[jaysonelliot]

The concept is great. Some constructive feedback on the execution:

I expected to be able to provide a lot more than just five books. The splash page says "Our AI analyzes your reading DNA to recommend books you'll absolutely love—guaranteed to match your unique taste." That makes it sound like it's going to dive deep into my reading history and really think about my taste. Five books is absolutely not "my reading DNA."

There are lots of ways to get a richer picture of the books a person loves. You could connect to Goodreads or Storygraph, or scrape their social media for books you've discussed, or let users upload a .csv exported from other sources like LibraryThing, their Amazon wishlist, or their own local lists they might keep on Obsidian, Notion, or wherever. My public library keeps my reading history automatically - that would be another good data source for my "reading DNA."

Right now, it's just an AI recommendation based on five books. I can do that with any LLM from ChatGPT to Copilot to Gemini. The recommendation I got was very basic, just obviously similar books from the same authors I entered or ones that are closely related.

People's tastes are complex. Even if you allow much larger data sets to create a person's reading DNA, that alone won't necessarily recommend books that are right for them. For example, I love PG Wodehouse, but I have no interest at all in Evelyn Waugh, James Thurber, or G. K. Chesterton. A great recommendation engine would ask me why I love a book, and try to tease out the reasons behind my reading list in order to recommend books that will be more accurate and unexpected than I could get from a simple ChatGPT query or my Goodreads profile.

A site like this needs to do a lot to stand out. It's an excellent concept, I hope you develop it into something special.

[quesera]

One huge challenge is in identifying and measuring the axes of appreciation.

Asking "What do you like about X?" is a tough way to extract good data. People usually cannot explain why they like things. And it's legitimately difficult to know sometimes.

Also, tastes are often context-driven/sensitive. A book that I loved when I read it last summer in Barcelona, or on the 16-hour flight to Auckland ... does not necessarily map to what I would enjoy reading right now. Or that I should pack for my trip next week.

I've tried to suss this out in music. Songs are theoretically more approachable than books/films/etc: Bite-sized consumption quanta, a fairly robust (but large) genre taxonomy, one basic grounding theory (not really, but a reasonable approximation for the culture within which I exist). Then you can split out by instrumentation, style, arrangement, tempo, etc and get some well-defined groups.

This doesn't work. It's over-analytical, and under-representative of human taste spectra.

The "best" engines use high-resemblance cohorts, but no one actually likes them -- they give lame obvious suggestions, and are terrible at surfacing surprises. They're OK at "good enough, sometimes" in the same way that turning on a TV for the 6pm news and sitting there on the same channel until Letterman signs off was "good enough" (i.e. horrifically bad!) back when serial TV was a thing.

There remains something ineffable about taste -- "It don't mean a thing, if it ain't got that swing". (Ironically, "swing" is now probably measurable! But the point remains for other as-yet-undefined axes.)

[jaysonelliot]

That's a great point. For example, my wife probably knows me better than anyone in the world. She's very good at seeing a book and knowing I'm probably going to like it. That includes your example of "this is a good book for a trip" vs "this is a good book to read at home, at night." But even she gets it wrong about 20% of the time.

In order to be able to really recommend something as multi-faceted as a book, movie, or song, you have to know a person on pretty much every level. I suppose seeing a person's entire social graph, search history, LLM history, media consumption history, and browser history might get you close, but it's still a Hard Problem™.