renard_roux , 11 months ago

I have a feeling that has more to do with the quality af the algorithms you’ve been exposed to than the base concept of them.

As a for-instance, I gave ChatGPT a list of 2 books¹ I was happy with a few days ago, and asked it to recommend similar stuff⁴. It gave me a list of 5 books², 3 of which I’d already read, and very much liked. Asked it to add those 3 to my input list and recommend some more, got 5 new ones³ (one I’d already read, and liked).

Så far I’m on book 2 of the recommendations, very much liked the first one, almost done with the second and it’s great.

I probably won’t get that level of recommendations from Bookwyrm at any point, but it would be nice to have something based on all the data in pumping into it, instead of having to guess which stranger to follow and hope they read something good, and that I’ll actually be there to notice.

——————

1 — input books

• Perfume by Patrick Süskind
• Jonathan Strange & Mr Norrell by Susanna Clarke

2 — first recommendations (® for already read, ®® for read after recommendation)

• The Shadow of the Wind by Carlos Ruiz Zafón ®
• The Name of the Rose by Umberto Eco ®
• The Picture of Dorian Grey by Oscar Wilde ®
• The Golem and the Jinni by Helene Wecker
• The Historian by Elizabeth Kostova ®®

3 — recommendations, second round (® for already read, ®® for currently reading)

• The Thirteenth Tale by Diane Setterfield ®®
• The Night Circus by Erin Morgenstern
• The Strange Case of the Alchemist’s Daughter by Theodora Goss ®
• The Angel’s Game by Carlos Ruiz Zafón
• The Book of Lost Things by John Connolly

4 — my initial prompt:

• Can you recommend any books similar to Patrick Süskind’s “Perfume”? I felt like it was similar, at least in tone, to “Jonathan Strange & Mr. Norrell”, which I also liked.