A.I. Imaging

Best Selling RPGs - Available Now @ DriveThruRPG.com
Status
Not open for further replies.
Yup. Art is also one of those fields where the client thinks they're paying you to be art director themselves, as opposed to just for the final product and letting you do your own thing following your own process. I had one guy say that he wanted me to draw a logo from scratch rather than work from premade icons and such as a base. Luckily that never went anywhere, cuz those are the worse clients to work with.
Yeah, I get that with copywriting. Some clients, especially inexperienced ones, don't seem to quite get that the client/contractor relationship is very different than the boss/employee relationship and I simply don't have to put up with the same level of crap. The client tells me their requirements. I do the project to those requirements, generally with an agreed number of rewrites. I won't work with micromangers.
I'm guessing people will likely complain, like you're robbing them if they know you're working from AI generate art as a template.
I'm absolutely sure they will. There's still clients that try and make people work on pay per hour. As far as I'm concerned, if the project is done properly it's none of their business if I take 20 minutes or 20 hours. But I've definitely come across potential clients who would think that I should reduce my rates because I'm faster than the bottom feeder end of the market. (And then sometimes I end up having to rewrite a piece they tried to save money on and they end up paying exactly the same price they would have if they'd come to me first, plus whatever they've paid the first person).
 
People complain no matter what. My father designed ships for a living and always preferred to work for commercial folks vs non commercial. The commercial folks knew why they wanted things even if what they wanted was wrong. You could correct them and not get them feeling insulted because all they really wanted was to make money.
Non commercial were alwaysaking wierd nebulous tradeoffs that only they could ultimately decide.
I find people in the creative industries tend to be either completely impossible or an absolute joy to work with. I like doing musician bios because most of the time the unofficial brief is just "I want to sound as cool as possible" so I can move away from the formulaic.
 
JR0BI3d.jpg
I really struggle with stuff like this - it's a bit uncanny valley - it looks both 'real' and so wrong.

When I knows it's art I don't bat an eyelid at Grom the Paunch or Shiva - but this gives me the heebie jeebies.
 
Dream is crap at modron tridrones, but pretty good about Godzilla after he taked a trip through the wh40k warp.
blank_tradingcard.jpg
Gonna try Elvis impersonator Godzilla next.
 
Just to throw this grenade into the room, for some reason I remembered this from Steve Albini today. (He's talking about the music industry, but I think there's obvious parallels).

As is true every time an industry changes, the people who used to have it easy claim the new way is not just hard for them but fundamentally wrong. The reluctance to adapt is a kind of embarrassing nostalgia that glosses over the many sins of the old ways, and it argues for a kind of pity fuck from the market.
 
Just to throw this grenade into the room, for some reason I remembered this from Steve Albini today. (He's talking about the music industry, but I think there's obvious parallels).

As is true every time an industry changes, the people who used to have it easy claim the new way is not just hard for them but fundamentally wrong. The reluctance to adapt is a kind of embarrassing nostalgia that glosses over the many sins of the old ways, and it argues for a kind of pity fuck from the market.

Artists are about the furthest away from "people who have it easy" in an industry as one can possibly get.
 
As is true every time an industry changes, the people who used to have it easy claim the new way is not just hard for them but fundamentally wrong. The reluctance to adapt is a kind of embarrassing nostalgia that glosses over the many sins of the old ways, and it argues for a kind of pity fuck from the market.

Perhaps, but modernity asks us to adapt at a pace that frankly does feel unnatural.

The vast majority of our ancestors in 1600 lived a life their ancestors in 1300 would have recognized. Work, culture, amenities, landscapes, relationships, philosophies, they didn't change that much for the farmers/farmer-support that made up 99% of the population.

We, on the other hand, have to put up with seismic change every decade, perhaps less if we use modern transport and administration to move around a lot in that desperate search for opportunity.

Nostalgia and homesickness used to be synonyms. Many of us just want to 'go home' to the decades that formed our tastes and comfort zones, like a captive dolphin wants to go back the the sea, no matter how many "sins" it contains.
 
Artists are about the furthest away from "people who have it easy" in an industry as one can possibly get.
True. Artists do not have it easy. Neither do coal miners. Neither did the men who built and maintained steam engines or those who knew how to work the rigging of sailing vessels. Occupations will come and go as technology changes. The pace is accelerating and will not slacken in the foreseeable future. Yes, this is painful.

What I find interesting is that 20 years ago most people would have assumed it was the manual laborer who would be replaced first by technology. Some of that has happened, of course, but it turns out that robots may be more expensive than AI. This means a transition is also happening from the other end, where those who rely on their intellect and creativity may face transition sooner than those who work with their hands.

There will come a day when there are RPG's designed entirely by AI. If those games are better than your current favorite (don't think they wont be), will you play them?
 
As is true every time an industry changes, the people who used to have it easy claim the new way is not just hard for them but fundamentally wrong.

The people who used to have it hard in the effected industry often do the same. Few coal miners rejoiced when modern open-cut operations made older pits uneconomic and they all lost their jobs.
 
There will come a day when there are RPG's designed entirely by AI. If those games are better than your current favorite (don't think they wont be), will you play them?
I very much doubt those RPGs will be better, at least not without actual AI (as opposed to machine learning). And we're nowhere near that and probably won't be in our lifetimes.

Look at translation. Machine language translation has been around for ages and there's not a single program who can do that better than a skilled human. It gives a translation that's "good enough", not good.

Same applies here. Largely this is likely to affect the bottom end of the market.
 
I very much doubt those RPGs will be better, at least not without actual AI (as opposed to machine learning). And we're nowhere near that and probably won't be in our lifetimes.

In the time since I started this thread, I've seen the quality and accuracy of AI art improve at a seeming exponential rate. I'm no longer as confident as you that "Baldur's Gate: Eternal AI-Generated Adventure" isn't coming within a decade. What purpose will GMs serve when your group of friends can co-op their way through endlessly novel scenarios that can achieve 90% of the satisfaction a tabletop game can, plus better visuals and faster mechanics? Maybe the best of us GMs will find jobs 'feeding' and tuning the AI prompts, the way visual artists will be reduced to cleaning up details on AI commissions...
 
There will come a day when there are RPG's designed entirely by AI. If those games are better than your current favorite (don't think they wont be), will you play them?


No. Just like I will never knowingly purchase a product that utilizes AI art.
 
AI art is not art. It's just a bunch of numbers put together to create an image. For something to be art, an actual human has to have created it.
Sure, a lot of human art is just a bunch of paints splashed on to a canvas. Or someone took a shit into a jar and called it art. But at least, a human made it.
This is not an invitation to start a "what is art" debate. Because that's futile.

The development and improvement of these AI generators, is moving at a very fast pace. If you think, there's already a lot of garbage coming out of the various entertainment industries. Just wait, until these AI's becomes good enough to be used to mass produce entertainment media. It will be a literal tsunami of garbage.
 
If you think, there's already a lot of garbage coming out of the various entertainment industries. Just wait, until these AI's becomes good enough to be used to mass produce entertainment media. It will be a literal tsunami of garbage.

Or perhaps we will use this sort of technology to fragment what little is left of mass culture even further. We'll be able to retreat even deeper into niche genres, only surfacing once in a while to touch a small part of the mainstream, steal a single idea from it, and use that as a seed to fuel our AI-generated music station / one-man anime studio / neverending choose-your-own-adventure book, etc.
 
The part that makes me think RPGs are exponentially harder is they have to refer back and forth and to table contents often. I don't keep up on AI too much but I don't get the sense they are good at understanding a whole concept so much as working well in discreet areas.
So I could see each section looking perfect but consistency failing.
 
Well AI isn't "intelligence" or "understanding" as it stands with today's tech. Its just a set of rules along the lines of "if A is true then B". Now the rules can be complex, layered, interspersed with random number generation, etc., but its all that if-then once you're at the base. None of these AIs will ever "understand", they're just a complicated "clever hans" setup.

The current aporoach is to set up a set of rules to write other rules based on input and then run tons of data through that. You can, if you grok the subject and the rules, write meta rules or weight the rules created towards stuff that lead to a particular outcome. Example: an AI programming class got an assignment to write this style AI for a rules-simple game. Most teams ran on a "highest score move at this time", but the winners wrote in a bit that looked ahead several moves and set up it's stuff to a pattern that would produce big scores in the late game. The winning AI looked like it "understood" the game better, but really the students looked deeper at the game it played and came up with a meta rule to make the AI favor certain late game board patterns. At no point does the AI understand anything about the winning or scoring, just that moves to produce certain results are rated better than other moves based on the current board and the number of turns left in the game so play those high rated moves.
 
Well AI isn't "intelligence" or "understanding" as it stands with today's tech. Its just a set of rules along the lines of "if A is true then B". Now the rules can be complex, layered, interspersed with random number generation, etc., but its all that if-then once you're at the base. None of these AIs will ever "understand", they're just a complicated "clever hans" setup
Just a slight addition. Not for pedantry, but because I think it enhances your point. Many machine learning algorithms don't even decide things based on simple conditional logic, but rather purely correlations.

I attended an interesting talk concerning a machine learning AI which supposedly managed to correctly identify melanomas with an accuracy exceeding a typical dermatologist. However it turned out when melanomas are identified clinically a high resolution image of them is taken, where as benign skin conditions* are kept at low resolution. The machine learning algorithm simply correlated high resolution/pixel count with the token "melanoma". To some degree it wasn't even looking at the details of the images.

Just a good example I thought as back in 2018 there a sequence of news reports about AIs outperforming dermatologists, but later studies were much more skeptical (e.g. here). And in many cases the AI's company won't release the AI for replication testing of their claims.

*Any dermatologists/oncologists here feel free to correct with accurate terminology.
 
What I find interesting is that 20 years ago most people would have assumed it was the manual laborer who would be replaced first by technology. Some of that has happened, of course, but it turns out that robots may be more expensive than AI.

Well, they did the labour workers. That’s where we got “saboteurs” from. They didn’t do all of them and their attempts to replace hospitality workers are laughable. But most cars and computers are assembled hugely by machine and by hand for a little bit. Amazon is one of the companies doing best in machine intelligence and computer vision (understanding).

Artists have struggled to make money but let’s not compare it to people in a workhouse. Or miners. Doing what you love is sadly a privilege and not a right in our dystopia

The Singularity is here. Anything that can be replaced, will be replaced.


But it’s not all doom and gloom We have guitars assembled by factories and people still pay thousands for a Lowden.

just sucks if you ain’t George.
 
But it’s not all doom and gloom We have guitars assembled by factories and people still pay thousands for a Lowden.

My girlfriend said something reassuring about all this:
"I still row as a hobby even though there are much faster and more efficient ways of propelling boats these days."
 
My girlfriend said something reassuring about all this:
"I still row as a hobby even though there are much faster and more efficient ways of propelling boats these days."
As a liveaboard sailor, I appreciate this. I also don't necessarily hate on people with motorboats, cars, planes*

*maybe a little on the planes
 
Just a slight addition. Not for pedantry, but because I think it enhances your point. Many machine learning algorithms don't even decide things based on simple conditional logic, but rather purely correlations.

I attended an interesting talk concerning a machine learning AI which supposedly managed to correctly identify melanomas with an accuracy exceeding a typical dermatologist. However it turned out when melanomas are identified clinically a high resolution image of them is taken, where as benign skin conditions* are kept at low resolution. The machine learning algorithm simply correlated high resolution/pixel count with the token "melanoma". To some degree it wasn't even looking at the details of the images.

Just a good example I thought as back in 2018 there a sequence of news reports about AIs outperforming dermatologists, but later studies were much more skeptical (e.g. here). And in many cases the AI's company won't release the AI for replication testing of their claims.

*Any dermatologists/oncologists here feel free to correct with accurate terminology.
Correlation and pattern matching, with the decisions it makes that wow the company in the lab being near disastrous if applied to the real world. I can see it now. Dermatologist takes high-res images of moles he knows are benign so the AI flags it, and the surgery gets done. Retires early.
 
Even as an artist - I gotta say, I'm really loving using AI art. Give it a few more years and it's going to be truly magnificent.
 
Even as an artist - I gotta say, I'm really loving using AI art. Give it a few more years and it's going to be truly magnificent.
There's been so much toxicity and negativity about it that many companies are vocally distancing themselves from it.

I see lots of claims of harm....but I don't see where the harm actually is? (I remember the first time I was told that one of my books was on that trove site. I was - holy crap, they scanned it and uploaded it? Wow.)
 
There's been so much toxicity and negativity about it that many companies are vocally distancing themselves from it.

I see lots of claims of harm....but I don't see where the harm actually is? (I remember the first time I was told that one of my books was on that trove site. I was - holy crap, they scanned it and uploaded it? Wow.)
I'm pretty open to it... I like the fact that it can give creators a voice if you will for their own creations that they would not have been able to have before.

I would certainly see it as an extra tool in the 'tool box' so to speak.
 
Status
Not open for further replies.
Banner: The best cosmic horror & Cthulhu Mythos @ DriveThruRPG.com
Back
Top