If that is the case then the magic 8-ball says
It will be a civil suit, so the plaintiff and the defendant.
Which is the reason for corporate secrecy- especially of formulas used in some indices and calculations.
"this is a choice" while ignoring the context that there may in fact be present labor abuses, which are listed in the article, which should be accounted for. Everything from no age verification (allowing for exploitive child labor where children are forced to look at the worst elements of the internet, which can likely cause mental trauma, its in the article that you claim is hyperbole), bans from work and no pay for no reason whatsoever (so they don't have to pay you by making excuses), skirting the law by using contractors who source labor from parts of the global south that have no other jobs available at the local level in many cases.
Unless you are a lawyer, you don't know any of this is a certainty. Even experienced lawyers are hard-pressed to say anything with certainty until the courts decide what is actually going to happen with the law. This will take years to conclude. In the meantime, all of this is conjecture.
Is anyone really surprised that companies like Getty Images and the NY Times would try to keep their relevance through lawsuits? It's what they've been doing for twenty years.
While I wouldn't represent myself or other in courts this doesn't mean I don't read up on the law to understand my rights and how to operate within the legal system. This also makes the time I do spend with my attorney more productive as I understand better the options they would be telling me.
Because many folks are reacting emotionally to what they imagine will LLM do. As opposed to how they actually work. Also, this is not a fault unique to those against LLMs; many of those in favor of LLMs also have unrealistic views of the technology.
Because they want to control the use of their content and make money off of it. Most large companies making a living off of copyrighted work would lock down and monetize every bit of content they make even if it gutted things like commentary, reviews, and public libraries.
What is ethical is to return what is taken from the public domain in a timely manner. To recognize that the work we create rests on the shoulders of those who dreamed before us that the right thing is to return what we have borrowed for our own purposes. Copyright was formed in the 17th century as a tool for censorship.
For the 100th time, I am not 100% against the tools. I am tired of repeating myself at this point on that subject. The tech itself is however completely unsafe for commercial use until legal matters are resolved, because no one can truly assess the potential risk until the damage is done. Telling people otherwise is putting them at risk, I am allowed to warn people that using the technology comes with inherent risks.
I will address this in a separate post because this is a completely different issue.
(star trek reference) 
It's really not. It doesn't study, reflect or gain muscle memory like a human. Second when you've got ais that give you "in the style of [named artist]" that's not pieces that would be considered ethical to pull out as an artist, unless it's transformative and an hommage or parody, or you have IP owne consent or it's CC. Copying is part of private study, but it's really a small part unless you're on a professionnal copy maker carreer (eg making copies of old paintings), most of it is exploratory, technical & intellectual.
Those things are changing, there was a documentary on AI learning a few years ago called Chappie.
How tho? If I wrote a whole manuscript and character relationships are relevant, I probably know it all, have notes, and prepared most of it before writing etc. I t doesn't make sense to need that post hoc, it's completely pointless. Do you need an AI to tell you how many classes and races there are in your setting?
I know many authors who don't keep statistical data on the setting they create. Then at some later time want to tabulate these numbers. With right setup involving LLMs this can be handled by analyzing the original text. As opposed to the previous method of going through the text, extracting the relevant data, formatting that data, and then compiling the statistics.
Get a subscription to ChatGPT and try it for it self. Since January, they updated it so you can point it at your own image and text, and you can do various things with it. The main problem it is very much a Version 1 feature and there are still limits on how much you can throw at it. For example, I tried feeding it all my Majestic Wilderlands stuff across four decades, and it choked on that much data. But it was able to handle say six documents I had in a folder about the Ghinorian Empire.
So it is different and not "just like humans", which is my point. No one denies computers are better at calculations, I'm saying that computations using Image generators are useless creatively - unlike say 3D softwares, digital painting software, etc
Comics style is a matter of working on short delays, often from pauses taken from magazines, etc, and a product of publishers wanting a relatively standardized product. Comic artists also do a lot of study that's technichal, character design, etc... but you wont see it much because they specialized in giving a given result. But often if they have a public side production you will see they do all sorts of stuff. Same for the crappy digital art made for viral social media spam etc.
No I'm talking about what artists do. But consumers will also catch up on it if it's too obvious, probably why you're aware of it. It is unethical and not seen well.
You stated that Image Generators learn just like humans, and this was logical demonstration it is not the same thing at all. I'm not saying it's valuable info, I'm saying it proves the process is completely different. Artists can even tell you why they like this or that, why they took this and ditched that, etc. Can a machine makes this sort of creative choices?
Is Chat GPT perfect? Does it detect plot holes? I doubt on either case.
What statistical data can be obtained thus AND be useful AND is not already covered by any basic writing software (eg word X apears Y number of times, indexation, etc). I m also skeptical on the fact it has no quality detector or creative decision making.
That's a possibly interesting feature, but :
Of course not it a random table generator for words (or pixeles0.
That would require it to understand context, which, in my opinion, isn't happening with this technology. But what it does well is respond to questions you ask it. And a good line of questions can lead to figuring out stuff about what you wrote.
I believe the example I posted just illustrated that. No writing software is going to spit out a list of realms you wrote about in a document. The current main limitation of ChatGPT and other LLMs is that at a certain point you are better off using a traditional tool like a spreadsheet. A LLMS can get you data but then fall apart analyzing the data. A good example is character generation in a LLM.
Wow, sorry to be critical, but you are making way too many assumptions here. I am very good at remembering enough about history to know where to go to find the details including fictional ones. But it is not a photographic memory, and I do not remember everything accurately.
Yes, it is accurate enough so you can find the actual passages to make sure what it really said. It is faster than rereading all the relevant documents, and I read very fast.
Right now, it is on par; once they overcome its limitations, it will be so much better there will be no comparison. Currently worldbuilding software can handle more. But also, the software has an upfront organizational penalty in that you have to get your document and data in just the right form for it to be useful.
Not sure what you mean by in-situ?
Very useful as it ability to search and summarize is at a level we haven't seen before. The longer the original document the more useful it is. Again I have to stress that the best way to use it to point you to where you need to look. You still need to read the original material. You are using it as a ToC or index on steroids.
If you don't have a need then you don't need to understand how it works then, I guess. As for me, I have thirty years of digital notes, and prior to that a decades worth of notes, in all kinds of format. And ChatGPT has proven very useful to me in sorting it out.
So perfection must be irrelevant for memory too.
How accurate is it at that and how much realms does any setting have that are relevant and the main author wouldn't remember? So I'll give you in some niche circumstance that might be somewhat useful. But given it does not understand context and only looks up wheighed associations, how is it different from ctrl+F "realm" ctrl+F "empire"?
Aids for disability esp transcripts & so on I think is a perfectly valid case use btw. But again the question is productivity. If you wanted to make a youtube video for exemple instead of transcripting one, would that be useful? You could record it then transcript it, and I guess have AI analyze tone or something... but that would tell you not much. Someone with good ears could tell you if the sound is good and you're well understood etc. A pro could fine tune it faster than you can decypher AI ramblings.
Being critical is good, I'd never criticize someone for it
Fair enough.
Ok so right now, this is speculative and existing systems do things better is the factual information here.
So is is or is it not accurate?
But ctrl+f can also do that.
I understand. Keep in mind I'm not saying it's not useful to you. I am questioning the idea it's going to be a decisive advantage in terms of productivity (unless quality loss) as it pertains to creative work. No offense but if you've got 40 years of notes strung about all over the place, you're already not competitive with the top organizational minmaxing (me neither) so I don't think this does much to prove the point. But the fact is not even being top minmaxing, just always having a PC, it would only add another useless step to an already efficient process.
There is a race to the bottom with digital art and I think it's actually good advice to try to resist it and that going viral on social medias by spamming is overrated. And I very much agree traditional art is not to be underestimated as a source of revenue.
If you want a more nuanced discussion of the wages in the countries impacted you can see it here:
basically the wages called "slave wages" in the articles turn out to be average wages in the countries where the work is being done so not horrible pay for what amounts to a job requiring minimal skills.
People living in these countries, from the comments in the discussion, generally see the income as a good thing and raising the standard of living leading to eventual rises in the overall wages for the countries if they last long enough.
The articles themselves say the pay is above the minimum wage and use baseless claims of "digital slavery" while showing only limited examples of force (prisons/internships) being used to support that. In low wage countries it appears this is a choice and pays relative to other jobs in country average.
If the reporters here did a better job of providing context and support I would be more moved by the information but as it is they seem focused on hyperbole.
I have addressed this in an earlier post. Artists can't speak without being anonymous because they literally risk losing their jobs. Today NDA's are also standard practice for most companies, meaning that an artist cannot talk about anything related to their work as it comes with the risk of getting sued by their employer. Think about that for a second.
Well, that's a lot of stuff covering a lot of different subjects. I am limited on time so I will do my best:Of course everyone wants professional-grade art. It's not that they don't want to pay for it, it's that with the best will in the world there isn't an artist cheap enough to provide art for small press RPGs and allow for any sort of economy of return. (some folks out there are super lucky and have an artist in their crew - thinking of the Jorune and Nightfall folks - I've never been that lucky.)
I pay artists. My latest book had three human artists as well as my own paltry efforts. It will be a year before that pays off just the cost of the art alone so the long tail means the book stuff is a hobby.
I'm not really bothered by the AI art thing. Messing with prompts has never gotten me exactly what I want but I'll use it for character portraits in situations where I'd never pay. Where otherwise I'd just be scamming google images for pictures of people to use as a 'face'. I'm not really bothered by the training thing because that's exactly how humans learn (by studying and not crediting or paying other artists).
I remember using the healing tool in Photoshop and thinking "omg what a cheat this is". Same with filters and masking. Artists will learn to use these tools and just get on with it. They'll use the prompts to create the generics and then use their talent to overpaint what needs overpainted. They'll get the rep for being fast and efficient. Those who can't keep up will be left behind. Cruel? Maybe - but then I learned desktop publishing in an RSA course aged 20 for fun....sitting beside a dude in his 50s who was told by his boss to switch from manual typesetting to using Adobe Pagemaker or find another job.
Either way, it ain't going away.
The big companies will license art and train their models on art that's properly licensed. What then will be the major objection? That computers are replacing people? We can all start beating our laptops with our wooden shoes in that regard.
Anecdote: I visited the labs at the local acute hospital back in November. They have immense machines that process literally thousands of blood samples a day. Operated by one lab tech. That one machine in one hospital serving less than half a million people replaced easily more than 100 lab techs.
Anecdote 2: The local university just ran a 2 hour seminar on how to use AI tools, notably chatGPT and NovaAI, to improve essays. The young lads I work with (recent grads) all say they'd not have passed their degrees without chatGPT writing code and essays for them.
It's entirely shit that computers are replacing jobs. It's entirely shit that they're replacing jobs where people **love** their job or where they don't get paid much and unfortunately rent costs money. It's shit, but it's hardly new. And no, I don't have any answers.
You mean like your phone being assembled by a low wage labour? So clearly that isnt tech.
Google "whataboutism."
That is not a point I raised.
I outlined the legal issues and the specifics. If you wish to make a specific point rather than a blanket warning, I am willing to discuss that point. Using fear, uncertainty, and doubt is not a great way to discuss the issues.
I think a lot of voices are being heard on this issue. As for the problems with the freelance market that has been under severe pressure for 15 years as a result of structural changes caused by digital technology. Generative AI/LLMs is just the latest in a sea of changes that impacted the creative industry.
In contrast, I am a freelancer who does cartography projects for other folks. So, I have some relevant experience in this matter.
We are in the midst of a hype. It will sink in that LLMs are random generators. In the long run, is that artists will start use the technology will be far more productive. Single artists will be able to handle larger projects at a higher level of quality. The negative side of this it will broaden the income gap between those with genius talent versus those with mediocre talent.
I will address this point in particular. What I advocate for and encourage is open content under an open licenses. When I talk about specifics, my advice is that everybody benefits when systems and rules are open content under open licenses. However, when it comes to adventures and settings, open those only if you feel you will gain a goodwill or marketing advantage. While there is an ethical side to what I say about open content, more importantly, open systems encourage the network effect that RPGs rely on in order to thrive.
The pathway out is to support a system where individuals and not fictitious people in the form of corporations are held accountable. Fortunately, we have one with how the internet and digital technology are developing. The barriers and capital costs have lowered enough to allow individuals and small groups to pursue ambitious projects that are competitive with offerings from the largest RPG companies like Hasbro. And individuals and small groups have a bigger stake in their personal reputations. Individual freelancers have a better chance at forging partnerships than just a work-for-hire relationship. Along with the fact that most folks are but a Google search away, like myself, as far as their reputation goes.
Respectfully replying:
I don't have any surefire solutions, as at the end of the day I am just a client to freelancers, nothing more. What I can do is advocate for those who risk losing their jobs and their livelihoods for speaking about it. A significant portion of freelancers are forced to work under NDA's thus the only way they can tell share their perspective is via an anonymous source.
100% agree completely. Cheers.
I will have to agree to disagree with this one here, as in my personal experience, boots on the ground, is that the tools do increase productivity, but at the cost of quality. Artists are able to work faster, but I have witnessed a significant decrease in overall quality of work created by artists that use the tools in their workflow. As a client this is a problem, as this costs me more time and money to fix the art myself or hire another freelancer to fix any problems with a piece.
Well then by all means, how about some transparency. Do you plan on using gen-ai images in place of artists? Inquiring minds would like to know... /s :p (OK I am joking, I won't assume how you operate, but you can't deny that you are not without bias on the subject as you stand to benefit financially from the current status quo).
Except you are talking about a completely different issue here. I support open game licenses as well, up until the OGL disaster last year, now I am looking into creative commons as an alternative to the OGL for my own SRD releases. The OGL debacle inflicted a significant amount of damage to my publishing business, as in the period of waiting for a final response from wotc, I was forced to take action to protect myself and my project by scrapping the 5e version I spent years designing, and killing it for an in-house system. I know that the OGL is now technically safe again, but after wotc pointed a gun at my head (and hung the Sword of Domacles over the entire industry), I will not give them another chance to pull that stunt again.
And I can't really disagree with any of this either, I think that we are generally on the same page here.
I will say again that the statement you responded to was a prediction of the future on my part. The current state of affairs, in my opinion, basically amounts to a modest automatic indexing/reference system that has to be double checked and a decent set of random tables like Judges Guild Ready Ref book, Midkemia's cities, and so on.
That list is reasonably accurate. It correctly listed all the realms are the focus of the document. But omitted those listed as external powers and those who were mentioned in the region's past.
