The problem with reading book lists like this is that nobody has time to read all the books. That's a ton of crap out there and I want HN to help me filter through them.

Thus the problem with existing solutions is NOT "limited recall" or "insufficient rules" or "no Amazon link".

And the problem with this "solution" is that there is no justification for why a book is great and applicable to my circumstances, and people have to trust your black box. Otherwise I'm likely to waste my time, just like reading books from any other crappy recommendation engine.

With a deep learning model reducing all the reviews to "book names" you've successfully removed the value of the book discussions themselves. Therefore, for me this engine and all similar engines are strictly worse than simply going through the actual big threads themselves, i.e. https://news.ycombinator.com/item?id=21900498

Edit: I've just seen the embedded comments by switching to a desktop browser. It's a nice addition. However, for me to make sure I'm not wasting my time going through arbitrary books and comments, I would need to know why a book is ranked highly compared to other books. And I want to be sure that ranking is tailored to me, at a very, very high accuracy.