pouët.net

@rexbeng/PA:

Yes. and in fact, "art-ing" is not (or at least should not be) a thing. Art is in the eye of the beholder, i.e. the audience.

You can't create "art" on purpose, other people decide whether it is, or isn't. Or in other words: "Art" requires "Context".

When a creation resonates with you, for whatever reason that you possibly cannot even explain, you may call it art. It's a diffuse and fuzzy term (in the true sense. I am quite aware of the "art dealer industry" that revolves around it, which is mainly about money and investments, though).


@Bifat: I was surprised to read your comment. In my opinion, "demos" are most certainly a kind of digital art, and they are even acknowledged as "world cultural heritage" by some (e.g. UNESCO, the United Nations Educational, Scientific and Cultural Organization).

I've always thought of them as "digital graffiti", in the sense that both are created without commercial intent and you can basically do what you want (within the boundaries of human decency, of course).

Regarding artistic quality: Everything, regardless where you look, is >95% "bland, overproduced garbage", IMHO (commercial and non-profit), nothing new under that sun.

@bsp: I'm surprised too. Some of your statements seem to contradict each other. I agree for example that you cannot create art on purpose, and that it has to resonate with the recipient.
But when you say demos are being "recognized" as cultural heritage - what does that have to do with art? We can certainly agree that demos are a form or craft.
Art is novelty in concept, idea, expression or execution that resonates precisely because the context is removed. Art communicates a new idea or possibility to perceive. That's why it's so rare, especially within the constraints of a form, and demos are especially limited in this regard.
Therefore AI can never create art, but only kitsch or good craft at best, and appeal to a common cultural background of all the things we have seen a million times. It's perfect to be employed as a painting tool. Because AI produces a rather generic, cheap painting color, or the "background that is to be expected here", it can be shown and used without restriction, and poses no danger. If we're lucky, it may even help us to sharpen our senses to recognize art.

The fun thing is: it is not -- neither a novelty, nor without context. Highly acclaimed people have created (on canvas) what is basically a filled rectangle and have been praised for it. Art really cannot work without context. And the term *is* fuzzy.

I don't think that "demos", as an art form, are limited per se. On the other hand, there is, undeniably, a certain "uniformity", "instant recognition value" to them (possibly due to a relatively small group of people "copying" / "mimicking" each other).

Last but not least, I agree that "AI" will never recognize "art", only humans do. It is just another tool that needs (purposeful) direction.

I am quite curious what the future will bring, given the HW industry's tendency to spend more die-area on NPUs than GPUs.

I cannot accept the circular argument that what has been declared art by some other definition can - according to this definition - only be art because of its context. For we have reasons to assume a black square can be art for other reasons too - or not. Certainly there is the context of human nature, and context is a gradual thing too. There is not just perfect art and utter garbage. But this contextual idea is so inclusive of garbage that it's of little use for the things that interest me.

The main problem I have with "AI", generally speaking, is not with the tech but how it's used.

I would have zero problems with a wise, compassionate entity like "The Oracle" (as pictured in the Matrix movies) but I fear that, in the long run, we are more likely to create vicious "Agent Smiths" that will eventually turn against us (ref: use of "AI" in warfare).

An "AI" needs parenting, and, as much as I hate to admit that, we, as a human species, are, at least currently, not good reference points to train any artificial intelligences (at least not for use in a feedback loop to our own society).

Somebody used Google Gemini to create a picture of a blue hedgehog as a game character. The result: 75% of the pictures looked like Sonic the Hedgehog. IMHO this clearly shows that the use of AI tools for image creation is copyright infringement!

What about the other 25%?

Both The_Sarge and grip shares much of my views on this.

But. What is demos really about? Showing of human skills and the passion that goes into those skills, or just showing *something* on screen?
It takes minimal skill to generate AI generative botshit. Almost none. Someone that uses AI to generate images don't make those images.

Why the fuck do we do demos? To make it easy? Making demos has never been about convenience, and outsourcing it to some blind idiot machine is nothing but convenience.
If you use AI to generate that kind of shit, you are a fucking lamer.

Why the fuck should I even bother spending shitloads of time on making something when others just type in a few words in a prompt and get something shat out in seconds? Why?
I see no point whatsoever in continuing making demoscene related stuff if others just cheese it. It's cheating and lame.


And, the worst thing with AI generated images is that it makes me question everything that is actually made by humans. If I notice the slightest detail that makes me suspect generative AI is involved I start to question the whole thing. It's not only demoscene related, I've started to do this with basically all digital art and it so fucking depressing. I can't trust shit anymore.
All this has made me very reluctant to do any kind of digital art myself. I don't see the point anymore, and if this infects the demoscene even more I see no point at all in even bother.

It is ruining so fucking much. It is ruining my own will to actually put any time and effort in creating digital art, and I know this is a huge problem for extremely many artists. It is a constant problem amongst artists to get their own authentic and human made art questioned.

It isn't only about the demoscene.


If you use AI generated images, you are actively working against real human artists, and if you do that, fuck you.

Would that change if suddenly everybody stopped using generative AI?

If no measures against is taken against, it will only get worse. Way, way worse.

Maybe one of that measure would be having artists be entirely open about their process, with documentation all the way. But then i guess that many artists won't like that, for various reasons.

Yes, that could be one measure, albeit a very weak one as it relies solely on honesty, and humans are the most, and probably only, dishonest organism that we know of.
Besides, why would anyone be dishonest about their art? Why would an artist be so ashamed of their art and the processes involved in making it?

Please don't go into "what is art" territory.


Without this, everything is screwed for sure. Even if this was accepted as standard, I'd personally still have a sh*tload of problems with AI. But this should be bare minimum. I understand that some people couldn't be arsed to document their processes. And that's OK. But in this case I see it as a safety measure for all our sakes.

I think it's not about dishonesty but more about giving away trade secrets.

Indeed, the whole "what is art" discussion is an entirely different discussion.

I don't think I've ever met any artist that think about their own processes and methods as trade secrets. On the contrary, most artists are very happy to share and talk about their art. The only "artists" I've talked to that see it that way have been grifters, scammers, and cheating liars.

Hi people ! I think most has been said, already ! So my 2 cents are probably gonna be useless blabbering.

1) there is a continuum between using no AI at all and completely depending on it. Some complex algorithms in painting programs could be considered as a form of AI, for instance. Using AI doesn't prevent from massively working on the picture afterwards.

So, if we define rules and a distinct category, they must be very precise.

2) as someone said a few posts earlier, from this point there will be a constant suspicion for some of us. This is my case. I am massively impressed by stuff I know were made 100% by hand and I think this is my definition of the demoscene even if I know there are some conveniences people can make use of without altering the whole concept.

For instance, I didn't buy the latest PolyPlay boxed Amiga games. I used to -- they were drawn by a talented graphist who died in a car accident. Now I suspect some recent covers are made by AI. I feel it's completely pointless to waste resources building a boxed game if the box has been drawn by a robot !

3) would we learn how to draw, code & write music if we had access to AI ? Learning is an extremely long process which gives results in the long run. Is it possible to keep motivated if everything we do look like crap compared to stuff made in a fraction of a second by AI ?

Personally, having a constant Internet access already kills my creativity, because I need some (temporary) void to focus. This is the next level IMHO.

4) I am sure numerous counter-examples can be given, of people making amazing use of AI. But maybe those people have already developed tons of skills before having access to this kind of technology ? Or is that a paradigm shift ?

I know of many examples of artists that have used generative AI in their own art in a way that I have no problems with. In those cases it has been used as a part in the process.

This artist is a good example: https://www.instagram.com/p/Ch7E3v1DqJ8/
She trained it with photos of herself and generated a bunch of fucked up images and then used her own skill to actually incorporate it in her paintings.

I have no problem with using generative AI in that way. She still used her own skills to actually produce the art and didn't rely on the blind idiot bot to do it for her. She painted it herself, with inspiration drawn from botshit fed with a dataset that she produced herself. She did the heavy lifting and I'm pretty fucking sure that she could do it without generative AI.

Feel the same, like I would add noise to an already tired discussion and probably repeat what others have said.

But this whole conversation makes you think what is that you personally enjoy about art. I am not artist and I do appreciate art as a craft, I am imagining someone painstakingly pixelling through a hard process. Others may think about the end result and the theme, aesthetics, composition, etc. Our different reaction to AI might be based on what we value from demos/art.

It made me wonder though about the context, outside of pixel art, when I see a demo that was done in software rendering I feel something deeper. I posed to myself this question, what if there are two demos, one software rendering and one GPU, but they both produce the exact same pixels frame by frame if we compare them (assuming you degraded the GPU demo to render in the same exact manner as the software rendered demo). Pragmatically it should be the same. But I feel the context that there is more involved, once from the coder to engineer the process to render each individual pixel and also imagining little minions of pixels all precisely engineered together to fill the whole. Even if the result is the same, one feels more valuable than the other. But that's just me, if I watch a demo and I know it's also software rendered or written in low level assembly, even if the result was 100% the same frame by frame to other ways of generating this, I would feel like the coder's soul is in those pixels. Technically it shouldn't matter, but there is something deeper in the creations you watch when you know that a different and more sophisticated process came to the same result.

@Optimus: I agree with what you say.

I value human creativity and the actual effort, and the context is an important part in how I value something.

I feel completely indifferent when looking, reading or listening something created by an AI. It would be interesting to delve deeper into the psychology of the phenomenon, but they just feel like.... nothing.

The novelty has worn off.

@Optimus : +1000. Yep, context, indeed and valuing the efforts put into the creation.

The spectator is not passive passive and is actually trained his/her whole life. Enjoying jazz music or operas for instance, is not innate !

As I see it, I value human creativity, skill, effort and time. When someone just puts out AI generated botshit with minimum creativity, with no actual skill, effort and time, they don't value my creativity, skill, effort and time, so why the fuck should I care?

I don't get these visceral responses at all. This looks like a textbook "in the moment response to a new technology". I'm not saying you have to like the use of AI such as is being discussed here - but this is a personal response, right? Feels like there's a big pile-on happening here, probably - I'd guess - because this whole conversation has touched a much broader raw nerve.

I'm just finishing a very last minute intro where I needed some text. I don't have time to find a gfx guy to do this for me. So I booted up ol' Grafx, and used a Windows font. It even perfectly scaled and anti-aliased it for me. I won't be crediting Microsoft OR Grafx. I look forward to the frenzy of downthumbing

It isn't only a personal response. It's the response of thousands of artists.