Choosing what to read
Nov. 14th, 2025 04:25 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
blogcutterIn his latest blog post, David Headley of Goldsboro Books in the U.K. asks: Do you trust an algorithm to tell you what to read? :
https://goldsborobooks.com/blogs/news/do-you-trust-an-algorithm-to-tell-you-what-to-read?mc_cid=79cf372777&mc_eid=47fc6ebfe4
He makes some valid, if rather obvious points about trapping you into reading what you've always read and closing off the delightfully elusive and ineffable processes involved in serendipity.
Here's how I would respond.
No, I wouldn't necessarily trust an algorithm. But depending on how sophisticated the algorithm is, it can be an interesting parlour game to find out what the recommendation system thinks I might like. If some of its suggestions sound appealing, I would certainly see if my library or independent bookshop has them available, and find out more about them.
If I'm able to devise my own algorithm or search strategy, or if I have some flexibility to refine theirs, that's great too. That's precisely what I used to do in my working life as a librarian, whether conducting a reference interview, looking in available reference sources, consulting with colleagues or following up on my hunches. It's what I still do now if I'm looking for reading matter.
Can hunches be automated? I think the jury is still out.
Of course, people tend to be consistently inconsistent too. What I'm in the mood for right now may not be what I feel like reading next week or next year. Moreover, telling people they should not or must not read something because it's erroneous, disgusting, hateful or obscene will often make them all the more determined to get their hands on it. Banning a book can be a sure-fire recipe for its success.
I do think sometimes very basic recommendations like "If you like Author A, you may enjoy Authors X, Y and Z" are somewhat useful. There are familiar comfort-reads and then there are challenging or interesting reads that you'd like to discuss with others and that make great book club selections.
Human reader advisory services are still a thing, both in bricks-and-mortar libraries and online. There are plenty of good independent sites out there too. Here's an extensive one I found recently and really like:
catherinerosegunther.com
How do you decide what's going to be your next read? Do you generally read one book at a time or do you generally have two or three on the go at once? And where do magazines, newspapers, newsletters etc. fit into the overall picture?
https://goldsborobooks.com/blogs/news/do-you-trust-an-algorithm-to-tell-you-what-to-read?mc_cid=79cf372777&mc_eid=47fc6ebfe4
He makes some valid, if rather obvious points about trapping you into reading what you've always read and closing off the delightfully elusive and ineffable processes involved in serendipity.
Here's how I would respond.
No, I wouldn't necessarily trust an algorithm. But depending on how sophisticated the algorithm is, it can be an interesting parlour game to find out what the recommendation system thinks I might like. If some of its suggestions sound appealing, I would certainly see if my library or independent bookshop has them available, and find out more about them.
If I'm able to devise my own algorithm or search strategy, or if I have some flexibility to refine theirs, that's great too. That's precisely what I used to do in my working life as a librarian, whether conducting a reference interview, looking in available reference sources, consulting with colleagues or following up on my hunches. It's what I still do now if I'm looking for reading matter.
Can hunches be automated? I think the jury is still out.
Of course, people tend to be consistently inconsistent too. What I'm in the mood for right now may not be what I feel like reading next week or next year. Moreover, telling people they should not or must not read something because it's erroneous, disgusting, hateful or obscene will often make them all the more determined to get their hands on it. Banning a book can be a sure-fire recipe for its success.
I do think sometimes very basic recommendations like "If you like Author A, you may enjoy Authors X, Y and Z" are somewhat useful. There are familiar comfort-reads and then there are challenging or interesting reads that you'd like to discuss with others and that make great book club selections.
Human reader advisory services are still a thing, both in bricks-and-mortar libraries and online. There are plenty of good independent sites out there too. Here's an extensive one I found recently and really like:
catherinerosegunther.com
How do you decide what's going to be your next read? Do you generally read one book at a time or do you generally have two or three on the go at once? And where do magazines, newspapers, newsletters etc. fit into the overall picture?
(no subject)
Date: 2025-11-14 10:45 pm (UTC)"Kalivas!" by Nick Mamatas: the algorithmic recommendations here are pretty decent—grimdark and edgy indie horror, includes things that other people have recommended to me, I would probably enjoy these things.
"To Leave a Warrior Behind" by Jon Tattrie: the algorithm mainly recommends books by the same author, who has written a lot of nonfiction and about Black history in Halifax. Vaguely interesting but not something I'd pick up. I am in fact reading this for podcast research, so whatever conclusions the algorithm makes based on my reading it are likely incorrect.
"Katabasis" by RF Kuang: Only other books by RF Kuang, and since that's not enough to fill up the sidebar, it also recommends translations of her work into Russian, which I 100% won't be reading.
My own latest book: Book by one of my friends (which to be fair is a good comp), book by Premee Mohamed (we write on a bunch of the same themes, so sure), generic-ass fantasy novel, fairy tale retelling by T. Kingfisher (I like her but there's probably not tons of crossover), realism road trip book (???), a bunch of books by friends of mine. Again, mostly things I'd enjoy, but kind of random for anyone else.
I think the main thing is that an algorithm doesn't have human intelligence behind it, just programming to sell more product. So YA romantasy readers are never gonna break out of that, nor will the genre grow because what sells is a thing that is basically like that other thing they liked.
(no subject)
Date: 2025-11-15 02:15 am (UTC)Amazon's book algorithm is odd when it comes to history. After reading a single WWII biography, it suddenly started recommending books by right-wing US political commentators.
https://armiphlage.dreamwidth.org/639834.html
Kobo's book recommendation algorithm is actively trolling me. It recommends other editions of books I have already purchased from Kobo. When I click the "Not Interested" button on recommendations, it recommends the next book in the series, or other books by the same author. Or even recommends it the following week, despite me explicitly telling me I am not interested. When I download a preview (or the whole book via Kobo Plus), and then delete it after reading the first few paragraphs, it then recommends everything the author has ever written.
(no subject)
Date: 2025-11-15 03:51 pm (UTC)I've never used Kobo's book recommendation algorithm. I used to have a Kobo e-reader (it may actually still be around somewhere). It was one of those "free with purchase" deals and came pre-loaded with a bunch of classic books for which copyright had expired. Started as a good space-saver and ended up as a waste of space.
(no subject)
Date: 2025-11-15 08:58 pm (UTC)Most of my reading is history with occasional fiction and poetry.
I haven't read a newspaper in years.
Newspapers
Date: 2025-11-17 02:51 pm (UTC)How do you typically get your news?