What upsets me most is how much of a treasure trove of data that amazon/goodreads are sitting on and yet their recommendations are still pitiful. A basic collaborative-filtering algorithm would be better than "OH WE THINK YOU WILL LIKE THIS BOOK BECAUSE IT IS IN THE SAME GENRE AS A BOOK YOU BOUGHT A YEAR AGO!".

I'll humbly plug https://abooklike.foo just in case people want better book recommendations. You can input books u like and it'll find others you may like. I can explain more about its mechanics if useful. Been working on it for around 6 years now as a side project. Anyway, just posting in case someone's scanning through the thread and happens to see : )