Turn off/Block AI features

[Bradles]

Originally Posted by DiapDealer

Those who wanted the ability are happy to jump through all the hoops to make it work...

As some of you may be aware, I proposed an AI Book Summary plug-in two years ago. The feedback was overwhelmingly negative, so I abandoned the project almost immediately. (I still use it for my own amusement on occasion.)

With that in mind, and with all due respect, who are these people who wanted the AI features, and where's the evidence?

[DiapDealer]

Originally Posted by Bradles

With that in mind, and with all due respect, who are these people who wanted the AI features, and where's the evidence?

As we're not in a court of law, I fail to see the relevance. There are two sides here. Their relatives sizes are immaterial. The point is that both sides have been accommodated.

[Quoth]

Originally Posted by DiapDealer

As we're not in a court of law, I fail to see the relevance. There are two sides here. Their relatives sizes are immaterial. The point is that both sides have been accommodated.

In your opinion!

[DiapDealer]

Originally Posted by Quoth

In your opinion!

For god's sake, YES!

But in all fairness, your opinion is that NO ONE should have an opinion on AI that doesn't match yours. Whereas I'm perfectly fine with people falling somewhere in the middle of the love/hate AI spectrum.

[JSWolf]

Originally Posted by DiapDealer

For god's sake, YES!

But in all fairness, your opinion is that NO ONE should have an opinion on AI that doesn't match yours. Whereas I'm perfectly fine with people falling somewhere in the middle of the love/hate AI spectrum.

I fall in the middle. But I find there's no sense complaining. It won't change anything. Besides, It's very easy to ignore the AI if you want and the code doesn't run unless you want it to.

Complaining is (IMHO) just useless.

[Anna H]

Originally Posted by Quoth

I'm not updating till this is scrapped. It's a thin end of a wedge and encouraging people to use these over-hyped LLMs, even though off by default. Destroying the planet, impairing mental health and filling up with plausible nonsense, or illegally copied sources and without accurate citations.

There are "AI" that works, but Generative AI / LLM isn't one of them.

Amen! There are many use cases for LLMs, but they are getting shoehorned into places where they are demonstrably WORSE than the tools that preceded them, or into places where they take far more resources than are needed, or are put to sues that are dangerous to their users' physical and mental health. The things this "AI" are supposed to do in Calibre are things that either don't need such a resource hog to do ("if you liked this book you'll like this other book" has been around since long before the current AI push), or that you don't need at all. It's bloatware. I really hope the developers will come to their senses and scrap it quickly!

[Sirtel]

Originally Posted by Anna H

...or that you don't need at all. It's bloatware. I really hope the developers will come to their senses and scrap it quickly!

How do you know what other people might or might not need? Are you a clairvoyant? Or do you just feel entitled to speak for all the millions of calibre users about what they need?

[DiapDealer]

There is no AI bloat within Calibre. The resource hogs are the llvms that already exist (either locally or somewhere online). If you don't configure calibre to utilize one of those pre-existing resource hogs, the amount of code added to calibre to make these feature possible doesn't even come close to rising to the level of "bloat." If you don't use it, you are unaffected by it. All rants against it are silly and misplaced (and yes, of course, that's only "my opinion").

[Quoth]

Originally Posted by DiapDealer

For god's sake, YES!

But in all fairness, your opinion is that NO ONE should have an opinion on AI that doesn't match yours. Whereas I'm perfectly fine with people falling somewhere in the middle of the love/hate AI spectrum.

No, you are not, because your opinions mostly seem aligned with corporations and you constantly denigrate people that have an alternate opinion, and even ignore facts.

See the analogous desire of people to have TVs that are not "Smart". It used to be that you could ignore "smart" by not giving wifi password, or not connecting cable or not logging in. The smart features encouraged people to use them and the makers, Google and others then "spied" on usage including what was connected / played via HDMI, not just channels watched or online content used.

I'm told that some Android & Fire TVs / Projectors now have no function unless you connect and log in at least once.

But you don't seem to care about privacy, environmental waste (AI LLMs), consumer protection etc. Other people do care and don't support the Corporate and so-called "Libertarian" viewpoint (which is about Corporate freedom, not personal liberty). It is absolutely the thin end of a wedge even if you don't care. It's better than Firefox's adding of "AI", because it's off by default.

You are welcome to have your own opinions, but that doesn't give a freedom to denigrate people wanting privacy, care of the environment and consumer protection. Frankly you rebut opinions in a bullying style rather than state your own.

[Loosheesh]

Off topic but, for some reason, this whole discussion got me thinking about dead horses (my brain is weird ), which in turn reminded me of something funny I read a while back in The Maxwell Daily Reader ...

A friend sent me “The Top Ten Strategies for Dealing with a Dead Horse,” which I think is hilarious:

1. Buy a stronger whip.
2. Change riders.
3. Appoint a committee to study the horse.
4. Appoint a team to revive the horse.
5. Send out a memo declaring the horse isn’t really dead.
6. Hire an expensive consultant to find “the real problem.”
7. Harness several dead horses together for increased speed and efficiency.
8. Rewrite the standard definition of live horse.
9. Declare the horse to be better, faster, and cheaper when dead.
10. Promote the dead horse to a supervisory position.

Like I said, off topic ... mods, please be merciful and remove if necessary

[Quoth]

A famous author's view. Note that ALL the LLM based AIs used by this option have illegally used content, so when the answer is correct on something not in the Public Domain, it's copyright violation.

https://whatever.scalzi.com/2025/12/...-for-anything/

[DiapDealer]

Originally Posted by Quoth

No, you are not, because your opinions mostly seem aligned with corporations and you constantly denigrate people that have an alternate opinion, and even ignore facts.

I'm sorry. But that just means you're not really paying attention. I personally have not found any use for AI as of yet. That may change in the future, but maybe not. Right now, I have it disabled in just about every place I've encountered it. The difference is, I'm not raving about it. I'm letting people choose differently than me. You don't. So what you've mistakenly identified as some sort of alignment with corporate love for all things AI is merely a willingness to let reasonable people make their own decisions about how they use AI (or don't use it) without preaching at them. People who tend to be unreasonable about it get my attention, however.

Which means a find offense with anyone who tries to bully an open-source dev into making sure that no one gets to make a decision like this for themselves. By declaring for everyone that it has no place in a piece of free software.

You've also completely missed the point in saying that I'm denigrating people. I'm not. I simply choose to challenge you whenever you invariably try to force your unreasonable aversion to anything called AI onto everyone else. Unlike you, there are many AI threads around here I take no interest in. But you can bet whenever you try to decide for everyone what does or doesn't belong in freely developed open-source software, I'm going to take notice. I'm not going to call you names, I'm not going to attack you personally, but I AM going to make sure you know--in no uncertain terms--that I think you're you're dead wrong. Nothing more, nothing less.

Stop assuming that people who disagree with you on this are mindless corporate wonks who want everyone to love all things AI. Talk about denigrating.

[jhowell]

Originally Posted by Quoth

Note that ALL the LLM based AIs used by this option have illegally used content, so when the answer is correct on something not in the Public Domain, it's copyright violation.

Not necessarily. Knowing something about a book is not proof of having illegally obtained a copy of it. AIs can be trained on sites that contain plot summaries and reviews.

Personally I have found AI to be very useful for coding. And Amazon’s new feature of being able to have AI recap a book you are reading up to where you currently are in it seems useful if you take up a book after a reading break. I don’t see any way that could be done without using AI.

On the other hand Amazon recently had a major blunder when they produced an AI recap of season 1 of Fallout with obvious mistakes. (Amazon pulls its bad AI video recaps after Fallout fallout)

[sushigal007]

Personally, if I left a book long enough that I'd forgotten that much of it, I'd just reread it. I don't trust AI not to mix it up with the movie version, or a fanfic, or to just straight-up hallucinate something else entirely. I don't trust AI with anything creative. And on that note, I was already not updating Calibre because the Generate Cover plugin was having font issues, so I'll continue not doing that, because now they're in the program at all, I also don't trust that AI 'features' will stay switched off in further updates.

[DiapDealer]

Originally Posted by sushigal007

so I'll continue not doing that, because now they're in the program at all, I also don't trust that AI 'features' will stay switched off in further updates.

Without a manually configured (by the user) external llvm connection/account, there is simply nothing that CAN be automatically "switched on". Please understand that.