pouët.net

How much will we appreciate code, gfx, and music created by humans, when most ppl go gaga over chatbot prompts?

That is to say, we can already not appreciate chatbots more than anyone who cares not one bit about humans who spend time creating vs. chatbots, who have no life and must spend no time to just websearch human-created stuff and remix it.

In my mind, the only thing we can do is question any work taken credit for.

In this sense, AI has already won; by remixing art undiscovered by most that they have no clue how to create themselves, human art is compared to remixes of the best that can be websearched, and most artists will obviously not all be the best artists.

This is the hand forced. How creative are we? Does the Demoscene typically reward creativity? Does it typically question work taken credit for?

Should we embrace the time spent by humans by requiring workstages for all contributions?

Currently, gfx and visual things are the greater threat. And it will take a very long time for an AI to create competitive code for a performance-limited platform. But music is relatively simple to get away with, with or without AI, and there's no reason why it couldn't remix the next song we 'like more than the original', i.e. the human-created one.

As with every one of my thread on Pouet, the theme is the same: Do we care?

And if the answer is "not really", the conclusion is inevitably and inexorably the same: the Demoscene will not be different from anything else, and Wikipedia will be in the right questioning its existence despite our vehement objections and the pages removed, while of course keeping lots of pages on Milli Vanilli, because they were once... a band... or something...?

Now hold on. Whatever you think of that band, the time they spent was likely greater than that of a chatbot. This should feed your hate of chatbots and the love of every genuine contribution to the Demoscene.

So do we just thumb up any misguided prod just cos it's clearly human made, warts and all? No, we don't. So then, what do we do with those who can't even show they made it?

My suggestion is to try our best to not let it be part of the Demoscene, and a general rule that is understood when contributing. Make what you want, but make it. This way, we avoid the silly bickering, wasting the time of the lives of humans. We can at the very least remove that bias towards AI, which they will obviously be completely oblivious and uncaring towards. AI and HI will never co-exist, because AI is "what it is".

Ok, but what about https://bjoernkarmann.dk/project/paragraphica

Hey photon,

Firstly, I think you bring up a critical issue for all creative communities, not just the Demoscene, as AI continues to permeate our lives in more profound ways.

I absolutely share your sentiment on the authenticity and creative passion that human coders, musicians, and graphic artists bring to the table, which is something AI lacks inherently. The AI may be capable of remixing or creating fresh content from a wide array of resources, but at the end of the day, it is fundamentally lacking in the lived experience, emotional depth, and creative intuition that drive human artists.

In the context of the Demoscene, the celebration of human creativity, the spirit of competition, and the enthusiasm to overcome hardware limitations make it unique and powerful. You're right - creativity is the heart of the Demoscene, and we should strive to preserve this ethos.

I agree that we need a clear distinction between human-made and AI-made work, especially if AI's role in content creation continues to expand. Verifying and validating the authenticity of work through workstages or other forms of proof could be a step in the right direction, especially if it helps to preserve the community's integrity.

In response to your question of whether we "thumb up" any human-made prod just because it's flawed and clearly not AI-generated, I would say it depends. I believe there's something to be said for appreciating the effort, passion, and creativity that went into a piece, even if it's not technically perfect. However, blindly praising all human work while disregarding quality is not the way to go, either.

AI may be able to mimic or recreate aspects of what we do, but it will never fully comprehend the meaning behind it. The code we write, the music we compose, and the graphics we create are expressions of our experiences, emotions, and ideas, something that an AI simply doesn't have.

In a way, the emergence of AI could make us appreciate the 'human' aspect of our work even more. Let's embrace this opportunity to reflect on what makes us unique as creators, and let's take steps to ensure that our Demoscene community continues to be a place where human creativity, passion, and effort are celebrated.

Keep pushing the conversation, photon. Your threads never fail to engage the community in meaningful dialogue about our craft and the future of the Demoscene.

More seriously, I agree wholeheartedly with the principle of providing work stages. In fact, this approach is already being employed by events such as the Ludum Dare Game Jam, which encourages participants not only to submit their final results, but also to document and share their personal and technical development logs throughout the process. This additional layer of evaluation enriches the competition, creating a broader understanding of the creator's journey and the challenges they faced.

Personally, I'm not fascinated by AI generating original works. I view it more as a form of fraud than authentic creativity. However, I am deeply intrigued by the possibilities offered by navigating through the latent space provided by AI (hence my post of my twitter account above). Rather than limiting our use of generative AI to simple, conversational systems, which though easy-to-use and accessible to the general public, I believe we should explore the vast potential offered by more complex forms of AI.

The demoscene, to me, has always been about pushing the boundaries and hacking the system, exploring every crevice and exploiting every loophole we can find. With this mindset, why shouldn't we apply the same principles to AI? The future of the demoscene may well depend on our ability to merge human creativity with artificial intelligence, and in the process, redefine what we understand as 'art' and 'creation'.

This is true on one hand. On the other it doesn't matter at all if the viewer does not know about it. We all use inspiration, may it be concious or subconcious. So a machine is way faster in finding the ingredients to cook something while the human might need way more time (and inspiration) to come up with a similar result. Plus it has a larger pool of sources than any human could handle. That's at least how I see it.

As long as is does not become the standard in e.g. gfx compos to use AI stuff in the process as "a tool" ... at least folks should be open about it then. But since people used to rip code and gfx in the past honesty depends on the person (tm). But as I said above, if the viewer does not know it does not matter what the story behind the result is (while judging/voting).

Workstages are a (sometimes) used option indeed, but I guess we need to be careful not to become guilty when hitting the submit button unless we have proof that we are not.

In the end we all will see how things will turn out. same as possible changes within the brains of humans and in society, if we want or not.

That's funny. In my book, how a demo is made is a big part of the value of said demo. That happens behind the stage so it's usually underappreciated.
There are proponents of "the end justify the means" aka "just use the best tools available".
Well, soon enough the tools will be smarter and more creative than us! I guess "0% AI" will replace "0% scan" and "Winners don't use crossdev (except to win)".

I found this video interesting: https://www.youtube.com/watch?v=CfVHXDty5Pk

Yes, the demoscene is one of these things that is (at least) as much about the process than the end result. And AI certainly allows to get a pretty close result to what we did before, with a lot less efforts.

So, you have two choices from there: complain that it's cheating, and set up rules to preserve things as they were before. Or, accept it, and use the fact that it removes some of the previously needed effort, and invest that time into going further and exploring new things.

To take other more demoscene-related examples, in the 90s there was a similar debate about 3D accelerated vs software rendered demos. At one point there was a lot of demos using 3D acceleration to do "3D flybys", which looked kind of cool initially, but got pretty boring after a few dozen demos doing it. But now there are a lot of very cool demos using GPU acceleration, in many different ways.

For now we are in the "3D flyby" era of AI stuff. We'll see where we get from there. Maybe it actually can't do much more than that and it will get boring after a while. Or maybe it opens a lot of new possibilities, and makes some of the old things irrelevant.

I have mixed feelings about it, as I see new people who had no skills in art suddenly creating really good pictures in great quantities. It's like the idea-guy taking over spaces and skill cannot match that. You either need to adapt or perish.

Meanwhile, I might have noticed the use in few prods, from demos, games, background artwork for youtube videos, so it's appealing to use it if you are not an artist and some of these, if I forget it's AI, they are pretty nice. But it gives me this bad feeling again, like what's the point. But since it's here to stay, it's out of my control, I'll continue doing what I am doing as I like it, even if I could be using all the modern tools. That's about coding too, why spend time coding assembly when you can do pico-8 stuff much more easilly?

But there is another thing that noticed. Artists of course are most affected, but I see the discussions on youtube and they are most pissed off. And besides the copyrights, they get it that it's a loss of meaning, if with few text prompts and no skills you can do all that, then the process of creating the art, what gives you meaning is lost. But on the coders side, at least from those tech bros on youtube, I see positive messages that read like "Finally, the bright future is here! We Will not need to Code again!". It's like in programming people don't get pleasure from the creative process, I guess it makes sense that most people want frameworks on top of frameworks and as far as away from the machine. I find some random computer science student sometimes that tells me "I don't really like programming". But you don't find an artist saying that. Which is strange. You never find a mathematician or physicist who is like "You my math bros, the future is bright because we will never look at numbers and symbols anymore, these are for nerds only lol!". Why is it only programming? (That's also one reason people avoid caring for performance, they just want to write one thing and leave it there).

That was a bit off, but why programmers don't like programming (negative stereotypes of code monkey, most are professionals but not really into?), but artist love creating stuff and where the first to see how meaningless it is?

"The struggle itself towards the heights is enough to fill a man's heart. One must imagine Sisyphus happy."

- Albert Camus



That remembers me the stir that was caused in the scene with the arrival of affordable scanners.

I think in compos you can only appeal to the honesty of the artists. In the 80's/90's many GFX artists copied pictures of Boris Vallejo. In principle, the human implementation of an image in pixel style is nothing different than having an AI do it. It's just faster.

However, this topic becomes problematic if you do not simply have an image created by an AI, but only partially use tools in which the AI works. Or actually uses AI images as inspiration and then paints or reworks them. Artists are already using Google image search for inspiration. What would be the difference to an AI? And as tools continue to be infused with AI (e.g. Adobe Firefly), it may eventually become impossible to create images that didn't involve AI. Unless you draw them on paper.

@Optimus: I could imagine that AI will also become interesting for artists in the future. Namely, if you only train an AI with your own works, for example. Then, as an illustrator, you could have a much larger output than before. Because creating illustrations is often a tough process that at some point is no longer fun, especially under time pressure. Or if the customer has constant changes.

The difference here is that programming offers promising career opportunities so it's not merely the domain of passion.

"Sorry but your Deluxe Paint has been taken over by an AI"

deluxe pAInt

Good to see you calling this out.
Please just think about it next time you are about to thumb down a prod like bitmap rasters or salty water.

.. both of which have a grand total of 1 thumb down

Isn't creativity challenging for almost any scene?

We know that the initial reception for many great artworks of the past has been cold.

Perhaps it's noteworthy that demoscene does not have professional critics who sometimes champion works that are shunned by the audience. Or is the role of Meteoriks jury similar enough..?

The point was possibly that both of those are by the same person who asked if the demo scene rewards creativity. Kettle, pot, etc.

indeed

Same for Batman Rises btw lol

And EON …

it’s only matter of time until i code like crazyass smash, too, deal with it lol.

https://www.tiktok.com/@elestepariosiberiano/video/7244576481030982938

i see you hatin but my prompt engineering skills are superiorlously even for creating notch in less than ninety seconds. end.

https://www.linkedin.com/posts/lukas-lundin_githubcopilot-openai-ugcPost-7080098159556726784-HQ2q

I can’t understand why everybody conflates LLM’s with AI. My problems with LLM’s are: (1) they are energy hogs, (2) they are centralized and owned by big capital and (3) they (chatbots, prompt-to-image bots) do not represent original nor creative use of technology, they bring nothing new or revolutionary to the table, they are rather mimicking more of the same old ‘art’. They are big data scraping / remixing machines. Not that a lot of human ‘artists’ are not basically the same thing. I see something like Mandelbulbs for example as way more creative than Midjourney/Dall-E/whatever. To me they represent exactly the opposite thing to scene/garage/diy ethos.

The main difference between "prompt engineers" and "actual artists" is control, I guess

I respect people who have control over their craft. Skilled coders, fractal artists who are not working generatively, musicians applying knowledge and quantitative techniques instead of randomly changing parameters until it's nice. Some coders using copilot may have control, but the majority doesn't. Unity allows control, but the majority of its user base has none. Same with diffusion- or language models.

Eliminating boilerplate is a legitimate use case of AI imo. No need to claim "looking it up in the docs and adding the example, with slightly different parameters", is an important part of writing code. Also no use in excluding this use case by ruleset.

Eliminating software design, or visual concept / composition from the creative workflow by using those same tools is lazy and redundant. If I'm consuming such a prod, I'm usually not able to tell the difference tho (in the case where the author has applied control beforehand). So for me: If it's pretty, and if it makes me happy, create it however the fuck you want. And if you're able to amaze me with low-effort diffusion model crap, please do it.