AI is going to be our biggest game changer
This discussion has been closed.
Adding to Cart…
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
My issue with these is that, while they do look good at a glance or from afar, they are still essentially gobbledygook. The background is all just a mish mosh and you can't really make any one thing out.
Again:
What can AI do for me in completing my artistic process?
Where is that AI, that can give me photo real skin shaders, when I name a shader skin?
All I needed to add would be my favorite skin maps.
(Just a little bit like in that Reality to Luxrender shader converter, but AI supported.)
Hair shaders are a huge challenge obviously.
Where is that Ai, that understands how human joints work?
E.G. a AI driven skelleton, that is placed inside a human mesh. Place the joints in the correct positions, and tell the AI skelleton about the character's weight, muscular, fat and bone size parameters.
The AI could render the correct Mesh distortion for the characters joint.
I know, that Cinema 4D 6.** had already weight mapping and a kind of JCM like feature back in the year 2000.
Why haven't we much progressed since then - 23 years later?
Same could be true for quadrapeds as well.
I don't want AI to do art for me by giving it key words and then it spits out something,
I want it to support my art in improving the tools.
My main reason for not remixing my renders with AI is that I play with Daz to get things I can't get any other way. I want technicolor and bokeh grain, I want distance blur, raytraced lights and reflections. That requires raytraced rendering engine. When I want something drawn they I'm drawing it.
Different techniques and media for different results.
@Masterstroke
Adding secondary motion to a characters movement is very tedious.
the alleged “AI engine” in cascadeure does it for you, saving alot of time.
Of course, not everything works... here's another render I fed into the AI, and while I like some of the results, none of them gave me what I originally imagined the AI would do to it... So there is a learning curve about what will work well and what will not.
(Sure, I could do the stuff I want in Photoshop, but that's not the point of this exercise.)
Yeah, that would be awesome.
Love the third one, cool art style. Again I'm wondering though if you show this to someone and they go wow, you did that? And you go well... sort of? Maybe that's the litmus test. Will you be proud you made that at all? Will you feel embarrassed explaining that it was AI? And what will the reaction be? I'm assuming at that point it might devolve into a discussion about what awesome things technology can do these days, rather than an admiration of your art skills. In fact, I think most people will conclude at that point that you did nothing, even ignoring that there was a base render you fed into it.
to be fair I feel that way about my DAZ renders when people who don't know what 3D art is ask me if I made it
Here's a quick take from filmmakers (as opposed to artists who work for them): We don't care where the artwork comes from. Our endgame is to tell a story. If AI makes that easier, then the choice is academic. As an artist, I am a captive of my storytelling and I'll use whatever tool as a means to that end. Since I've become involved in using 3D to tell stories, I have noticed that there are two kinds of filmmakers. There are artists who make films and there are directors who make them. Often the difference between them is that many artists make films for other artists, while directors make films for the public at large. Artists' films emphasize their craftwork, while director's films emphasize the story. So in this sense, AI is viewed in opposing ways. For many artists, AI is a direct affront to their value in the filmmaking process. For directors, AI is another valuable employee that makes the process easier.
I think, the important thing about AI is, how much of the creative process is under your control.
A super A.I. is a make art button, but you don't have any control over it, so it just doesn't give you the satisfaction, that it is actually YOUR art.
On the other hand it is all yours, but you'll spend a hell of a time to make it right, really right.
A.I. could help you in a process where things might become to complex in order to be done in your little home studio.
(and keep in mind, some other 3d artist, might sneer about you, because you are using pre fabricated DAZ figures, that are even rigged for you.)
In the end it will be about finding the right balance.
Yes, I also like this third image:
Just wonder how much time, it would take to recreate such image directly in Daz Studio.
Nailed it!
That is totally fair. And to take that one step further, I don't think anyone would care much if our creation tools incorporated AI to help us with the arduous process of creating 3D art. I wonder if generating an image outright by telling an AI what you want isn't removing you from the creation process so much that you're literally not involved anymore. Hence, there can be no artistic satisfaction in having created anything.
I'm thinking of using my Daz budget and signing up for a 12 month subscription to Midjourney. I can creates thousands of high quality images versus the limitations of Daz. I produce art in a competitive environment and it's taken me 14 years to build a portfolio with 8,000 images. With Midjourney I can easily create 10,000 images in a year. Regardless, there will be limitations. I currently use live models and photographs to create composites with Daz renders in order to achieve the results I need. I may need to do the same with Midjourney renders. Either way, Daz is extremely limited not so much in the time it takes to do a render but the lack of available products. Take away the pretty faces and skimpware, there really is not a wide range of genres to work with.
Remember that the neural net-generated art is not going to be protected by copyright in itself. If it is composited with other work you may be OK, but get legal advice first.
My agency has looked into this and is allowing AI generated artwork under certain conditions. I believe there are lawsuits out there so things may be changing. The bad for someone like me who has spent decades producing art, now any idiot can create art better than a professional and generate tens of thousands of images to boot. The glut that's coming may destroy the individual contributor.
That.
And the blind crowd rushes headlong for the edge of the cliff.
I wonder if generating an image outright by telling an AI what you want isn't removing you from the creation process so much that you're literally not involved anymore. Hence, there can be no artistic satisfaction in having created anything.
Will be decided by the individual, not with the opinions or judgements of others.
many artists make films for other artists, while directors make films for the public at large.
I think this is dang close, but not correct or in need of correction.
You meant to say there's a segment of Filmmakers that make films [for] primarily to impress other filmmakers.
Filmmakers make films FOR studios. Studios [all owners implied] make films for a consumer base, which = audience.
-----------------
Why is a "Director" messing with 3D or AI?
Why is a "Novelist" making Book Covers?
Why is a "Filmmaker" doing Animation?
The truth is there is a spirit of DIY in all of this. The one-person FILMMAKER. The One-person Publisher. The one-person Whatever.
So the better line to draw in the sand is people who want to Do-It-Themselves and those that want to Outsource.
If AI is the new outsource then so be it.
The pride for much of this was the "I did it all by myself!".
That's where the credit comes from.
--------------------
Now the weird part for me has been this obsession with AI from NON-ARTISTS.
They keep showing up trying to give the bottom line and then drop the mic.
Even (most) artists are not against AI- they are against their work being used.
They are against competition from a THING they basically trained.
They are against NON-ARTISTS parading IN THEIR SPACES as Artists. (That's about as emotional as it gets)
-----------------------
We also know that NON-Artists//end consumers don't care (all that, that) much about where the art comes from.
They just want the end product. And the Pride is OWNERSHIP not AUTHORSHIP.
It's very difficult to argue with someone about outsourcing the one thing they are there to do.
It is and I would take no satisfaction in doing that. I would also not willingly let an AI touch my art after it's made. My work is my work, warts and all. If that makes me less competitive, that's fine. At least it's my work. Even my Daz work is my work because of what I've done to make it look the way I want it to look. If I haven't made all the assets, that doesn't actually matter because I end up adjusting, modifying, and changing everything. The final items are no longer what came out of the stock package. The lighting and composition are mine, too. Then add time for test renders and more adjustments. It's a lot more than typing in some text and hoping the AI does something desirable.
deleted
When an Ai does it, it just doesn't feel like MY work.
I believe Midjourney also makes your generated images and prompts visible to all other users, unless you pay a fee on top of the subscription. While the dataset it was trained on is not public, at least earlier versions were trained on LAION-5B. They also have a list of words you cannot use in the prompt and some sort of NSFW detector if you upload your own images as a base.
If you really need to use AI, maybe give Stable Diffusion a try first. It was also trained on the same "dirty" dataset, but at least it's open source and all your pictures would be private. In addition, it has plugins that let you edit the images. There were new plugins for SD released some days ago that would give some control. Things like coloring an already made sketch, generating a depth map from an input picture, or generating something using a pose reference.
As other people mentioned before, there are several ways to train a custom model on your own images with Stable Diffusion (it still depends on the base model that was trained on LAION-5B I believe). But, even with modifications it's really murky who owns the rights to those images.
Ironically, I have seen SD models trained on images generated with Midjourney to get the Midjourney look.
Before very long there will be be tools to direct and manipulate the Ai in exactly the same way as any art medium.
Deleted. I try to show what can be done by combining DAZ items and AI and I get insulted? Fine.
Nothing of this is your work except for the original render, though. That's the price paid for the easy mode - you are as much as an artist as something operating a washing mashine is a washing anything.
LOL, then so is every single person who purchases products from this site and simply renders images with them without also manually drawing or painting large portions of it. I would still do postwork and edits on any AI generations to make them my own, as I think all who use AI should.
All have lost the expression in your render (and the top-left seems to have taken his pecs breastwards).
That's what postwork is for, but apparently no one has any imagination around here anymore.
Everyone whos only using premade scenes and click presets, yes.
But you managed to eliminate even the lightning part because even this is now determined mostly by the AI and whatever was there on render is now masked under tons of machine generated brush strokes.