As a hobby game programmer stable diffusion looks exciting, the idea of quickly generating some assets for a game jam is very appealing. But as a hobby musician it strikes me hard. I want to write, compose and play my own music, so I would understand someone feeling the same with regards to hand drawing arts. This whole stable diffusion is exciting on one side, but on the other makes me wonder what we will be left with once this technology reaches higher capabilities?
Initially I had similar misgivings (I make visual art professionally and music as a hobby) but then I realised AI will not prevent me from creating my own stuff, it might just give me more options to use.
A big part of my motivation is enjoying the artistic process itself, and I still get to choose my process.
In terms of a productivity boost, it can give me the same boost as everyone else. Possibly even more as I have decades of experience.
The biggest negative effect I anticipate is an endless stream of low effort content, bit TBH I think we were already there before the AI arrived. I feel this is more a matter of mental hygiene and culture and how we use technology, rather than the technology itself.
I can't talk about visual arts, but music market is actually appreciating a step back to a more organic/human sound. Although I must say it's actually a mixture of musicianship, recording techniques, better and fair use of electronics sounds and so on... So it actually might be like you said, that technology could empower artistic process. But, if we're already in an endless stream of low effort content(and I agree), how would you make yourself get noticed in an ocean of good enough crap once AI hits the public?
appreciating a step back to a more organic/human sound. Although I must say it's actually a mixture of musicianship, recording techniques, better and fair use of electronics sounds and so on..
its funny, people have been saying this every few years since the 1980s
In terms of a productivity boost, it can give me the same boost as everyone else.
Be careful with that, you may end up being more productive but also be expected to do more with fewer human resources or teammates. This will definitely work for some but I can't see this not becoming a race to the bottom for the creators. There will be tons of benefits too but just like with automation, it could be a double edge sword.
I used to work in the animation industry. My friends who are still in it tell me that there is constant pressure from the studios to combine multiple roles into one, without any increase in pay. They've got a strong union so there's a lot of pressure to keep union shops as places with a lot of ways for you to bounce around and spend a year really delving into the fine details of one particular part of the craft, while still being able to afford to live in LA.
I have been thinking a lot about low effort content. Because low effort is quick, dirty and might be akin to junk food. Will people gobble it up or go meh after the initial honeymoon period?
There's gonna be a lot of 'meh', mostly because of lack of intentionality, but also through oversaturation. Basically people will gobble it up until they choke, and then will feel sick and blame it on all art, not just AI art.
I'm in an odd position on the subject myself: I've long wanted to get facile and slick in my art creation, able to do any genre or style, but it backfires. The only thing that keeps me going is, pursuing really eccentric pursuits. Now I think that's a blessing, because ability to be facile is seriously devalued now.
If you're a human artist very derivative of Greg Rutkowski right now, you're more screwed than I could possibly imagine. And yet, the reason someone would do that is because the style is popular on an extremely basic level: a far cry from trying to be a Basquiat from scratch.
I think a very real concern is, can AI adopt and popularize a trending style so fast that it obscures the initial artist from which the AI is trained? If novelty is what's needed, how small of a seed is enough to spawn 10000 AI derivatives, which may themselves be popularizations of the original concept?
Maybe the future of AI is to proliferate any new innovation so profusely that it inevitably chokes out the innovator and hybridizes with 'what's commonly popular' which is the guts of the neural network that makes AI what it is.
Maybe in some fields this has already begun to happen with a little assistance from AI-guiding humans. Take it up a level: what about AI prompt crafting? Can you define, not just what will be commonly accepted as popular, but what will be innovative and trendy, in a neural network?
Yeah. A bunch of interesting points. You also seem to describe a sort of future trendwatchers.
In Star Trek there still seem to be writers and creators of holo deck programs, even though you could probably have the computer serve a blend of different stories. Perhaps “a craft” will still be seen as admirable.
I forget where I saw it, but there was a musician who made a good point: he said, I didn't do all the hard work of learning to play my instrument in order to impress you with my talent, I did it because I have all these songs in me that I desperately wanted to be able to express.
I can understand that. In the visual domain, especially: how often haven't I thought about a story, my own or someone else's, and wished I could show people. I can also understand the dismay of some artists, who did all that hard work, and suddenly find that it's even harder to reach people with those songs, pictures and stories bubbling in them, because now so many others are able to express theirs too.
I didn't do all the hard work of learning to play my instrument in order to impress you with my talent, I did it because I have all these songs in me that I desperately wanted to be able to express
Markets for many kinds of art have already been driven to near zero, but I wonder if AI will drive them negative: if you want a human audience to give you their much-contended attention, you'll have to start paying them.
Vanity publishing is already a thing, it might become bigger. But I don't think the price as such will become negative, because paying people to listen to you sort of defeats the point. You want to believe that the things you have to say are worthwhile, after all.
It looks bad for career artists. Maybe it's an important step in terms of human expression and communication though, that visual and musical art won't require years of skill to convey your ideas or emotions effectively.
I don't have much of a solarpunk outlook on this, I don't think it's going to be that beneficial. It'd be nice to be wrong though.
I don't know. I've put out an album largely driven by modular synth, meaning that I've got many percussion elements driven literally by a machine, defeating the need for me as a human to play drums in perfect time (which I'm not good at)
But in doing that, I've spent years learning how the micro-timing of musical grooves work (some of it's very obvious, like how a heavy snare backbeat will lag slightly behind what the perfect time is, almost to the point of being a 'flam')
As a result, I was able to make an album that conveyed certain kinds of groove not immediately accessible to novice musicians (or drummers)… but given the same information, any schmoe could push a button and have a preset in his DAW produce the same effect.
I think to some extent if the person doesn't really understand the purpose or need for such an effect, their grasp of how to implement and craft it will be pretty loose. If you don't know why you're doing the effortless thing you're not going to guide it very well and your results will be kind of generic.
Rather than thinking of years of skill, maybe call it years of focus, or years of purpose? To some extent, we collectively respond to creations with a profound sense of purpose. If that purpose comes out of AI it will have to be something from the AI, and not simply a blind reflection of us and our crudest drives.
That's actually my point, without the requirement of years skill building, what's left for us to do? Also, imo, what makes art so human is also failure, it's part of the game. e.g. a successful musician releasing a bad record, or a not more relevant musician that after years of experimentation reinvent himself and publish a great record. I don't have a clear outlook on this too, only questions about the future
Yes, as at a hobby level this is true. The problem is that fewer will embark on journey if low effort and a click of a button would compete with their years of hard work. Yes, they can use AI too but that's not what they signed up for in arts, there was a benefit of being able to be in a completely different space, away from the computers and engineering. Anyway, the cat's out of the bag, we shall see how it goes and I hope I am wrong.
Maybe it also feels a bit like mid 90s, when the Internet and Linux came up. Before I had just QBasic and Dos/ Windows. But the with Linux, you got so many compilers and with the Internet (and with CDs on Magazines) you got sooo many programs. For almost everything you would like to do, somebody wrote already a program.
Comments
As a hobby game programmer stable diffusion looks exciting, the idea of quickly generating some assets for a game jam is very appealing. But as a hobby musician it strikes me hard. I want to write, compose and play my own music, so I would understand someone feeling the same with regards to hand drawing arts. This whole stable diffusion is exciting on one side, but on the other makes me wonder what we will be left with once this technology reaches higher capabilities?
Initially I had similar misgivings (I make visual art professionally and music as a hobby) but then I realised AI will not prevent me from creating my own stuff, it might just give me more options to use. A big part of my motivation is enjoying the artistic process itself, and I still get to choose my process. In terms of a productivity boost, it can give me the same boost as everyone else. Possibly even more as I have decades of experience. The biggest negative effect I anticipate is an endless stream of low effort content, bit TBH I think we were already there before the AI arrived. I feel this is more a matter of mental hygiene and culture and how we use technology, rather than the technology itself.
The silver lining is I guess that the low-effort content is of much higher technical quality than it was before, for better or for worse.
I can't talk about visual arts, but music market is actually appreciating a step back to a more organic/human sound. Although I must say it's actually a mixture of musicianship, recording techniques, better and fair use of electronics sounds and so on... So it actually might be like you said, that technology could empower artistic process. But, if we're already in an endless stream of low effort content(and I agree), how would you make yourself get noticed in an ocean of good enough crap once AI hits the public?
its funny, people have been saying this every few years since the 1980s
If it's "good enough", is it really crap?
I consider "good enough" stuff to be decently suitable for the general public not for the excellence
Be careful with that, you may end up being more productive but also be expected to do more with fewer human resources or teammates. This will definitely work for some but I can't see this not becoming a race to the bottom for the creators. There will be tons of benefits too but just like with automation, it could be a double edge sword.
Yes.
I used to work in the animation industry. My friends who are still in it tell me that there is constant pressure from the studios to combine multiple roles into one, without any increase in pay. They've got a strong union so there's a lot of pressure to keep union shops as places with a lot of ways for you to bounce around and spend a year really delving into the fine details of one particular part of the craft, while still being able to afford to live in LA.
I have been thinking a lot about low effort content. Because low effort is quick, dirty and might be akin to junk food. Will people gobble it up or go meh after the initial honeymoon period?
There's gonna be a lot of 'meh', mostly because of lack of intentionality, but also through oversaturation. Basically people will gobble it up until they choke, and then will feel sick and blame it on all art, not just AI art.
I'm in an odd position on the subject myself: I've long wanted to get facile and slick in my art creation, able to do any genre or style, but it backfires. The only thing that keeps me going is, pursuing really eccentric pursuits. Now I think that's a blessing, because ability to be facile is seriously devalued now.
If you're a human artist very derivative of Greg Rutkowski right now, you're more screwed than I could possibly imagine. And yet, the reason someone would do that is because the style is popular on an extremely basic level: a far cry from trying to be a Basquiat from scratch.
I think a very real concern is, can AI adopt and popularize a trending style so fast that it obscures the initial artist from which the AI is trained? If novelty is what's needed, how small of a seed is enough to spawn 10000 AI derivatives, which may themselves be popularizations of the original concept?
Maybe the future of AI is to proliferate any new innovation so profusely that it inevitably chokes out the innovator and hybridizes with 'what's commonly popular' which is the guts of the neural network that makes AI what it is.
Maybe in some fields this has already begun to happen with a little assistance from AI-guiding humans. Take it up a level: what about AI prompt crafting? Can you define, not just what will be commonly accepted as popular, but what will be innovative and trendy, in a neural network?
Yeah. A bunch of interesting points. You also seem to describe a sort of future trendwatchers.
In Star Trek there still seem to be writers and creators of holo deck programs, even though you could probably have the computer serve a blend of different stories. Perhaps “a craft” will still be seen as admirable.
I forget where I saw it, but there was a musician who made a good point: he said, I didn't do all the hard work of learning to play my instrument in order to impress you with my talent, I did it because I have all these songs in me that I desperately wanted to be able to express.
I can understand that. In the visual domain, especially: how often haven't I thought about a story, my own or someone else's, and wished I could show people. I can also understand the dismay of some artists, who did all that hard work, and suddenly find that it's even harder to reach people with those songs, pictures and stories bubbling in them, because now so many others are able to express theirs too.
Markets for many kinds of art have already been driven to near zero, but I wonder if AI will drive them negative: if you want a human audience to give you their much-contended attention, you'll have to start paying them.
Vanity publishing is already a thing, it might become bigger. But I don't think the price as such will become negative, because paying people to listen to you sort of defeats the point. You want to believe that the things you have to say are worthwhile, after all.
Vanity Publishing is basically the entire economic force behind Social Media, and even traditional media.
Inventor: I made a thing! I want people to know about it so they will want it, and pay me for my cool thing.
Public: We can't notice you over the million other things to pay attention to.
Inventor: Hey Mr. Zuckerberg, here's $100. MAKE THEM NOTICE ME.
It looks bad for career artists. Maybe it's an important step in terms of human expression and communication though, that visual and musical art won't require years of skill to convey your ideas or emotions effectively.
I don't have much of a solarpunk outlook on this, I don't think it's going to be that beneficial. It'd be nice to be wrong though.
I don't know. I've put out an album largely driven by modular synth, meaning that I've got many percussion elements driven literally by a machine, defeating the need for me as a human to play drums in perfect time (which I'm not good at)
But in doing that, I've spent years learning how the micro-timing of musical grooves work (some of it's very obvious, like how a heavy snare backbeat will lag slightly behind what the perfect time is, almost to the point of being a 'flam')
As a result, I was able to make an album that conveyed certain kinds of groove not immediately accessible to novice musicians (or drummers)… but given the same information, any schmoe could push a button and have a preset in his DAW produce the same effect.
I think to some extent if the person doesn't really understand the purpose or need for such an effect, their grasp of how to implement and craft it will be pretty loose. If you don't know why you're doing the effortless thing you're not going to guide it very well and your results will be kind of generic.
Rather than thinking of years of skill, maybe call it years of focus, or years of purpose? To some extent, we collectively respond to creations with a profound sense of purpose. If that purpose comes out of AI it will have to be something from the AI, and not simply a blind reflection of us and our crudest drives.
That's actually my point, without the requirement of years skill building, what's left for us to do? Also, imo, what makes art so human is also failure, it's part of the game. e.g. a successful musician releasing a bad record, or a not more relevant musician that after years of experimentation reinvent himself and publish a great record. I don't have a clear outlook on this too, only questions about the future
People play chess even if they don't have a chance at winning against an AI. And in Art there is no winning anyway.
People will continue to paint, like they did after the invention of the photograph and photoshop. Same for music, films, etc.
Yes, as at a hobby level this is true. The problem is that fewer will embark on journey if low effort and a click of a button would compete with their years of hard work. Yes, they can use AI too but that's not what they signed up for in arts, there was a benefit of being able to be in a completely different space, away from the computers and engineering. Anyway, the cat's out of the bag, we shall see how it goes and I hope I am wrong.
Maybe it also feels a bit like mid 90s, when the Internet and Linux came up. Before I had just QBasic and Dos/ Windows. But the with Linux, you got so many compilers and with the Internet (and with CDs on Magazines) you got sooo many programs. For almost everything you would like to do, somebody wrote already a program.
Do the extended creative process for yourself
Just like a hipster/enthusiast with a more manual photography and development process
some people might appreciate it, the market likely wont but thats already the same before AI