Why would they add LLMs to Calibre?

It looks like it's so you can "discuss" a book with the overgrown autocomplete, or ask it what to read next if you are incapable of independent thought. I could, sort of, understand making a plugin for this sort of nonsense, but building it in, as a brief check of their site suggests has happened, is so utterly ridiculous that it defies rationality.

I liked Calibre, I've even contributed code to it, but it looks like I'll be investigating this fork now.

Why would they add LLMs to Calibre?

So it can scan text.

I have not heard of this program until just now, but presumably they're not integrating an LLM so you can stop and chat with fictional characters. It's gotta be an overpowered answer to 'here is some plaintext, populate its metadata.'

The guy behind it isn't really known for caring what users think. Looks like he's even disabled issues for Calibre on Github lmao

Depending on how you use it, you might be able to get by with Calibre-Web-Automated which has a kobo-sync snapin.

Kovid Goyal

Used to be a fairly accessible guy.

Its a niche fork of a niche product that doesn't make money for anyone involved in making it. They don't care.

Considering it's already the first result searching "calibre without AI" it'll probably be fine. No one will use this who doesn't already know about calibre.

What would the difference be between it having one user and it having a million users? I doubt they are banking on getting mad donations, there are easier ways to make money.