AI is forcing the hand of the Demoscene.
category: general [glöplog]
Quote:
Seems people like that there is a human behind the art and that the art has a backstory of some kind.
Quote:
Art needs context. Without context it’s a mere decor.
————————————————————
Quote:
Latest LLMs can use up to 1 million tokens in a prompt. So you could actually provide very elaborate context. Maybe it doesn't work yet so well in current gen image/video generators, but it's coming.
So you claim that in these million tokens I can provide the LLM with context of Van Gogh’s life. His feelings. His sense of rejection. His relationship with Theo. Not only to have it sense the accumulated feelings and thoughts of a lifetime, but also make it understand this context. That’s what was meant by “context” and “human backstory”. I don’t think even tech bros pushing this tech are this brave, this eager to market this cr*p.
Quote:
I would differentiate between popular art and more ambitious art sometimes called high-art.
According to my ad-hoc definition, popular art is artistic content deliberately created to reach the largest audience possible, while high-art is more about pushing the boundaries of creation with the usual audience assumed to be educated in art history and sophisticated enough to understand the depth of the creation, context etc...
This doesn't really mean any of the kind is higher or lower quality.
In fact, popular art has pretty high quality bar, because quality helps to reach widest audience possible. However, it is not as demanding when it comes to originality. For example, it's quite common that very generic song gets to the top charts,because the singer has already a strong brand. This is where AI could potentially shine (if done right). It could create high-quality, but pretty generic content. Maybe useful in gamedev too to cut cost.
For high-art, originality is much more important, and AI struggles with it, even more than one would expect from such systems. Extensive prompt engineering won't help you that much, so yeah, I was stretching it a little bit in the previous posts. But who knows, it may improve in the future.
Good point. AI definitely threatens the "compofiller" style of productions. To be honest: They had it coming: We've been flooded with bland garbage over the past years and if an algorithm takes over that role, so be it. *Real* art will not so easily be replaced by a robot.
last but not least: this has major potential under "Human Supervision".
kind of reminds me of the 90ies when ppl were discovering Photoshop and it took quite a many years for them to finally do something actually interesting with it ;-)
kind of reminds me of the 90ies when ppl were discovering Photoshop and it took quite a many years for them to finally do something actually interesting with it ;-)
Quote:
AI definitely threatens the "compofiller" style of productions. To be honest: They had it coming: We've been flooded with bland garbage over the past years and if an algorithm takes over that role, so be it. *Real* art will not so easily be replaced by a robot.
This is so true. Outside the scene too. Instead of art, We humans succumbed to churning out soulless cr*p for quite a while, so it’s not going to be that hard for AI to emulate that cr*p.
Quote:
We've been flooded with bland garbage over the past years and if an algorithm takes over that role, so be it.
And your contribution to turn that tide iiiiis....?
Quote:
kind of reminds me of the 90ies when ppl were discovering Photoshop and it took quite a many years for them to finally do something actually interesting with it ;-)
I have yet to see something actually interesting produced with Photoshop. ;-)
@Gargaj: Still working on it. And for the record, I do not professionally flood the gen-pop with my garbage (and I hope you'll excuse my admittedly much-left-to-be-desired hobbyist attempts in the staging areas, like pouet or GS)
Quote:
So you claim that in these million tokens I can provide the LLM with context of Van Gogh’s life. His feelings. His sense of rejection. His relationship with Theo...
Actually, yes, 1 million tokens is enough to put quite a few books about Van Gogh's life as context.
But, this only works in high-end LLMs so far. In text2image creating the mechanism to "understand" prompt is more challenging, as extensive textual context is not part of the image obviously (see https://github.com/openai/CLIP). Still, I'm sure many people work to improve text2image prompt length and understanding as we speak.
Not books. Human life. Human experience. Art relies on human experience. For interpretation of experience you need understanding. This thing can not understand. It’s a glorified autofill, and no matter how much you improve it it won’t be able to really understand. On the other hand, I don’t think it’s nearly complex enough to trigger some sort of self-evolution.
Yeah, but there is human "in the loop". If the human would be able to provide long and extensive prompt that captures all those experiences and understanding you are talking about, it should be sufficient for a next-gen text2image AI to translate this into image. It would also imitate human-way of translating it. Obviously, it's not the same as experiencing first-hand as a human living the life, but it might be unfortunately sufficient to pass the test of fooling high-art snob into think this is art.
OK. It seems to me if there was a human who could examine, process, and transfer to the text prompt the wholeness of Vincent’s personal human experience, that human would be an artist, perhaps surpassing Vincent in sheer talent. Vincent could only do it with paint.
But I get it. It doesn’t have to be perfect. It doesn’t have to be real. It has to be ‘good enough’ and ‘real enough’. That’s the trademark concept of nowadays AI. It doesn’t have to understand. It has to mimic understanding. It doesn’t have to create art. It has to mimic art. OK. What I would like to know is why? Why? I don’t give a f*ck how? Seems that’s all they care about. The how? What about the why?
But I get it. It doesn’t have to be perfect. It doesn’t have to be real. It has to be ‘good enough’ and ‘real enough’. That’s the trademark concept of nowadays AI. It doesn’t have to understand. It has to mimic understanding. It doesn’t have to create art. It has to mimic art. OK. What I would like to know is why? Why? I don’t give a f*ck how? Seems that’s all they care about. The how? What about the why?
We've been there 12 years ago: Lux (Warp / Opal Records).
A great album and it's not completely auto-generated: He listened to it over and over and tweaked the algorithm parameters until it sounded "right".
A great album and it's not completely auto-generated: He listened to it over and over and tweaked the algorithm parameters until it sounded "right".
@bsp
So you claim that generative and LLM AI generated is the same?
So you claim that generative and LLM AI generated is the same?
of course not. this was just the "avant-garde" of things to come (but the primary work flow remains the same)
Quote:
Why?
To undermine power of individual. Put people against each other. Take control. Enslave.
But seriously, for business people it's about short-term profit from raising stocks. For engineers, a fun, challenging problem to solve.
and while that album is hauntingly beautiful I whole-heartetdly agree with your assessment that AI (or any other generative algorithm) will NOT replace real musicians anytime soon.
Kara Comparetto, Béatrice Martin, Patti Rudsil, Lena Raine -- if you are looking for talent and see these names, hire them w/o second thought !
Kara Comparetto, Béatrice Martin, Patti Rudsil, Lena Raine -- if you are looking for talent and see these names, hire them w/o second thought !
(and as always with these kinds of "greetings" lists -- I instantly regret that I said that b/c I missed so many other of the greats, like Joel Corelitz)
"The target of the Jihad was a machine-attitude as much as the machines," Leto said. "Humans had set those machines to usurp our sense of beauty, our necessary selfdom out of which we make living judgments."
Frank Herbert was very much a system critic - give this old interview with him a listen, it's good stuff: https://www.youtube.com/watch?v=AT2R9L_r4oM&t=1s
So you are saying AI tech attracts pathological personalities, because it gives a promise of ultimate power?
We can only hope the promise is false :-P
We can only hope the promise is false :-P
It will be like always. If it comes from the scene, you know it (anyway) and it will be good.
To be honest, I ignore the mainstream, but of course I had to try it out and the only output I made was demoish.
VI is great for learning.
To be honest, I ignore the mainstream, but of course I had to try it out and the only output I made was demoish.
VI is great for learning.
Howdy all!
It's been forever, don't even remember when it was the last time I've posted, hope you're all doing great. Last party I attended was Breakpoint 2004 and the Art of Code circa 2012? maybe where I was handing out with Gargaj (still waiting for him to call me to web an interactive game... perhaps I should send a follow up email... :P )
Long time ago I made a small attempt to get out of being a spectato, put a team together and brought to life some ideas, "I wanna be a demoscener" with Mogwai Productions, "A scene'rs brain" and ACT 3 that was our last demo as Reversed Engineers, but I am aware this is what I do best.
I have been inactive and just spectating the last 22! years crazy how time flies. I was never a coder, I was never a musician, only some good effort 2D guy but I have the passion to bring people together and make something great, lots of imagination and ideas on how to get to accomplish certain things either by providing tools to the right ppl or help them any way I can.
Lately I am feeling the urge to make one last demo (for now), I feel I want to close a chapter but my past attempts to get the old team together or built a new one didn't go as planned.
This time around I am thinking to leverage some AI tools to assist me in bringing my vision to life but I wanted to see what the active Demoscene players have to say about it.
It will be mostly as a Wild Demo and I do understand that the creative process of writing music from scratch, drawing, coding is an amazing and unique creative process that must remain alive, at the same time, should ideas stay locked in the drawer? now that we have means to bring them to life even if they are not exactly in the natural approach that we used to know?
I am building a demo using AI (WebGL) and I either get ideas from videos, images but never input those in the tool, just getting the inspiration and describing everything from imagination with the exception on something I saw on shadertoy that I like to incorporate in the wild demo, totally changed though in terms of aesthetic.
I would love to hear your thoughts, comments, suggestions
-sybeX (Reversed Engineers, aka Rev:Eng)
P.S.: I still want to be a demoscener :P
It's been forever, don't even remember when it was the last time I've posted, hope you're all doing great. Last party I attended was Breakpoint 2004 and the Art of Code circa 2012? maybe where I was handing out with Gargaj (still waiting for him to call me to web an interactive game... perhaps I should send a follow up email... :P )
Long time ago I made a small attempt to get out of being a spectato, put a team together and brought to life some ideas, "I wanna be a demoscener" with Mogwai Productions, "A scene'rs brain" and ACT 3 that was our last demo as Reversed Engineers, but I am aware this is what I do best.
I have been inactive and just spectating the last 22! years crazy how time flies. I was never a coder, I was never a musician, only some good effort 2D guy but I have the passion to bring people together and make something great, lots of imagination and ideas on how to get to accomplish certain things either by providing tools to the right ppl or help them any way I can.
Lately I am feeling the urge to make one last demo (for now), I feel I want to close a chapter but my past attempts to get the old team together or built a new one didn't go as planned.
This time around I am thinking to leverage some AI tools to assist me in bringing my vision to life but I wanted to see what the active Demoscene players have to say about it.
It will be mostly as a Wild Demo and I do understand that the creative process of writing music from scratch, drawing, coding is an amazing and unique creative process that must remain alive, at the same time, should ideas stay locked in the drawer? now that we have means to bring them to life even if they are not exactly in the natural approach that we used to know?
I am building a demo using AI (WebGL) and I either get ideas from videos, images but never input those in the tool, just getting the inspiration and describing everything from imagination with the exception on something I saw on shadertoy that I like to incorporate in the wild demo, totally changed though in terms of aesthetic.
I would love to hear your thoughts, comments, suggestions
-sybeX (Reversed Engineers, aka Rev:Eng)
P.S.: I still want to be a demoscener :P
Quote:
I am thinking to leverage some AI tools to assist me in bringing my vision to life but I wanted to see what the active Demoscene players have to say about it.
don't
Quote:
It will be mostly as a Wild Demo and I do understand that the creative process of writing music from scratch, drawing, coding is an amazing and unique creative process that must remain alive, at the same time, should ideas stay locked in the drawer? now that we have means to bring them to life even if they are not exactly in the natural approach that we used to know?
if you can't make it, learn how to do it, or leave it behind. "AI tools" are not the solution. Just don't.
Quote:
I am building a demo using AI (WebGL) and I either get ideas from videos, images but never input those in the tool, just getting the inspiration and describing everything from imagination with the exception on something I saw on shadertoy that I like to incorporate in the wild demo, totally changed though in terms of aesthetic.
please stop
Quote:
I would love to hear your thoughts, comments, suggestions
i suggest you don't taint your return to the scene (or closure as it may be) with genAI.
Just don't.
Quote:
if you can't make it, learn how to do it, or leave it behind. "AI tools" are not the solution. Just don't.
Or do what I did even without AI...
If you cannot improve on your last productions in any way, simply fucking leave. Your content isn't wanted. Go somewhere else
sybeX you should have came back last year, there were a lot more people shilling for AI slop then!
