Aha! I was posting about a thing I made for just this purpose the other day - it is called Leafle. You talk to it about books you’ve read and enjoyed or not, and it uses that plus the fact that llms have very extensive knowledge of pretty much all books, to suggest ones you’ll enjoy (or in fact, that will meet your need - reading isn’t just about ‘enjoying’ after all).
It can suggest them to buy, or as audiobooks, or it can find ones currently available to borrow from your local library!
This for me is one of the last great unsolved problems on the internet. I've yet to find a book recommendation site/algorithm/social thing/whatever that actually works well enough for me to trust it and invest my time in logging my own reads to improve the results etc.
I've had far more success personally in growing my reading list with books I ended up loving by following a few people on Instagram who share their reading groups choices. Similarly, anytime anyone whom I vaguely trust recommends me a book I just order it immediately on Amazon and add to the input pile.
It's vitally important for me to keep the next read pipeline well stocked - I get through a lot of books, and if I run out for whatever reason it can take me weeks to get the habit going again.
Virtually every other form of media, I can find a recommendation system that works well enough, even if it's not perfect. Books have always been an outlier for me personally.
I find the best recommendations to be the works cited in other works you like. You can press pause in a movie when the camera goes over the bed side table book, or write down the movie they go see in the book you read. Gives a growing list to discuss with your librarian, although some time you need to branch out. Nice idea that’s mentioned there, shame it is only for non fiction books, would be interested in a fiction equivalent!
I use embeddings to create suggestions on what to read when it comes to blogs and articles. It semantically matches articles from sites I provide (like HN of course) against previous things I have read and a set of topics/themes I am interested in. It does this periodically and provides a suggested list of what to read next. Implemented this a few weeks ago and its throwing up pieces I might have otherwise missed.
I am however concious that sometimes the most interesting things in a given day are not related to anything else I am reading or would read - this is what I get through scrolling through HN serveral times a day.
I always have a few novels on the go at the same time. As I finish each one I put it on the 'read' pile and pick another complementary book from my bookshelf whose last rereading I can't recall. This way I work my way through all the books I own before starting again a year or more later. No need to buy any more books ever...
My favorite rubric for this is: give yourself permission to put the book down and never pick it up again at page (100 - your current age). Which means that, yes, once your turn 100, you're allowed to judge a book by its cover.
Start somewhere.
Go places.
It's fun.
JW
It can suggest them to buy, or as audiobooks, or it can find ones currently available to borrow from your local library!
It is at https://leafle.nanosheep.net if you’d like to give it a go :)
I've had far more success personally in growing my reading list with books I ended up loving by following a few people on Instagram who share their reading groups choices. Similarly, anytime anyone whom I vaguely trust recommends me a book I just order it immediately on Amazon and add to the input pile.
It's vitally important for me to keep the next read pipeline well stocked - I get through a lot of books, and if I run out for whatever reason it can take me weeks to get the habit going again.
Virtually every other form of media, I can find a recommendation system that works well enough, even if it's not perfect. Books have always been an outlier for me personally.
I am however concious that sometimes the most interesting things in a given day are not related to anything else I am reading or would read - this is what I get through scrolling through HN serveral times a day.