really need Knuth, but mostly you don't.")
"Has read TAOCP" has low prior, but after reading this I have >50% confidence that the poster has read enough TAOCP to meaningfully compare it.
For instance, a user recommending CLRS over TAOCP is an odd piece of advice, since TAOCP is distinctly not a textbook. The user purports to have read TAOCP, an assertion with a low prior-probability.
Unfortunately, I do not think we can rely on the advice of anonymous internet contributors. With no way to enforce the rules, LessWrong fails to do due diligence and is merely spreading the opinions of those that reply.
So, to remedy that, I would add a 4th rule: You must state where your authority on the subject, and on the choice of textbooks comes from.