I’ve been using a somewhat different approach lately.
Instead of looking for “books similar to X”, I use an AI as a kind of personal book-hunting assistant.
The important part is that it already knows what I like — not just genres, but the things I tend to dislike as well. For example, I can say something like:
I’m looking for something with a strong hook, preferably progression or an unusual premise. No LitRPG grinding, no crafting-heavy stuff, no generic fantasy-school setting. I’m also quite happy to read web novels and translated fiction.
Then I let it search.
Google’s AI search is surprisingly useful for this too. You can give it a fairly weird description of what you’re looking for and ask it to find actual books, web novels, reviews, Reddit discussions, Royal Road recommendations, etc. It’s much closer to describing the kind of reading experience you want than searching by genre or author.
And there’s another trick I like: once I find something promising, I ask the AI to follow the trail from that book — what else is mentioned by its readers, what the author recommends, what other books are repeatedly compared to it, and so on.
So instead of:
book → recommendation algorithm → another popular book
I get something more like:
book → discussion → obscure recommendation → another discussion → completely unexpected book
That tends to work much better for me.
I actually ended up putting this idea into a little Android EPUB reader I’ve been working on. It can keep the reading/discussion side in the same place, and I’m experimenting with using AI to help discover books rather than just recommend another bestseller.
Here’s a short demo of what I mean:
https://www.youtube.com/shorts/c7rxSK0f6vQ
It’s still very much a work in progress, but I’ve found this approach considerably more useful than the usual “people who bought this also bought...” recommendations.