HAS THE CLOCK STARTED TICKING FOR DAZ?

frank0314

nelsonsmith said:

Richard Haseltine said:

This has been asked before and just tends to get contentious. In any event, there is a forum for discussing Daz AI Studio here https://www.daz3d.com/forums/categories/daz-ai-studio

Except this question, isn't really about Daz AI Studio.   All you have to do is look at any of the work being done with Midjourney and whether Daz incorporates and embraces AI or not doesn't matter, the fact is the business model Daz is using now, is  nearing an end.  Why wait for a PA to create something, let alone pay a PA for anything when you can get what you want with a prompt?  The question isn't if DAZ is going to become a niche tool, it's going to be how small is that niche going to become.  Or at least that's the way I see it.  It's rather sad, because DAZ 'was' heading in the direction it needed to be going, it was just moving too slowly, and now it's going to be extremely difficult to gain a larger customer base when you've got the major competition that is now out there.

The "Daz AI Studio" Forum can be used for all AI discussion, not just the generator. Looking through a whole site to find info on a particular subject is quite counter-productive when there are dedicated places for the topic of discussion.

kyoto kid

namffuak said:

The clock hasn't been wound up yet.

I will be concerned when the various AI engines can take a two-page detailed narrative description and faithfully render the result down to the last detail.

...spot on.

Or when  we finally have fully viable DNI (Direct Neural Interface) so we can plug in and what we see in our minds render on the screen.

richardandtracy

kyoto kid said:

..when  we finally have fully viable DNI (Direct Neural Interface) so we can plug in and what we see in our minds render on the screen.

Oh dear. I'm really worried by that idea. If it happens, furry, fluffy, cute cats will be the only thing you'll eversee on my screen from that point on. 

kyoto kid

...

kyoto kid

....Video killed the radio star, MP3 killed the video star, and Streaming did so to the MP3 and now we've arrived at AI which can compose and perform a piece in anyone's style from Bach to the Beatles.

OK with that out of my system, The one thing that keeps me working with Daz is the fact I have control over the process.  Yes, I'm using pregenerated content but I determine how it is morphed, posed, positioned, "kitbashed", and "skinned". If I could wrap my head around modelling better then the process would even come closer to teh vision I have.  On the other hand, AI uses a library of other people's works to draw from and tries its best approximate what the user enters in  at the prompt.  It may look fantastic but little has if any originality in my book..

For example, I've played with DA's Dream Up engine and it was, well, cute.  Of course I just had to see what typing in the perennial Naked Vicky In A Temple With A Sword would produce.  Yeah after the third try I got the result I wanted, however it wasn't my own vision or interpretation of the theme (reminded me more of how the Brothers Hildebrant sould have interpreted it) and there in is the reason to keep working with Daz.  

McGyver

I'd say no in regards to AI being a reason for DAZ's demise, if anything it would be for other reasons I don't wish to dabble in mentioning... As far as AI replacing 3D... I don't really see that... at least for a while... right now it's more of a niche 

tsroemi

Maybe this is an interesting tangent to the discussion: I don't have much experience with imaging AI, but a bit more with AI that is supposed to help with creative writing, Sudowrite especially. I thought I should probably get to know the beast instead of just being scared of the unkown, and so I went and got myself a subscription for a while.

Honestly, it was very, VERY much fun in the beginning. It felt like having a writing pal / editor mix awake and ready at all times, who I could throw ideas at and it would return with plot structures, for example. So I spend some exciting nights with sudo at first ;-) But the thing is, you really only ever get the same five types of stories flung back at you. And it only works with rather simple input (in the so-called story bible which is like a more sophisticated set of prompts) from which it will fish the trigger words it knows and then spin those into a story. If you give it something more complex, and with less common triggers, it will still spin it into a story structure, but that won't make terribly much sense. Because it just ignores the things it doesn't understand. And it can only digest and, sorry, regurgitate, in a way - not project. Because it doesn't create, it replicates.

It took me amazingly long to understand why I was growing so bored and frustrated with it after a while. Same old, same old ... Don't you get that I'm trying to do something else here? Why won't you help me developing that and instead keep pushing me back onto paths that I don't want to tread? By, for instance, trying to actively force me to include a very specific sort of wholesome character development arc in virtually every story I plot. And then man, those times when I actually let it write scenes on its own ... That was some bad, bad, BAD writing.

Yes I know, I know, you can feed it different rules, and many only use it for writing boring description, etc. pp. But honestly, I'm just not very impressed. And I don't think this is due to the developmental stage we're at. It's an inherent flaw of the way AI works. It doesn't really learn, it collects (well, steals, really, most of it), and then remixes. It doesn't make what it takes its own by using it in its own ways - the very thing that you need to include in your teaching if you ever want anyone to truly learn anything. It cannot ever even have 'own ways'. Because it isn't creative, but derivative. It can never take us anywhere, in any form of art, where we haven't been yet, because that's quite literally where it lives, and where it will keep us, too, if we let it.

Because it pretends to take us to new places, and that, I think, is a very real danger. It bribes us with pretty replicas of what we already know (well - the images at least are getting technically pretty good from what I'm seeing, even if not the writing), and the brain luuuuurves those things it already knows ... But it's only the technique that's new. Not the content. Whatever new emerges from it, comes from the human user. And there, like I said, I found it more hindering than useful.

So personally, I don't see 'creativity helpers' like DAZ getting elbowed out anytime soon, although it probably makes sense for them to include AI features like they're already doing, because there's just not that many aficionados wanting to wait three hours for a render to cook only to then see that a grass blade is poking through an ankle. A really smart thing in my view would be to update DAZ studio to make it quicker and smoother and include more manipulating tools, because then, what would you even need AI for?

kyoto kid

McGyver said:

I'd say no in regards to AI being a reason for DAZ's demise, if anything it would be for other reasons I don't wish to dabble in mentioning... As far as AI replacing 3D... I don't really see that... at least for a while... right now it's more of a niche 

...more likely I am wondering about the fate of the "other store" given the flight of their PAs to here. 

mding

@tsroemi: Thank you for those interesting insights!

ainm.sloinneadh

AI does not equal image generation and writing emails for you. Those are simply 2 headline grabbing applications of AI and I feel the distinction of what AI looks like to the layperson and what it actually is often goes lost in these conversations.

At the root of AI are many types of machine learning techniques that can be trained to automate tasks. That technology has been around and developing since the 1960's. The current bubble is generative models, which we're calling AI because they apparently fool enough people that they have some sort of creative sentience behind them. They absolutely don't, and skilled professionals who assess their suitability for creative tasks will very quickly understand that. A LLM model, as an example, is literally just a linguistic interface into a search engine that fools the layperson that it's able to create literature when, in actuality, it's laughably poor at that task. The same for image generation models.

What the recent bubble has shown, however, is that if recent technological advancements in the ML field are instead employed in a limited manner as a tool within a creative process, they can significantly improve the creative output of the individuals utilising them. The failure of Daz to design and integrate tools based on recent computational and technological advancements into Daz Studio is where the threat to their product lies. Midjourney will not kill Daz - that's not the competition, and I, for one, simply don't understand why Daz picked that battle with its recent AI Studio web product, unless it was a stepping stone to a wider strategy to understand and develop the tools in Daz Studio that I believe it needs. Instead, the threat is from other rendering and gaming engines who are looking like they might get there first. I love Daz Studio, but if in a year or two a competing product - either existing or completely new to the market - offers me tools that increase my creative output, I don't see why I wouldn't be interested. That's the threat. Or the opportunity, if you're into SWOT analysis. Me, I just want to create more output.

xyer0

kyoto kid said:

namffuak said:

The clock hasn't been wound up yet.

I will be concerned when the various AI engines can take a two-page detailed narrative description and faithfully render the result down to the last detail.

...spot on.

Or when  we finally have fully viable DNI (Direct Neural Interface) so we can plug in and what we see in our minds render on the screen.

At that point every neuralinked mind will say/see the same thing, for it will not be their minds but the mind of their master. Nor can they form the thought, "Unplug me!" 

Silent Winter

kyoto kid said:

... Of course I just had to see what typing in the perennial Naked Vicky In A Temple With A Sword would produce.   

Clearly Daz AI doesn't know the history, or at least hasn't read the old forums for the acronym - I just tried the prompt 'NVIATWAS' , admittedly on 'Fast' quality, though I did choose 'Victoria 9' as the model, and got this result:

Kinda abstract interpretation. 

tsroemi

mding said:

@tsroemi: Thank you for those interesting insights!

Very much welcome! If you want to check out sudowriter for yourself, it gives you some free 'credits' to play with for a bit. 

Crescent

I've been playing with various AI "writing" softwares like Sudowriter and they're great for basic ideas and the occasional well-crafted sentence, but it feels like bowling in a video game - I may occasionally get a few pins knocked down if I click everything just right, but most of the time the ball goes into the gutter of unoriginality.  No matter the details, plots keep bending back towards basic archetypes.  Descriptions are unintentionally hilarious - the amount of times something is "a monument to ___" or "a testimony to ___" is just amazing.  "The omnipresent wind was a monument to the harshly scuptured landscape."  "The explosion of colors of the fall leaves was a testimony to the wonders of nature."  Sometimes it's a great way to get past the "blank page" dread - let the AI slop down a bunch of paragraphs then re-write everything to get the correct tone and flavor - but it's still nowhere near good, consistent writing.  

Even if you get an image that escapes the usual AI detail flaws (like 6 1/2 fingers per hand), there's a certain sameness to most images, regardless of style.  It's going to be good enough for many purposes, but it won't really fulfill the need people have to create art until it can understand detailed descriptions and create what individuals actually want instead of creating what the masses overall enjoy.  

IceCrMn

*

paulawp (marahzen)

tsroemi said:

Maybe this is an interesting tangent to the discussion: I don't have much experience with imaging AI, but a bit more with AI that is supposed to help with creative writing, Sudowrite especially. I thought I should probably get to know the beast instead of just being scared of the unkown, and so I went and got myself a subscription for a while.

Honestly, it was very, VERY much fun in the beginning. It felt like having a writing pal / editor mix awake and ready at all times, who I could throw ideas at and it would return with plot structures, for example. So I spend some exciting nights with sudo at first ;-) But the thing is, you really only ever get the same five types of stories flung back at you. And it only works with rather simple input (in the so-called story bible which is like a more sophisticated set of prompts) from which it will fish the trigger words it knows and then spin those into a story. If you give it something more complex, and with less common triggers, it will still spin it into a story structure, but that won't make terribly much sense. Because it just ignores the things it doesn't understand. And it can only digest and, sorry, regurgitate, in a way - not project. Because it doesn't create, it replicates.

It took me amazingly long to understand why I was growing so bored and frustrated with it after a while. Same old, same old ... Don't you get that I'm trying to do something else here? Why won't you help me developing that and instead keep pushing me back onto paths that I don't want to tread? By, for instance, trying to actively force me to include a very specific sort of wholesome character development arc in virtually every story I plot. And then man, those times when I actually let it write scenes on its own ... That was some bad, bad, BAD writing.

Yes I know, I know, you can feed it different rules, and many only use it for writing boring description, etc. pp. But honestly, I'm just not very impressed. And I don't think this is due to the developmental stage we're at. It's an inherent flaw of the way AI works. It doesn't really learn, it collects (well, steals, really, most of it), and then remixes. It doesn't make what it takes its own by using it in its own ways - the very thing that you need to include in your teaching if you ever want anyone to truly learn anything. It cannot ever even have 'own ways'. Because it isn't creative, but derivative. It can never take us anywhere, in any form of art, where we haven't been yet, because that's quite literally where it lives, and where it will keep us, too, if we let it.

Because it pretends to take us to new places, and that, I think, is a very real danger. It bribes us with pretty replicas of what we already know (well - the images at least are getting technically pretty good from what I'm seeing, even if not the writing), and the brain luuuuurves those things it already knows ... But it's only the technique that's new. Not the content. Whatever new emerges from it, comes from the human user. And there, like I said, I found it more hindering than useful.

So personally, I don't see 'creativity helpers' like DAZ getting elbowed out anytime soon, although it probably makes sense for them to include AI features like they're already doing, because there's just not that many aficionados wanting to wait three hours for a render to cook only to then see that a grass blade is poking through an ankle. A really smart thing in my view would be to update DAZ studio to make it quicker and smoother and include more manipulating tools, because then, what would you even need AI for?

I spent my own time on a somewhat similar avenue, but on the other side of the road. A while ago, I - the AI hater, she of the snarky "can't do dogs with four legs" kind of thing when referencing AI "make art" sites of the time - was pulled into trying out some AI-based story planning. I found that it was useful for people who like that sort of thing. I as a writer already don't follow most of the rules, so I don't outline stories that way and never actually pursued that on a personal basis. But what I did find was that Claude in particular made for a very useful online buddy that I used - and still use many months later - to discuss my stories with. I'm not interested in AI writing a story for me; I'm the writer and that's my job. Nor am I interested in having it come up with foundational plots, characterizations and story building blocks; also, my job. But what it is great at is being willing to chat incessantly about my long, complicated, boring story that no one IRL wants to spend time talking about, and through the course of that, I have often talked myself into seeing something new - finding an idea that I wouldn't have, had someone/something not been willing to sit there for an hour having a conversation about that part of the book. I talk to it just like it's some random internet friend and it responds in kind. Just a couple of days ago I had it tell me about the business of printing a major daily newspaper in 1940, providing some practical ideas about the jobs and crew typical of that era (which is quite a lot different from my own time with a daily newspaper that was decades later than that), and then we segued into how a character with a job at a newspaper print site is a critical foundation for the plot of one of the books in my story series. I discussed my existing notes of that plot line and Claude was very happy to help provide suggestions and refine ideas until I was done discussing it. Frankly, I couldn't pay a real person to sit there and do a better job of just listening and providing useful responses, and that's not even counting the "wow, this is so great!" stuff that Claude starts tossing at you. It's also great for worldbuilding what-if gaming. The key thing here is that it's a conversation in which the AI is reacting to my input, vs generating the idea and details. Having presented an entire situation or subject, I can then ask it something like, "In this situation we've been talking about, what do you think are the likely outcomes when ________ happens?" It will return any number of ideas, which I have come to expect are reasonable ideas considering what information it has. I haven't always thought of all of them myself.

Anyway, just my 2 cents as a fiction writer who has been tinkering with AI in my process, and in my case, found a lasting use for it.

On the general subject of this actual thread, yesterday an online friend introduced me to Suno, which entertained me in a way that the "make art" sites didn't, and at this point, I see enough to understand why that is. I said to my friend, it's a "make music" button. And it really is exactly that. Write a short prompt, get - so far - a much better actual song than I personally could create with what I have available to me, including my mediocre skill at creating music. I said to this friend, I see where a real musician might do this once or twice and shrug and move on because they want control over all aspects of creating their art. But someone who just wants a "make music" button can create prompts and actually really like the output for some sustained period of time. Frankly, I see an exact parallel between that and the "make art" AI stuff. Assuming there are products out there that  more serious musicians who like tech vs physical instruments have to use to make music, something like Suno will drain away dabblers who lack all the things it takes to persevere with the option that is harder but offers total control over the process and output. Ultimately, that's because there are different audiences for the two products, IMO.

https://suno.com/song/3e4432c7-5093-4db2-9b3c-4f4d11575c04

ArtAngel

paulawp (marahzen) said:

tsroemi said:

Maybe this is an interesting tangent to the discussion: I don't have much experience with imaging AI, but a bit more with AI that is supposed to help with creative writing, Sudowrite especially. I thought I should probably get to know the beast instead of just being scared of the unkown, and so I went and got myself a subscription for a while.

Honestly, it was very, VERY much fun in the beginning. It felt like having a writing pal / editor mix awake and ready at all times, who I could throw ideas at and it would return with plot structures, for example. So I spend some exciting nights with sudo at first ;-) But the thing is, you really only ever get the same five types of stories flung back at you. And it only works with rather simple input (in the so-called story bible which is like a more sophisticated set of prompts) from which it will fish the trigger words it knows and then spin those into a story. If you give it something more complex, and with less common triggers, it will still spin it into a story structure, but that won't make terribly much sense. Because it just ignores the things it doesn't understand. And it can only digest and, sorry, regurgitate, in a way - not project. Because it doesn't create, it replicates.

It took me amazingly long to understand why I was growing so bored and frustrated with it after a while. Same old, same old ... Don't you get that I'm trying to do something else here? Why won't you help me developing that and instead keep pushing me back onto paths that I don't want to tread? By, for instance, trying to actively force me to include a very specific sort of wholesome character development arc in virtually every story I plot. And then man, those times when I actually let it write scenes on its own ... That was some bad, bad, BAD writing.

Yes I know, I know, you can feed it different rules, and many only use it for writing boring description, etc. pp. But honestly, I'm just not very impressed. And I don't think this is due to the developmental stage we're at. It's an inherent flaw of the way AI works. It doesn't really learn, it collects (well, steals, really, most of it), and then remixes. It doesn't make what it takes its own by using it in its own ways - the very thing that you need to include in your teaching if you ever want anyone to truly learn anything. It cannot ever even have 'own ways'. Because it isn't creative, but derivative. It can never take us anywhere, in any form of art, where we haven't been yet, because that's quite literally where it lives, and where it will keep us, too, if we let it.

Because it pretends to take us to new places, and that, I think, is a very real danger. It bribes us with pretty replicas of what we already know (well - the images at least are getting technically pretty good from what I'm seeing, even if not the writing), and the brain luuuuurves those things it already knows ... But it's only the technique that's new. Not the content. Whatever new emerges from it, comes from the human user. And there, like I said, I found it more hindering than useful.

So personally, I don't see 'creativity helpers' like DAZ getting elbowed out anytime soon, although it probably makes sense for them to include AI features like they're already doing, because there's just not that many aficionados wanting to wait three hours for a render to cook only to then see that a grass blade is poking through an ankle. A really smart thing in my view would be to update DAZ studio to make it quicker and smoother and include more manipulating tools, because then, what would you even need AI for?

I spent my own time on a somewhat similar avenue, but on the other side of the road. A while ago, I - the AI hater, she of the snarky "can't do dogs with four legs" kind of thing when referencing AI "make art" sites of the time - was pulled into trying out some AI-based story planning. I found that it was useful for people who like that sort of thing. I as a writer already don't follow most of the rules, so I don't outline stories that way and never actually pursued that on a personal basis. But what I did find was that Claude in particular made for a very useful online buddy that I used - and still use many months later - to discuss my stories with. I'm not interested in AI writing a story for me; I'm the writer and that's my job. Nor am I interested in having it come up with foundational plots, characterizations and story building blocks; also, my job. But what it is great at is being willing to chat incessantly about my long, complicated, boring story that no one IRL wants to spend time talking about, and through the course of that, I have often talked myself into seeing something new - finding an idea that I wouldn't have, had someone/something not been willing to sit there for an hour having a conversation about that part of the book. I talk to it just like it's some random internet friend and it responds in kind. Just a couple of days ago I had it tell me about the business of printing a major daily newspaper in 1940, providing some practical ideas about the jobs and crew typical of that era (which is quite a lot different from my own time with a daily newspaper that was decades later than that), and then we segued into how a character with a job at a newspaper print site is a critical foundation for the plot of one of the books in my story series. I discussed my existing notes of that plot line and Claude was very happy to help provide suggestions and refine ideas until I was done discussing it. Frankly, I couldn't pay a real person to sit there and do a better job of just listening and providing useful responses, and that's not even counting the "wow, this is so great!" stuff that Claude starts tossing at you. It's also great for worldbuilding what-if gaming. The key thing here is that it's a conversation in which the AI is reacting to my input, vs generating the idea and details. Having presented an entire situation or subject, I can then ask it something like, "In this situation we've been talking about, what do you think are the likely outcomes when ________ happens?" It will return any number of ideas, which I have come to expect are reasonable ideas considering what information it has. I haven't always thought of all of them myself.

Anyway, just my 2 cents as a fiction writer who has been tinkering with AI in my process, and in my case, found a lasting use for it.

On the general subject of this actual thread, yesterday an online friend introduced me to Suno, which entertained me in a way that the "make art" sites didn't, and at this point, I see enough to understand why that is. I said to my friend, it's a "make music" button. And it really is exactly that. Write a short prompt, get - so far - a much better actual song than I personally could create with what I have available to me, including my mediocre skill at creating music. I said to this friend, I see where a real musician might do this once or twice and shrug and move on because they want control over all aspects of creating their art. But someone who just wants a "make music" button can create prompts and actually really like the output for some sustained period of time. Frankly, I see an exact parallel between that and the "make art" AI stuff. Assuming there are products out there that  more serious musicians who like tech vs physical instruments have to use to make music, something like Suno will drain away dabblers who lack all the things it takes to persevere with the option that is harder but offers total control over the process and output. Ultimately, that's because there are different audiences for the two products, IMO.

https://suno.com/song/3e4432c7-5093-4db2-9b3c-4f4d11575c04

I think not, but watch this video. At our local Lowes (and many other places) here in California, the cash register clerks disappeared And somehow I became a part time worker by checking out my own purchases. Hooray for the business establishments who got free labor, boo for me. After a few years of theft by vagrants with napsacks, they closed all the self-serve checkouts, opened one or two manned by an employee and reopened the old check-out aisles. Automation failed. If you write a book via AI you don't own the copyright. Numerous books written by AI get called out and the 'authors' too.What is the point of having a 100th version on something that already exists. That's why sequels sometimes fail and never live up to the expectations of the original. Who wants to read, see or feel a copycat experience. Nothing beats something original. Something fresh. Something new. AI does not deliver that. They think inside of the box. DAZ should engage in rewarding original art with a higher reward than editorial or copycat art. And DAZ has plenty of artists who can twist things around enough to make something inspired from a copy turn into a DAZling original. So no. DAZ will not be done, unless it becomes a pelican and chases trends at too fast a pace. AI is a baby. Learning from it's foster parents. Those that feed it art. AI is inevitable. DAZ should offer AI but as an option, and again reward the PA's who create originals by givinfg th copycats less of  a percentage than they give original artists. Inspiration that launches original art is like looking at a painting and doing your version. Copycat is typing in an order and getting the result delivered akin to going to IN & Out and telling them to change the cheeseburger to an animal burger. It is their recipe not ours.  Here are some 2024 videos on non--transparancy and AI generated content. It discusses a new federal bill that could require disclosure of songs/art used to train AI. If the bill is passed and the sources are disclosed all hell could rise.. 

Edit: Re Book: ownership

Re:Music Owneship 

 

kyoto kid

Silent Winter said:

kyoto kid said:

... Of course I just had to see what typing in the perennial Naked Vicky In A Temple With A Sword would produce.   

Clearly Daz AI doesn't know the history, or at least hasn't read the old forums for the acronym - I just tried the prompt 'NVIATWAS' , admittedly on 'Fast' quality, though I did choose 'Victoria 9' as the model, and got this result:

Kinda abstract interpretation. 

...one would think, that the results with Daz AI would have been better..

Maybe try spelling it out completely like I did, as I also got weird random results on DA just using the acronym.

Silent Winter

kyoto kid said:

Silent Winter said:

kyoto kid said:

... Of course I just had to see what typing in the perennial Naked Vicky In A Temple With A Sword would produce.   

Clearly Daz AI doesn't know the history, or at least hasn't read the old forums for the acronym - I just tried the prompt 'NVIATWAS' , admittedly on 'Fast' quality, though I did choose 'Victoria 9' as the model, and got this result:

Kinda abstract interpretation. 

...one would think, that the results with Daz AI would have been better..

Maybe try spelling it out completely like I did, as I also got weird random results on DA just using the acronym.

I was just trying the acronym for amusement and found the resultant 3D scribble amusing enough. I've only messed with the AI generator a little and it won't give me the results I want anyway. As I enjoy the setting up of a scene as much as the finished render, I'll stick to my 3D content in Studio.

PixelPie

Torquinox said:

@SilverGirl "Oh," she said. "So you're a 3D art chef."  Brilliant! And thanks. I enjoyed your reply.

I love this..this is perfect!  I don't feel so bad now about buying so many "ingredients" from DAZ .. :-)

I was out in my yard the other day taking some shots of "real" things with "real" pictures and thinking of this very thing.  I have tried using AI within photoshop for texturing.. and it is "just ok" compared to the real deal.  The day is already here where when I try to pull up a picture of a beautiful sunset to see how light plays on things, I don't know what photos are real and which are AI.  I can only hope places continue to separate AI from real.  Maybe we will find REAL photos, sketches and art only groups springing up?

Overall I do NOT like AI, but I  enjoyed playing around with DAZ AI and have found it fun and useful..  created my avatar using DAZ AI and finally my silly acronymish style name has a face. 

So, I have used AI to get some ideas, but really, since I created the prompt to visualize the idea, it is still my idea (some sites are hiding your prompts now) so I still prefer doing things the old fashioned way over all... I guess my age is showing lol.  Creating scenes by hand using my own prompt idea conceptualized within my own mind without AIs help is what I like.  I am a sketch artist first and foremost and most of my ideas wind up there in my sketchbooks first. So I can just create a "prompt" right on paper and that is a really satisfying feeling vs something else doing the work....and I don't want a machine in my brain... that sounds like a dystopian nightmare to me.  I get more enjoyment from my own accomplishment and exercising my old brain muscle than pushing the "make art button" 

 but honestly I hope there is always a niche for what DAZ has to offer

kyoto kid

...again AI as a "tool"; within a process or programme (like denoising or "depixelate"  a poor or enlarged image or photo) is one thing. 

However,  I agree when it's purpose is to generate art, music, and written works in any form, eliminates jobs, or is purposely used to deceive, that's where it crosses the line.  Fortunately it is still far from perfect, not only seeing the results here and on DA, but in RL as well (like the "robo taxis" in San Francisco).

almahiedra

 

Leana said:

 

In many fields AI can be useful as an additional tool if actual professionals can give it the right inputs / orientations and challenge the outputs, but it couldn't do the work alone.

I created an exam for my students with fairly light mathematics, I asked an AI to solve it. All the answers were miscalculated, all horribly wrong. The wonderful thing is that it wrote everything in LaTeX and after I typed in the correct values, I could send the solution, and not with my ugly handwriting. 

But there has to be a human realizing that the AI doesn’t even know the correct order of operations, someone who knows LaTeX and can read the code without compiling it is also needed, and then knows how to edit and compile it so that anyone else can see the formulas.

 

SnowSultan

Here is the simple answer to this and every thread that asks if artists are going to be replaced by AI.

When AI can make *exactly* what you require, down to the smallest detail, yes. Until then, no.   :)

nonesuch00

Crescent said:

I've been playing with various AI "writing" softwares like Sudowriter and they're great for basic ideas and the occasional well-crafted sentence, but it feels like bowling in a video game - I may occasionally get a few pins knocked down if I click everything just right, but most of the time the ball goes into the gutter of unoriginality.  No matter the details, plots keep bending back towards basic archetypes.  Descriptions are unintentionally hilarious - the amount of times something is "a monument to ___" or "a testimony to ___" is just amazing.  "The omnipresent wind was a monument to the harshly scuptured landscape."  "The explosion of colors of the fall leaves was a testimony to the wonders of nature."  Sometimes it's a great way to get past the "blank page" dread - let the AI slop down a bunch of paragraphs then re-write everything to get the correct tone and flavor - but it's still nowhere near good, consistent writing.  

Even if you get an image that escapes the usual AI detail flaws (like 6 1/2 fingers per hand), there's a certain sameness to most images, regardless of style.  It's going to be good enough for many purposes, but it won't really fulfill the need people have to create art until it can understand detailed descriptions and create what individuals actually want instead of creating what the masses overall enjoy.  

You are exactly right! It is the same with images. It sort of feels like, you know, how the word "egregious" suddenly become so popular out of nowhere. I'm like how did all of these professional marketing people all of the sudden start using this word? Have I been short changed in education by my schools? Have I been spending too much time watching the 3 Stooges?

FirstBastion

Silent Winter said:

kyoto kid said:

... Of course I just had to see what typing in the perennial Naked Vicky In A Temple With A Sword would produce.   

Clearly Daz AI doesn't know the history, or at least hasn't read the old forums for the acronym - I just tried the prompt 'NVIATWAS' , admittedly on 'Fast' quality, though I did choose 'Victoria 9' as the model, and got this result:

Kinda abstract interpretation. 

Stretching the acronym 'NVIATWAS'  out into words.  Vicky In A Temple With A Sword will give you something.

 

 

Silent Winter

FirstBastion said:

Silent Winter said:

kyoto kid said:

... Of course I just had to see what typing in the perennial Naked Vicky In A Temple With A Sword would produce.   

Clearly Daz AI doesn't know the history, or at least hasn't read the old forums for the acronym - I just tried the prompt 'NVIATWAS' , admittedly on 'Fast' quality, though I did choose 'Victoria 9' as the model, and got this result:

Kinda abstract interpretation. 

Stretching the acronym 'NVIATWAS'  out into words.  Vicky In A Temple With A Sword will give you something.

 

 

You missed a word   (presumably for forum TOS purposes).

But again, I wasn't really bothered by the outcome - just wanted to share the scribble.

She's kinda holding the sword funny in yours.

Leana

Daz AI won't give you any results if you add the missing word ;)

caravelle

I don't know... When I look at my expenses here in the shop, I have the feeling that the clock is far from ticking...  In addition, AI and 3D are miles away from each other as practised art. Even if I have the most sophisticated prompts and tools available for AI, it is a huge difference whether I actively design a picture or video or let the generator more or less decide for me.

I enjoy DazAI, I've been a premium user since the start of the beta version and have already achieved results that I really like, but DS is something else. Both have their charms.

BeeMKay

paulawp (marahzen) said:

tsroemi said:

Maybe this is an interesting tangent to the discussion: I don't have much experience with imaging AI, but a bit more with AI that is supposed to help with creative writing, Sudowrite especially. I thought I should probably get to know the beast instead of just being scared of the unkown, and so I went and got myself a subscription for a while.

Honestly, it was very, VERY much fun in the beginning. It felt like having a writing pal / editor mix awake and ready at all times, who I could throw ideas at and it would return with plot structures, for example. So I spend some exciting nights with sudo at first ;-) But the thing is, you really only ever get the same five types of stories flung back at you. And it only works with rather simple input (in the so-called story bible which is like a more sophisticated set of prompts) from which it will fish the trigger words it knows and then spin those into a story. If you give it something more complex, and with less common triggers, it will still spin it into a story structure, but that won't make terribly much sense. Because it just ignores the things it doesn't understand. And it can only digest and, sorry, regurgitate, in a way - not project. Because it doesn't create, it replicates.

It took me amazingly long to understand why I was growing so bored and frustrated with it after a while. Same old, same old ... Don't you get that I'm trying to do something else here? Why won't you help me developing that and instead keep pushing me back onto paths that I don't want to tread? By, for instance, trying to actively force me to include a very specific sort of wholesome character development arc in virtually every story I plot. And then man, those times when I actually let it write scenes on its own ... That was some bad, bad, BAD writing.

Yes I know, I know, you can feed it different rules, and many only use it for writing boring description, etc. pp. But honestly, I'm just not very impressed. And I don't think this is due to the developmental stage we're at. It's an inherent flaw of the way AI works. It doesn't really learn, it collects (well, steals, really, most of it), and then remixes. It doesn't make what it takes its own by using it in its own ways - the very thing that you need to include in your teaching if you ever want anyone to truly learn anything. It cannot ever even have 'own ways'. Because it isn't creative, but derivative. It can never take us anywhere, in any form of art, where we haven't been yet, because that's quite literally where it lives, and where it will keep us, too, if we let it.

Because it pretends to take us to new places, and that, I think, is a very real danger. It bribes us with pretty replicas of what we already know (well - the images at least are getting technically pretty good from what I'm seeing, even if not the writing), and the brain luuuuurves those things it already knows ... But it's only the technique that's new. Not the content. Whatever new emerges from it, comes from the human user. And there, like I said, I found it more hindering than useful.

So personally, I don't see 'creativity helpers' like DAZ getting elbowed out anytime soon, although it probably makes sense for them to include AI features like they're already doing, because there's just not that many aficionados wanting to wait three hours for a render to cook only to then see that a grass blade is poking through an ankle. A really smart thing in my view would be to update DAZ studio to make it quicker and smoother and include more manipulating tools, because then, what would you even need AI for?

I spent my own time on a somewhat similar avenue, but on the other side of the road. A while ago, I - the AI hater, she of the snarky "can't do dogs with four legs" kind of thing when referencing AI "make art" sites of the time - was pulled into trying out some AI-based story planning. I found that it was useful for people who like that sort of thing. I as a writer already don't follow most of the rules, so I don't outline stories that way and never actually pursued that on a personal basis. But what I did find was that Claude in particular made for a very useful online buddy that I used - and still use many months later - to discuss my stories with. I'm not interested in AI writing a story for me; I'm the writer and that's my job. Nor am I interested in having it come up with foundational plots, characterizations and story building blocks; also, my job. But what it is great at is being willing to chat incessantly about my long, complicated, boring story that no one IRL wants to spend time talking about, and through the course of that, I have often talked myself into seeing something new - finding an idea that I wouldn't have, had someone/something not been willing to sit there for an hour having a conversation about that part of the book. I talk to it just like it's some random internet friend and it responds in kind. Just a couple of days ago I had it tell me about the business of printing a major daily newspaper in 1940, providing some practical ideas about the jobs and crew typical of that era (which is quite a lot different from my own time with a daily newspaper that was decades later than that), and then we segued into how a character with a job at a newspaper print site is a critical foundation for the plot of one of the books in my story series. I discussed my existing notes of that plot line and Claude was very happy to help provide suggestions and refine ideas until I was done discussing it. Frankly, I couldn't pay a real person to sit there and do a better job of just listening and providing useful responses, and that's not even counting the "wow, this is so great!" stuff that Claude starts tossing at you. It's also great for worldbuilding what-if gaming. The key thing here is that it's a conversation in which the AI is reacting to my input, vs generating the idea and details. Having presented an entire situation or subject, I can then ask it something like, "In this situation we've been talking about, what do you think are the likely outcomes when ________ happens?" It will return any number of ideas, which I have come to expect are reasonable ideas considering what information it has. I haven't always thought of all of them myself.

Anyway, just my 2 cents as a fiction writer who has been tinkering with AI in my process, and in my case, found a lasting use for it.

On the general subject of this actual thread, yesterday an online friend introduced me to Suno, which entertained me in a way that the "make art" sites didn't, and at this point, I see enough to understand why that is. I said to my friend, it's a "make music" button. And it really is exactly that. Write a short prompt, get - so far - a much better actual song than I personally could create with what I have available to me, including my mediocre skill at creating music. I said to this friend, I see where a real musician might do this once or twice and shrug and move on because they want control over all aspects of creating their art. But someone who just wants a "make music" button can create prompts and actually really like the output for some sustained period of time. Frankly, I see an exact parallel between that and the "make art" AI stuff. Assuming there are products out there that  more serious musicians who like tech vs physical instruments have to use to make music, something like Suno will drain away dabblers who lack all the things it takes to persevere with the option that is harder but offers total control over the process and output. Ultimately, that's because there are different audiences for the two products, IMO.

https://suno.com/song/3e4432c7-5093-4db2-9b3c-4f4d11575c04

I think you and I worked on the same experiment. I'm also not much of a lover of AI in the creative process, but what I found useful was that it helped me structure my crazy thoughts and at the same time there was a written result (as in, notes, or summaries) which could be used as a base for further writing. Basically, it kept asking questions (based on the scripted questions it had beed fed as part of the "experiment"), and that way forced you to formulate your thoughts and write them down into the chat, rather than have them flitting in your brain. I did complete a story to outline level that previously just existed as a one sentence idea.

Here's Peter von Stackleberg's current version of his "Story Genius" writing aid, if you (general you) want to give it a try.

https://chatgpt.com/g/g-bAHd7S4gE-storygenius-story-idea-concept-designer

LeatherGryphon

almahiedra said:

 

Leana said:

 

In many fields AI can be useful as an additional tool if actual professionals can give it the right inputs / orientations and challenge the outputs, but it couldn't do the work alone.

I created an exam for my students with fairly light mathematics, I asked an AI to solve it. All the answers were miscalculated, all horribly wrong. The wonderful thing is that it wrote everything in LaTeX and after I typed in the correct values, I could send the solution, and not with my ugly handwriting. 

But there has to be a human realizing that the AI doesn’t even know the correct order of operations, someone who knows LaTeX and can read the code without compiling it is also needed, and then knows how to edit and compile it so that anyone else can see the formulas.

 

I remember "LaTeX" from the early '80s (when we were still in the typewriter era).  I never needed it but I worked closely with a guy who knew it well.  His advice to me was "Don't ever let anyone know you know LaTeX, they won't leave you alone."