“The room was very messy” is outline enough to judge.
If that’s all the input it needed…
But that’s not the question.
The question is: how good of an idea is it to multiply them (artisans) by 100 overnight.
“The room was very messy” is outline enough to judge.
If that’s all the input it needed…
But that’s not the question.
The question is: how good of an idea is it to multiply them (artisans) by 100 overnight.
He still doesn’t figure, this was just a quickly typed example …
Was that GPT or not ?
Either you lied or I don’t have more to figure.
Which is it ?
Many are called but few are chosen.
In any case I am not worried about this.
That’s a good point, thanks. I was about to add that actual unlicensed images, that miraculously show up in “generated” content (including visible watermarks), are certainly not public domain. But you already did that in your footnote.
So that’s the “direct infringement” (if we want to call it that way). There are other questions: Is the usage of unlicensed images for the purpose of training the algorithm allowed, even if not one pixel of them ends up in the “generated” content?
In my opinion it is not. Another concern is the method of acquiring these files. Of course it’s not illegal “to look at them” (as someone wrote), to preview them and to use them in (internal) mockups, etc., but scraping millions of them without consent… I know, I know, search engines. But luckily I don’t get paid to read all those borings TOSs.
Also, the US Copyright Office can only speak for the US. As I said, this will keep courts busy around the world. I wouldn’t go so far to give a global go-ahead.
Sounds like knocking somebody down in an unmarked car (as in: removing the license plates first).
But now all will be chosen.
That’s where I think there is naivety.
In five years, everyone you know who ever wished they had the talent or time to write a book – but so never did – will have.
No talent nor time required.
“Show don’t tell” is an old idea first promulgated in the 1920s. Its exponents, for example Hemmingway, ignored it. It is an idea that has had its time being up there with Maslow’s Heirarchy of Needs and the Seven Stages of Grief in that someone thought it a good idea and others jumped on the bandwagon but not of any real use or worth.
Another interesting use could be to train the AI with App manuals, like I did quickly with a few paragraphs of Scrivener user manual, just to see how it would work:
For it’s not of any real use or worth, the writing community, strangely, is widely discussing it, creating videos and seminars, etc.
Well, it’s no “must” to use this writing style but it’s another perspective. I’d call it a “you can, if you want to”.
Are you predicting that lumber jacks in Chicoutimi will be publishing works of the same quality as Balzac, Dumas, and Hugo?
Yeah and the Biritsh public put The Birdie Song and Bob the Builder to number 1 on the pop music charts. Just because everyone does it does not in-and-of-itself mean something is right. Only a few centuries ago peple did not wash their hands after going to the toilet; it was common practice but it was wrong.
And as someone who participates in “the writing cumminity” I know that some of those discussions are about whether or not it is a good idea.
I am predicting that quality won’t matter this much as the saturation will make it unlikely that enough potential buyers will know, versus the effect on the value of a book.
Less sales at a cheaper price. That’s what anyone with his\her name not already made by then should expect.
And I’m not even getting into the perceived value of a book as one’s accomplishment…
It is as simple as offer and demand. And anyone who’s into it will be effortlessly able to offer. Good or bad.
Perfect. So feel free to make your own decision. I’m not saying, everyone has to write this way, neither do I, but for some parts and descriptions, it’s useful. Everybody’s free to chose.
Until some AI tool decides otherwise.
As a linguist I observe longitudinal change in language; lexically, grammatically, structurally. The current brad of quasi-AI based tools ossify language of at least a few decades ago; they take no account of contemporary usages.
You’re talking about a broad definition of “derivative work.” In the classic sense, it’s writing a book about what Batman did on his day off. Someone else owns Batman, so it’s an fringement absent their license.
An argument can (and no doubt will) be made that using protected work to derive value is a kind of derivative work. If the court likes that argument, then AI training will be viewed pretty much like music sampling is viewed now.
Bear in mind that style, like an idea is unprotectable. In the eyes of the law, you can’t steal an idea for a book, because an idea isn’t property and can’t be stolen. Style is the same way.
I think it’s a mistake to confuse artistic pride with legality, as some in this thread have done. Artists have employed apprentices for centuries, Best-selling authors have ghostwriters/co-writers and draft-writers. Best-selling comic strips are often drawn by teams of artists under the name of the creating artist.
It’s all legal. Is it ethical? Everyone (in a jurisdiction) is subject to the same laws, but ethics are individual.
I still believe that content is king and that quality counts. And, the public will gravitate to higher quality if it exists.
My preceptor once gave the example of the “famous-famous Jalebiwalla” in Delhi. [A Jalebi is an Indian delicacy fried in ghee (clarified butter).]
Many others tried to undersell him by using inferior ingredients and fry them in oil instead of ghee. But even though his jalebis were much more costly he had a steady stream of customers lining up to buy them while the competitors had virtually no customers.
My preceptor remarked “there will always be a ready made market for the authentic thing.”
I wouldn’t say that books are a particularly scarce good now, or that it is particularly easy for even excellent books to find readers.
In the US at least, the Copyright Office has ruled that AI-generated works are not copyrightable. Which makes their commercial value approximately zero. On the one hand, that means that human-authored works can’t compete on price, but it also means there’s no financial incentive to distribute AI-generated works.
They, the customers, had something to compare with, having tasted the good stuff before.
Else, they could have been happy with the cheaper version and look no further…
Which is what I say will happen.
Or they’ll blow their budget on crap before they even get a chance to land on your book. (Assuming it is good.)
That is worthless if you can’t get them to read it first.
And you don’t even have the smell to help with that.
I read somewhere, but can’t remember where, that now AI companies, in order to assuage artists, are now licensing their “style” so that if someone want to do something in the style of someone else the original style creator gets paid. This cooled a lot of tempers. This of course only applies to the living.
Interesting.
Please expand. How exactly one “forces” the other.
Why can’t they sell it, say, 10$ ?