Often, too, there aren't exact translations for some words or phrases, so you'll have some artistic license based on how the translator decided to translate something. 

Oh, I'm well aware of all that.  There comes a point, though, where the differences are stark enough that you can't just shrug them all away as translation quirks, and that's what I'm talking about here.

As to your point, though, I saw an interesting example just a few days ago.  It's from the LOTR movies.  Check it out, and pay attention to the German version:





At 3:45, Theoden says "ein Bluttag," which literally translates to "a blood day" or "a bloody day" even though the English version is, "a red day."  Now, one can make the case that the day is red because of the blood that's about to be spilled, so the two words would then be interchangeable in this context, but I think it's more than that.  Remember that Tolkien was a master philologist, and if he wanted to limit that phrase to literal blood, he could have easily done it.  He would have just called it a blood day.  No, I think the use of red here is meant to be more all-encompassing: it includes the blood that's about to be spilled, sure, but it also includes rage, the "seeing red" of a man who has completely lost control and given into his anger and bloodlust.  It's as much about passion, particularly the passion of mortal combat and the berserker spirit, as it is about literal blood.  Perhaps even more so.  Like so much of the rest of the book, it's as much spiritual as it is material.

Yeah, I think the German translator dropped the ball here.   

On the subject of Tolkien, never forget that the guy not only invented several languages for his story, but the languages all work according to the rules of language, and he then proceeded to write poems in those languages, and the poems have rhyme and meter as poems should.  It's folly to underestimate him.

I equate some of this as people wanting to substitute money for effort, which is a legitimate choice in many areas of life. Pay someone to clean your gutters instead of risking your own life on a ladder; pay a mechanic to fix your car instead of trying to fix it yourself, etc.

It's not technophobia as such. It seems to be the same old thing: Google is your friend--but read deeply in the entries. Don't just pick the first entry on a page.

I've been pondering when we'll need to put no AI disclaimers on our books.

But, where do you draw the line?  There are lots of things that get labeled as "AI."

If you use AI to check your work for typos, grammatical errors, comma usage, etc., would you need to say you used AI tools to write your book?  Or if the AI tool can point out unclear sentences or maybe even find plot holes or characters whose names get changed halfway through the book?  Would you have to disclaim that?

I don't have any objection to using AI tools in the above described manners.

What about if I use a photograph I have taken and ask AI to re-do it as a comic book style illustration?  I don't see much of a problem there.  The photograph is mine.  The AI tool is using that photo and creating a derivative work.  Or, what if I do the reverse and take line art I have drawn and use an AI tool to make it photorealistic?

In my case, I have a section in my books where I cover the production details for those who might be interested.  So, in those cases, I would include those details in that section.  But I don't know that an AI disclaimer would or should be required in those cases.

My objection to AI is where there is the potential for copyright infringement or plagiarism.  I mean, we know the arguments that "oh, no, it doesn't store anything" but still it manages to sometimes spit out exact duplicates of copyrighted materials used in the training data.  And I don't know that we need disclaimers for generative AI as much as perhaps a ban.

If I write a book, there's a chance I might accidentally write something I've read before.  But, the odds are it would be a line or sentence fragment as opposed to an entire paragraph or chapter.  And, there is the chance that I might recognize it when reading or editing the story and catch it before I publish.  And, if not, a line or sentence fragment accidentally duplicated from someone else is unlikely to be a problem unless it's something really iconic or something, in which cases the odds are good it'd be caught before release.  But, I cannot plagiarize something I have not read.  I might by chance duplicate a line or sentence fragment from something I have not read, but the odds of writing an entire paragraph or chapter nearly word for word from a work I have never read is pretty slim.

On the other hand, with generative AI, there's a chance that it could reproduce a line or paragraph or even a good chunk of a chapter from someone else's work.  And, if I've never read it, I would have no chance of even recognizing that it was copied from elsewhere and, thus, no chance of removing it prior to publication.

And that's where I continue to see generative AI as problematic.

Of course, those problems could be mitigated if there was an AI tool that could "read" your work (that was created with AI "assistance") and check against known works for any matches.  An AI tool that could do that might be well worth it.  It could even help guard against accidental plagiarism where you've unknowingly inserted something from memory rather than writing it yourself.

It's not about how many books can be produced. Or even about who produces them.

It's all about sales.

You can flood Amazon with millions of books, and the top 200k won't change much. Most of the millions wont even get a rank.

It is about sales and it is a numbers game.

Put aside books for the moment.  I see this in other industries.  There are people/businesses that have numerous different variations on the same product that they list on their websites or other sites for sale.  And you look at them and you may scroll through lots and lots of things you would never buy.  And, yes, a lot of them do not sell.  But, some do sell.  The thing is, they don't know what is going to sell and they find ways of producing lots of variations and such at a low, low cost in the hopes that some of them will be hot sellers.  If you put up one item, you have one chance that someone may buy it and you can hope that tens or hundreds or thousands of people may buy it but you have only one thing to attract them.  On the flip side, if you have 100 things, you have a greater chance of (a) selling something and (b) having something a large number of people will buy.

Now, if you have one item that five people buy versus having 100 items where 20 of them sell only once, then you have made 20 sales versus 5.  Now, compare that to having 100 items versus 1,000 items where maybe you'll get 20 versus 200.  But the more items you have, the more chances that one of those items will sell a lot more.  Maybe you have 1,000 items and 900 never sell, but 90 sell one each, 9 sell a dozen each and one gets 100 sales.  There's 298 sales.  And maybe you'll get lucky and one will get 1,000 sales or 10,000.

Why would it be any different for books?  You might argue that the AI generated books aren't well-written, but I've come across a number of high ranking books that were poorly written and selling well, and those books were written by people before AI was even a thing.  So, maybe an AI written book is badly written or maybe it's well-written.  Either way, that doesn't necessarily mean it will or won't sell.

Plus, it's well-established that the more books you have available, the more chances you have to sell.

Now, is it something to worry about?  Maybe, maybe not.  Even without AI, there's nothing stopping someone from hiring dozens of ghostwriters and churning out a lot of books.  AI might be cheaper and faster so more people may be doing it, but you can't worry about what others are doing.  You just have to focus on your own stuff.

Quote from: Jeff Tanyard on December 27, 2024, 10:50:55 AM

And on a personal note, I found Voltaire's Candide and Captain Bligh's account of the mutiny on HMS Bounty to be far easier reads than Hugo's Les Miserables, even though the former two are from the 18th century and the latter is from the 19th century.  "Older" doesn't necessitate "more difficult to read."  Difficulty varies by author, not by century or even millennium.  Plenty of ancient Roman and Greek stuff is easily readable for today's readers.

How much of that is due to the translation?  With the exception of Captain Bligh's work, the examples you've given were originally written in a language other than English.  For example, Candide was written in French.  English translations were made in 1759, 1762 and 1947.  And I'm guessing today's readers aren't reading Roman and Greek works in their original Latin or Greek.

Good question.  I don't know the answer.  My presumption, though, is that a short and simple sentence in English would also be short and simple in the original French, and a long and convoluted sentence would also originally be long and convoluted.  Same thing with long-winded paragraphs and chapters in general.  I suspect Les Miserables would be an absolute chore in any language, and I highly prefer the musical and the 2012 movie (both of which I love, by the way).

Welcome back.  Haven't seen you post in a while.