AI crap in compo entries?
category: general [glöplog]
Has someone already decided to create a demo about it?
Just a crappy fastintro.
Quote:
potential artists saw Unity and UE and got used to that standard, and now have no use for the scene
I wonder if the scene lost some of the less technically interested artists to Flash even before that? I feel that by 2007, the decline had already been going on for a while, but the availability of commercial engines (and the scene's hostility toward them) certainly didn't improve the situation. Thankfully AI will solve the problem and make everyone happy. :D
Quote:
Quote:One can possess such a knowledge but be discouraged from using it [...]
So, I submitted that pic at Syntax 2023 pixel graphics compo. It has pretty much all that technical mambo-jumbo you are describing and more (storytelling, composition etc). And yet when it goes to peoples' preferences, it trails behind a (yet another) cutesy cat face and barely manages to stay ahead of a (yet another) skull with an 'aftereffects' layer.
Now, I know both those pics are from respected people in the C64 scene who do know their stuff when it comes to producing nice pixel pics. I don't want to disregard their work, and in no way to complain about that compo result.
I only bring this up because you are talking about how a graphician should feel and his motivation. Put AI aside for a second. Should I be getting discouraged from entering compos with a platform other than C64 or Amiga, then? Should I be getting discouraged from entering compos with anything outside the 'girl-face, skull, cute animal' all-star triptych? Should I get inline with 'scene politics' that go by the norms of 'platform vote', 'group vote'?
Nah, because the prior reason I do my pixels is the personal satisfaction. Then, it's the people I can relate to. Finally, it's the fact that, if I didn't make/release those pixels within the demoscene, I wouldn't have made them in the first place. I am not an illustrator nor a graphics designer (I do work in the broader graphics design area, but I don't consider myself to be one).
Will AI crap in the scene make any difference at all with regards to what I do and why I do it? Nope, I will continue to do the stuff that make me happy. Ofcourse it's not the same for everyone. Some scene graphicians have their concerns and they do express it. Within that 'how to regulate it, then' context, I wrote a few ideas some pages ago, and that's all I could contribute to the matter, I think. So, please don't just assume that every scene graphician should have absolutely the same concerns and/or same way of thinking. And please keep in mind that potentially de-motivating things are happening in the demoscene since ever and it's the very people involved in the demoscene who chose to not give more than 2 cents about it.
Yes sorry about this. I guess cats just happens to me whether I like it or not. Sometimes I feel haunted by them. Honestly, I try to do different things sometimes but the cats keep calling me.
This Syntax cat was part of another picture called GIMMIE FOOD with a white cat. This Syntax cat was black and not happy and is called BAD MOOD. They were both released at the same weekend on two different parties. So they are connected.
Sometimes simple fluffy things wins compos. I respect your art very much, probably it's more art than my art. But it doesn't matter, we both did it ourselves with our skills and because we liked it.
Absence/Gargaj: Gargaj is so spot on with this, another factor was/is the tooling/polygonal pipe and 3d cards themselves.
The late 90s was full of fun experiments with software rendering on the PC (3d or not), early 3d cards were horrible in that they were basically good for rendering lit/transparent triangles very quickly but not much more. Shaders came around but took a while to get useful. ESPECIALLY if you didn't keep up with high end desktop cards, the rest of my life at uni was so much easier with an laptop but that limited me to older/buggier APIs.
Combined with this was also that 3d content pipelines became more complicated, I had a bunch of artists around but around that age I got busier and by the time I got a new 3d pipeline done most people had also gotten busier (this was even before Collada, things might've been different if we had something half sane like GLTF back then).
If there is one thing we should've changed culturally back then it should've been to recognize how awful drudging BUT ALSO important pipeline work was and coalesced our efforts on opensource 3d parsers/plugins/reference players for more advanced tools than .3ds (I do remember some .ase parsers) so coders could've focused on wrangling the 3dcards for effects and had art to go with it.
The late 90s was full of fun experiments with software rendering on the PC (3d or not), early 3d cards were horrible in that they were basically good for rendering lit/transparent triangles very quickly but not much more. Shaders came around but took a while to get useful. ESPECIALLY if you didn't keep up with high end desktop cards, the rest of my life at uni was so much easier with an laptop but that limited me to older/buggier APIs.
Combined with this was also that 3d content pipelines became more complicated, I had a bunch of artists around but around that age I got busier and by the time I got a new 3d pipeline done most people had also gotten busier (this was even before Collada, things might've been different if we had something half sane like GLTF back then).
If there is one thing we should've changed culturally back then it should've been to recognize how awful drudging BUT ALSO important pipeline work was and coalesced our efforts on opensource 3d parsers/plugins/reference players for more advanced tools than .3ds (I do remember some .ase parsers) so coders could've focused on wrangling the 3dcards for effects and had art to go with it.
Quote:
and the scene's hostility toward them
What hostility? We rolled over clean.
Hey! Sorry about what exactly? :)
The thing is, I really don't enjoy it when people go generalize with "this and that will strip people from all the qualities" when the demoscene has not much to do with or doesnt care about all those qualities in the first place. The demoscene has it's own standards. As you say, sometimes simple fluffy things win compos. Or just good pixel-overs. Etc.
The reason I used Syntax 2023 compo is plainly because it's 'gone into history books'. So facts instead of going with some random hypothesis to get my point through. That's all there's to it.
I get how you are concerned about the use of AI in the demoscene, but I really do believe there's not much to worry about within this obscure geek community.
Also, I don't regard my pixeling as 'art'. Thanks for the nice words though. :)
The thing is, I really don't enjoy it when people go generalize with "this and that will strip people from all the qualities" when the demoscene has not much to do with or doesnt care about all those qualities in the first place. The demoscene has it's own standards. As you say, sometimes simple fluffy things win compos. Or just good pixel-overs. Etc.
The reason I used Syntax 2023 compo is plainly because it's 'gone into history books'. So facts instead of going with some random hypothesis to get my point through. That's all there's to it.
I get how you are concerned about the use of AI in the demoscene, but I really do believe there's not much to worry about within this obscure geek community.
Also, I don't regard my pixeling as 'art'. Thanks for the nice words though. :)
Quote:
So, I submitted that pic at Syntax 2023 pixel graphics compo. It has pretty much all that technical mambo-jumbo you are describing and more (storytelling, composition etc). And yet when it goes to peoples' preferences, it trails behind a (yet another) cutesy cat face and barely manages to stay ahead of a (yet another) skull with an 'aftereffects' layer.
Dear Rexbend. I know demoscene might be oversaturated with skull images, but I find my Insomnia image visualizing quite clearly how I feel most of the time. As a person with Parkinson’s, I suffer from chronic insomnia. This causes my eyes hurt, a constant buzzing sound in my head and brain fog. It’s sometimes like having lightnings going thru my skull.
Many of my personal stuff and experiences find their way into my images, even if they are just the usual demoscene images to some. I want to add that the image uses QCH mode by Copyfault with four times more vertical color resolution. He also kindly coded 32 hires sprites to the top and bottom borders so I was able to expand the image beyond it’s usual limits. So the image is special for me technically too.
PS. I like your style and find it unique. Also, I have never used After Effects.
https://demozoo.org/productions/342438/info/24378/
Added one hand drawn sketch of our production.
Added also the making of text.
Added one hand drawn sketch of our production.
Added also the making of text.
@Sulevi, I'm deeply sorry about your condition.
Like I already wrote to The Sarge, the point I was trying to make isn't about the reasons that drive graphicians pixel whatever it is they like to pixel. Our personal fixations drive us, or conditions in our lives even. I have my share of those.
There's a couple of things I wish to ask about your condition, but I gather this is not the place. I am sending you a pm at CSDB.
And it's rexbeng. :)
Like I already wrote to The Sarge, the point I was trying to make isn't about the reasons that drive graphicians pixel whatever it is they like to pixel. Our personal fixations drive us, or conditions in our lives even. I have my share of those.
There's a couple of things I wish to ask about your condition, but I gather this is not the place. I am sending you a pm at CSDB.
And it's rexbeng. :)
It's funny because Toxic Modulo 2 made me think of another story-led production on the Amiga: 911 by Limbo. And then I realised you made that one 25+ years ago, Greippi!
I don't think the story in 911 is any superior to TM2 and it's also a bit slow, but to me the art style of 911 while primitive in comparison had its own personality and made me smile. That art style felt like it was "ours", truly born from the demoscene or at least from the computer underground movement. When we opt for Gen AI, I think we lose the singularity of the demoscene which made it interesting in the first place (think the cool unreadable logos, golden fonts, elaborate ASCII art - even if mostly borrowed from other places, the demoscene gave them their own twist). With Gen AI we get instead an uninspired rehash of other art forms (American comics here, but quite often also manga/anime, etc.) without much added value in my opinion. I think this is what makes it hard for many people to be enthusiastic about that new direction.
I don't think the story in 911 is any superior to TM2 and it's also a bit slow, but to me the art style of 911 while primitive in comparison had its own personality and made me smile. That art style felt like it was "ours", truly born from the demoscene or at least from the computer underground movement. When we opt for Gen AI, I think we lose the singularity of the demoscene which made it interesting in the first place (think the cool unreadable logos, golden fonts, elaborate ASCII art - even if mostly borrowed from other places, the demoscene gave them their own twist). With Gen AI we get instead an uninspired rehash of other art forms (American comics here, but quite often also manga/anime, etc.) without much added value in my opinion. I think this is what makes it hard for many people to be enthusiastic about that new direction.
Quote:
What hostility? We rolled over clean.
This gold nugget from a previous thread springs to mind:
Quote:
Unity and unreal prods belong to the wild compo. They are not demos. Deal with it lamers.
Probably just a vocal minority, but still.
^^^ what he said.
Demomakers always existed, and Unity or Unreal are, by all demoscene standards, a demomaker. Such things were never accepted in demo compos. There are no AMOS demos, there are no Flash demos, and there's no reason to recognize Unity/Unreal demos either, even if they look good. Just my two cents.
Demomakers always existed, and Unity or Unreal are, by all demoscene standards, a demomaker. Such things were never accepted in demo compos. There are no AMOS demos, there are no Flash demos, and there's no reason to recognize Unity/Unreal demos either, even if they look good. Just my two cents.
@tomcatmwi: But then Tooll, Werkkzeug and cables are also demomakers.
Quote:
There are, at THE Amiga party, GERP. =)There are no AMOS demos
And i see no reason to discriminate against any of them.
Why yes, the existence of generative productions complicate things a bit. However, a production made with someone's own creator tool was always viewed differently from a prod made with someone else's tool. I'd say, using a demomaker should only be acceptable if the tool was made by you (or your group), or you made such a significant contribution or addition to to the original tool that it can be considered a respectable product. Well, but then, is a large block of code in Unity a significant contribution?
This matter takes us into several directions. Is using someone else's library or module (say, a music player such as MIDAS) acceptable? How about using your own AI model to generate assets or even code?
So yes, your point is valid, on a technical basis. And mine is valid too, on a traditional basis. What now?
This matter takes us into several directions. Is using someone else's library or module (say, a music player such as MIDAS) acceptable? How about using your own AI model to generate assets or even code?
So yes, your point is valid, on a technical basis. And mine is valid too, on a traditional basis. What now?
We're all standing on the shoulders of giants.
Maybe agree to name those giants... and their dogs. =)
Maybe agree to name those giants... and their dogs. =)
Conclusion: this matter is a fucking mess, let's leave it to future generations.
All 3 people!
Quote:
How about using your own AI model to generate assets or even code?
Nobody in their right mind would use AI generated code tho
- it removes all the fun from making demos
- results are wildly unreliable (skill level: intern at best)
- style level decreases massively
True, but it's a theoretical possibility. Here's a wilder one though. What if, in the future, it becomes possible to generate pictures with AI with sufficient speed to turn them into animation instantly? That would be an ai being able to churn out at least 25-30 images every second. Could an AI like this be considered a demo?
Quote:
True, but it's a theoretical possibility. Here's a wilder one though. What if, in the future, it becomes possible to generate pictures with AI with sufficient speed to turn them into animation instantly? That would be an ai being able to churn out at least 25-30 images every second. Could an AI like this be considered a demo?
Sure, why not?
But if you haven't trained the AI with your own data you are just presenting something someone else did as your own.
What if you trained an AI to just output a well known movie? Lets say The Terminator.
While it might be an impressive feat it's not really something you can present as your own.
I don't see how it is different if it is trained on so much data that no-one can tell what the source material is.
Here is an assertion:
It is less lame to take art that is in public domain to use in your demo than to use AI generated art trained on data from artists that didn't consent and can't be properly attributed.
Quote:
With Gen AI we get instead an uninspired rehash of other art forms (American comics here, but quite often also manga/anime, etc.) without much added value in my opinion. I think this is what makes it hard for many people to be enthusiastic about that new direction.
This is also my biggest issue with AI and the reason why I don't worry about it "dominating" the scene or other artforms, at least not in the current state:
It's soulless. The results are typically bland and quite uninspired.
This is fine for "generic" usage, like ads or sth. but it just doesn't leave a lasting impression. If you have seen some of them you start recognizing it and the soullessness of it.
(TM2: entire page 5 process + extras)
https://demozoo.org/productions/342438/screenshots/
https://demozoo.org/productions/342438/screenshots/
I've seen the Demozoo screenshots of TM2 and its development linked above, but I'm still nowhere near understanding HOW and WHERE AI played a role in this! All I see is comic book panels that look human-created, so apart from the strange plot in the speech bubbles, where else did AI have a role? Please explain, cos I'm lost.
