It sounds like most of us will be redundant.
1 Like
We are already seeing artists giving up. How many times have you seen questions like “Should I go to art school now”, or “Is learning art worth it”? And remember these people are passionate about art. Its just that the circumstances currently are such that spending to learn art does not seem like a worthwhile investment. But at the end of the day, I think this disagreement stems from how both of us understand the artistic process. So I don’t think arguing further will change any of our opinions.
I personally don’t feel like it is not possible for an AI to judge how “good” is a new style based on artworks using older styles. If you trained an AI on renaissance artwork, would it be able to tell if anime art is good? But again you can argue that every artwork contains all human emotions and preferences engraved into it. It just seems very unlikely to me.
But what I think both of us can agree on is that creating new art styles is going to take a huge amount of time, effort, and computation. Way more than the current remixing. AI companies do not understand these concerns. Artists do, but they hate AI.
Don’t overgeneralize. Some artists do, some don’t. Some companies do, some don’t. Some AI companies work directly with artists and studios. Some artists love AI art.
IMO the whole hate on AI just seems like it’s huge to you, because you are in that social media echo chamber that does hate on it.
4 Likes
The biggest AI company Stability AI is owned by a hedge fund manager. And I really haven’t seen any well-known artists promoting AI art. There might be some enthusiasts who care about both art and AI, but they don’t have the resources to pul something like this off.
The biggest ones are probably Micorsoft and Google. Having a hedge fund manager as a company CEO btw. is not automatically evil.
Also, as I said just look outside your information bubble, and also https://en.wikipedia.org/wiki/No_true_Scotsman
Please provide a counter-example to prove your point. For the sake of avoiding argument, lets bind ourselves to artists who have worked for famous movie and game studios
Just listen to some of the recent interviews done on CG Garage or Andrew Price podcast, you’ll learn about quite a few. Also check out corridor crew.
Corridor Crew have been constantly in support of using AI as actual tools and not the finished products as advertised by Stability. I will watch the interview. Using them as tools will not stop art form prospering.
1 Like
So will AI kill art or not? Make up your mind 
Continuing discussion in Direct messages
1 Like
That is not unusual for situations where changes are unavoidable.
If an AI was trained on renaissance art and we would have some sort of exploration system in place, it would likely first start to slightly diverge from renaissance art (keep in mind we train it to find something unique, which still is “good” according to renaissance art). From time to time, we let humans look at the results and they could judge whether those are reasonable directions, worth to explore further. If they are, those examples are added to the “good” dataset, such that a further exploration is possible. Over time, such a system can easily get quite far away from the original. I don’t see, why it shouldn’t be able to come up with styles that are similar to anime at one point.
Of course this would take a huge amount of time to get it running. But as similar systems already exist, I am sure this is going to happen. I am very uncertain about the computational requirements of a well optimized system like that. I assume with the current remixing (I would be very interested to see actual examples of remixing!!!), you are talking about current image generation models. Those are not well optimized at all, that’s why I don’t agree with your statement.
What concerns? Computational/energy requirements?
1 Like
I meant concerns about art not evolving
What makes “good” is not set in stone but a reflection of society. Renaissance artists certainly wouldn’t have liked Anime art today. The art we are seeing today stems from hundreds of years of millions of artists creating art influenced by billions of personal experiences of society then and now. If we want the same level of evolution we will basically have to simulate human civilisation itself.
1 Like
That’s why there is human feedback in the system…
That’s just ridiculous. Those systems are build to create something for humans, that’s why they are the judges. The real world is the judge and we most likely are not looking for huge changes in the form of renaissance → anime, but very small changes which still resemble with what we like, yet are unique.
Art will evolve no matter what. For most of the human history art was not a profitable profession. Even today I think that only small fraction of all the artists make a living out of art.
1 Like
Did you ever, ever hold a pen as a kid, and live through what you were drawing while you were drawing it?
Literally, in my experience the goal for a kid isn’t to make any “pretty picture”. It’s about that wild, vidid fantasy world within, ideas dancing, tumbling, swirling, bouncing … a kid drawing is a kid connecting to that world.
The pen is but a focus, a magic wand that can make this happen, make a kid’s fantasies come to life right in front of their eyes, there, right while you’re drawing it, right on that slice of paper! Can’t you see the dragon moving, hear it roar?
For god’s sake, just be silent and listen while kids are drawing! Essentially it’s a game for them, they even create their own soundtracks … 
AIs can’t contribute anything of value to that, on the contrary, they’re taking away everything! Because it has never been about perfection and final images, but about the inner dialogue, the process! However poor the drawing, if no adult or AI is spoiling it for them, in a child’s eyes each one is perfect.
2 Likes
You will need humans from literally all walks of life. And certainly, these many people will not volunteer for a seemingly obscure job. And you are already getting high-quality art, so what is the need? Why invest so many resources?
We have discussed this already. In my opinion, fast AI will kill the next genration’s will to learn art fundamentals. In @DeepBlender’s opinion AI art will help them to get introduced to art. Any further discussion is not going to change any of our opinions. Only time will give the answer.
I mean yes…? humans are the final judges, right? I can’t understand if you are for or against this point.
Too old I guess, but
The thing that bothers me I guess, is that the future generations won’t feel that change. Babies don’t start from living in the Ice Age, but I sometimes wish people could live long enough that they can have their own time going through human history. Sure you could read books, but that’s where teachers start to put a clear line between good and bad, even if in reality, it was a gray zone. I think that something greater than any pattern of words is going into the void everytime mankind progresses. Feeling.
The hard work so many have put into creating the dream like reality that is today, kinda made people useless. Working for the future? I guess. But somehow I feel that that might bring a downfall someday, when humans become meat blobs and can’t sustain themselves anymore. Everything is a cognitive tradeoff but we seem to have too much of a debt, when considering an AI assisted future.
2 Likes
What I meant was that the reward for the kid, if it was able to create a “pretty picture” would be quite good. The more it contributed to it, the better. This sort of tool might be useful for an initial boost or if they are stuck and are not getting better.
In my opinion, being able to explore what kind of visual styles I would find interesting is fascinating. Or easily exploring certain directions.
There is no being for or against this, because all those are literally made for humans and as such, they are the judges.
1 Like
One thing I have to say on this subject, other that eventually almost nobody will feel that learning how to manually draw or sculpt/modeling will be something worth investing time on, is that with AI generation tools one is not really learning much in terms of fundamental skills of general validity. If some new advancement comes out or one service dies, you’re left with nothing. It’s not like going from physical (analog?) to digital painting and viceversa.
6 Likes