So the whole AI movie thing is really moving along now…

oldsynthguy

@NeuM said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

Why must everything be X, Y or Z or it’s not “art”? This attitude runs contrary to the entire history of art. These things are tools. That’s all.

Exactly. They’re not, as you referred to them, ‘artists’ in their own right. A tool is not an artist. Therefore, they do not create art. Would you refer to a camera or a tube of oil paint as an artist?

@AudioGus said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

I use these things all time, for me they are just coarse collage machines capable of outputting epic amounts of source material that can then be reprocessed. It takes a human to get results of any real value out of them.

That’s my point. They are artists tools, not artists in themselves. They do not create art, they produce content based on previously scraped work from human artists, re-arranged and styled up by a bit of code. So they’re unlikely to think up the next art movement/genre, or punk/glam/prog/rock & roll genre.

What a human illustrator/artist/musician does with that content to make it more their own work, maybe that is art. Maybe that will spark a new creative ‘wave’. But for me the unauthorised scraping of existing work by the software itself seems definitely on dodgy ground, legally.

NeuM

@oldsynthguy said:

@NeuM said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

Why must everything be X, Y or Z or it’s not “art”? This attitude runs contrary to the entire history of art. These things are tools. That’s all.

Exactly. They’re not, as you referred to them, ‘artists’ in their own right. A tool is not an artist. Therefore, they do not create art. Would you refer to a camera or a tube of oil paint as an artist?

@AudioGus said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

I use these things all time, for me they are just coarse collage machines capable of outputting epic amounts of source material that can then be reprocessed. It takes a human to get results of any real value out of them.

That’s my point. They are artists tools, not artists in themselves. They do not create art, they produce content based on previously scraped work from human artists, re-arranged and styled up by a bit of code. So they’re unlikely to think up the next art movement/genre, or punk/glam/prog/rock & roll genre.

What a human illustrator/artist/musician does with that content to make it more their own work, maybe that is art. Maybe that will spark a new creative ‘wave’. But for me the unauthorised scraping of existing work by the software itself seems definitely on dodgy ground, legally.

As long as a duplicate of copyrighted material is not being made and being passed off as someone else’s work, it’s not a problem. Works of criticism and satire are still covered by fair use laws (at least in the US).

And that isn’t to suggest all of these things are completely clear. There are a number of lawsuits in the courts moving forward as we speak, so the answers are not all here yet.

AudioGus

@oldsynthguy said:

@NeuM said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

Why must everything be X, Y or Z or it’s not “art”? This attitude runs contrary to the entire history of art. These things are tools. That’s all.

Exactly. They’re not, as you referred to them, ‘artists’ in their own right. A tool is not an artist. Therefore, they do not create art. Would you refer to a camera or a tube of oil paint as an artist?

@AudioGus said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

I use these things all time, for me they are just coarse collage machines capable of outputting epic amounts of source material that can then be reprocessed. It takes a human to get results of any real value out of them.

That’s my point. They are artists tools, not artists in themselves. They do not create art, they produce content based on previously scraped work from human artists, re-arranged and styled up by a bit of code. So they’re unlikely to think up the next art movement/genre, or punk/glam/prog/rock & roll genre.

It definitely is not existing work "re-arranged and styled up by a bit of code". Diffusion models (that are not over trained) operate at far more of a conceptually atomic level than that.

AudioGus

@oldsynthguy said:

What a human illustrator/artist/musician does with that content to make it more their own work, maybe that is art. Maybe that will spark a new creative ‘wave’. But for me the unauthorised scraping of existing work by the software itself seems definitely on dodgy ground, legally.

Nitpicking, but the 'software' does not do the scraping. The models were made after the scraping was done using conventional scraping methods,

wim

I wonder how long it will be before laws are written by AI?
It's probably already happening, now that I think about it, at least in the draft stage.

Then we move along to legal cases being decided by AI. I can easily imagine a school of thinking that would hold that to be fairer than trial by jury.

Gsus

If AI is so good why is the forum so bad? Why hasn’t the forum software improved? Not enough examples of good forum software to ingest?

And btw, using ‘corpus’ makes you look like a pretentious, or was it pompous, elite 🤣

Krupa

@AudioGus said:

@Krupa said:
I still think that incorporation of it into more controllable tools is the key though, prompting can only take so far the creativity and control needed for engaging outcomes. Without this, it’s going to be impossible to direct content in a deep or meaningful way, maybe if the large language models are incorporated into ‘actors’ then it could be interesting, a dialogue between director and actor, establishing motivation and subtext into their performance… still seems a long way off to me though.

As far as still images are concerned there are extremely powerful tools for controlling the outcome. Whether or not the outcome is engaging entirely depends on the human using it. Prompting is just the tip of the iceberg most newbs do not get beyond, img2img workflows and ControlNet are insanely powerful. But then people say 'well that isn't the AI then is it?', I suppose not. But most people's use of the word AI doesn't even seem to fit the reality of what these things are and how they operate anyway.

Yeah for sure, I was using control net and stable diffusion with a blender script a couple of years back, what I mean is that the tools aren’t evolving that quickly. I expected that by now I’d be using it reliably as a render engine, but as far as I can tell it’s still a hotchpotch of strung together hackiness. Maybe I should engage myself to it more, but then again, I just wasn’t loving the output as much as the stuff I already make. And that for my clients, I can’t guarantee that the material is fully legal for them to use on their platforms…

wim

"AI, craft me a piece of legislation that will deliver the electoral votes of the state of Ohio, taking into account all recent and historical polls, historical election results compared to public spending, all other factors relevant to election results over the last 30 years ..."

AudioGus

@Krupa said:

@AudioGus said:

@Krupa said:
I still think that incorporation of it into more controllable tools is the key though, prompting can only take so far the creativity and control needed for engaging outcomes. Without this, it’s going to be impossible to direct content in a deep or meaningful way, maybe if the large language models are incorporated into ‘actors’ then it could be interesting, a dialogue between director and actor, establishing motivation and subtext into their performance… still seems a long way off to me though.

As far as still images are concerned there are extremely powerful tools for controlling the outcome. Whether or not the outcome is engaging entirely depends on the human using it. Prompting is just the tip of the iceberg most newbs do not get beyond, img2img workflows and ControlNet are insanely powerful. But then people say 'well that isn't the AI then is it?', I suppose not. But most people's use of the word AI doesn't even seem to fit the reality of what these things are and how they operate anyway.

Yeah for sure, I was using control net and stable diffusion with a blender script a couple of years back, what I mean is that the tools aren’t evolving that quickly. I expected that by now I’d be using it reliably as a render engine, but as far as I can tell it’s still a hotchpotch of strung together hackiness. Maybe I should engage myself to it more, but then again, I just wasn’t loving the output as much as the stuff I already make. And that for my clients, I can’t guarantee that the material is fully legal for them to use on their platforms…

Yah for animation and video rendering I don't see the point yet. For concepting it is amazing.

AudioGus

In other news ChatGPT-4 passes the Turing Test

https://www.ecoticias.com/en/chatgpt-4-turning-test/7077/

oldsynthguy

@NeuM said:
As long as a duplicate of copyrighted material is not being made and being passed off as someone else’s work, it’s not a problem. Works of criticism and satire are still covered by fair use laws (at least in the US).

@AudioGus said:
It definitely is not existing work "re-arranged and styled up by a bit of code". Diffusion models (that are not over trained) operate at far more of a conceptually atomic level than that.

I’m not overly concerned about the intricacies of the software coding, or legal standpoint to be honest. I take issue with the suggestion of artificial ‘creativity’ and production of ‘art’, and the underhand way in which artists and musicians have been treated.

AudioGus

@oldsynthguy said:

@NeuM said:
As long as a duplicate of copyrighted material is not being made and being passed off as someone else’s work, it’s not a problem. Works of criticism and satire are still covered by fair use laws (at least in the US).

@AudioGus said:
It definitely is not existing work "re-arranged and styled up by a bit of code". Diffusion models (that are not over trained) operate at far more of a conceptually atomic level than that.

I’m not overly concerned about the intricacies of the software coding, or legal standpoint to be honest. I take issue with the suggestion of artificial ‘creativity’ and production of ‘art’, and the underhand way in which artists and musicians have been treated.

I hear yah. Yah if I was retired I'd be all 'Great, screw all this digital crap anyway' and just be honkin horns, bip bapping drums and splashing paint.

NeuM

@oldsynthguy said:

@NeuM said:
As long as a duplicate of copyrighted material is not being made and being passed off as someone else’s work, it’s not a problem. Works of criticism and satire are still covered by fair use laws (at least in the US).

@AudioGus said:
It definitely is not existing work "re-arranged and styled up by a bit of code". Diffusion models (that are not over trained) operate at far more of a conceptually atomic level than that.

I’m not overly concerned about the intricacies of the software coding, or legal standpoint to be honest. I take issue with the suggestion of artificial ‘creativity’ and production of ‘art’, and the underhand way in which artists and musicians have been treated.

Like I said, legal cases are already in motion so there will be a decision at some point… and then there will be new challenges after that. LOL. Court cases could continue long after the Singularity has come and gone.

AudioGus

@NeuM said:

@oldsynthguy said:

@NeuM said:
As long as a duplicate of copyrighted material is not being made and being passed off as someone else’s work, it’s not a problem. Works of criticism and satire are still covered by fair use laws (at least in the US).

@AudioGus said:
It definitely is not existing work "re-arranged and styled up by a bit of code". Diffusion models (that are not over trained) operate at far more of a conceptually atomic level than that.

I’m not overly concerned about the intricacies of the software coding, or legal standpoint to be honest. I take issue with the suggestion of artificial ‘creativity’ and production of ‘art’, and the underhand way in which artists and musicians have been treated.

Like I said, legal cases are already in motion so there will be a decision at some point… and then there will be new challenges after that. LOL. Court cases could continue long after the Singularity has come and gone.

..and countries that give no Fs will have a HUGE advantage the whole time.

wim

It occurs to me that true creatives will in fact be the least impacted by AI. In fact, I expect they'll be inspired to break out in new ways. Look at every artistic revolution and you'll find that it's creatives instinctively breaking the mold when things get stale and overly homogenous. They can even leverage off of AI as they've leveraged off of existing norms forever.

The ones that are going to be more impacted are people like me who aren't especially creative but may have skills making their own interpretations of genres they like. Some of us have skills more like AI itself than a deep connection with originality and creativity. It'll be hard for those who try to make a living doing that. Thankfully I only entertain myself, not try to make a living from creativity.

That's OK. This is all nothing new in that respect. The pace of it is what is dizzying. You watch though. Breakouts and revolutions will appear and new artistic eras will be ushered in. AI won't be able to compete with the true creativity engrained in the human spirit. I really believe that.

NeuM

@AudioGus said:

@NeuM said:

@oldsynthguy said:

@NeuM said:
As long as a duplicate of copyrighted material is not being made and being passed off as someone else’s work, it’s not a problem. Works of criticism and satire are still covered by fair use laws (at least in the US).

@AudioGus said:
It definitely is not existing work "re-arranged and styled up by a bit of code". Diffusion models (that are not over trained) operate at far more of a conceptually atomic level than that.

I’m not overly concerned about the intricacies of the software coding, or legal standpoint to be honest. I take issue with the suggestion of artificial ‘creativity’ and production of ‘art’, and the underhand way in which artists and musicians have been treated.

Like I said, legal cases are already in motion so there will be a decision at some point… and then there will be new challenges after that. LOL. Court cases could continue long after the Singularity has come and gone.

..and countries that give no Fs will have a HUGE advantage the whole time.

We're already seeing that with generative video services coming out of China. They are getting a huge leg up on those companies which are restrained in their development by threats of possible lawsuits.

AudioGus

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

Genres may stem more or less as a narrative from traceable singular sources but an accumulated audience (and a large amount of subsequent imitators) is required for the genralisation (genre) to form and be visible from a cultural perspective. Typically a cultural funnel is used in the delivery system (stations/networks/critics etc). This new content to taste ala carte paradigm that AI, algorithms and modern delivery systems offer somewhat flies in the face of the idea of a genre/following or subcultures forming. It is likely going to eventually be about hyper individuality. The initial wave of course is nostalgia regurgitation but eventually most people will be in their own private idaho with ultra personal content that no one else but the individual experiencing it cares about. Much like dreams I guess.

So my initial question of what have humans done lately in this regard (new music, new genres) was sincere and wasn't about AI so much as it was about the change in culture itself. Subcultures, genres, styles etc simply are not what they used to be even in terms of cultural function. The things I saw 'recently' (ps. I'm old) don't seem to stick like earlier genres used to, due to this disposability of art in general, even pre AI/ML. So I always ask what has come around lately just to see if I've missed something.

NeuM

@AudioGus said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

Genres may stem from traceable singular sources but an accumulated audience (and a large amount of subsequent imitators) is required for the genralisation (genre) to form and be visible from a cultural perspective. Typically a cultural funnel is used in the delivery system (stations/networks/critics etc). This new content to taste ala carte paradigm that AI, algorithms and modern delivery systems offer somewhat flies in the face of the idea of a genre/following or subcultures forming. It is likely going to eventually be about hyper individuality. The initial wave of course is nostalgia regurgitation but eventually most people will be in their own private idaho with ultra personal content that no one else but the individual experiencing it cares about. Much like dreams I guess.

Call me crazy, but I would instantly subscribe to a service that would give me all new seasons (with all of the original cast) of the original Star Trek set back in the time period in which the series was first created, on demand. That would be pretty cool.

And I could see the surviving members (or families of original cast members) being completely OK with getting new residuals checks for roles those actors could no longer conceivably play.

AudioGus

@NeuM said:

@AudioGus said:

@oldsynthguy said:

@AudioGus said:

@oldsynthguy said:

@NeuM said:
That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

‘Other artists’? Are AI systems people now?

You’re wrong anyway. Until AI ‘artists’ come up with something that is actually ‘new’, they’re simply regurgitating the same stuff they’ve scraped in a slightly different way, albeit with a few extra digits.

Of course ‘human’ artists and musicians acknowledge and are inspired by their contemporaries and music from the past, but if there wasn’t that spark of unique creativity from innovative individuals, we’d still be banging rocks together, wailing, and spraying hand-prints on cave walls.

Where’s the ‘new’ AI art styles they’ve invented? ‘New’ AI music genre? With all this computing power and interest, where’s all the new stuff?

I would also like to see the ‘new’ human art styles that were invented over the past two or three years. ‘New’ human music genres?

Maybe they’re all too busy typing in prompts to think of their own stuff. Maybe modern tools have made them/us lazy. Maybe we’re too busy making tools to do it for us instead.

Whataboutism and human apathy aside, if you look back at a recent decade, for example the 70’s, and the waves of new musical genres that appeared, it’s pretty staggering what a mere human brain and a guitar can come up with.

So considering the colossal amount of computing power and trillions of scraped existing material - why has AI not come up with any new art or music yet?

Genres may stem from traceable singular sources but an accumulated audience (and a large amount of subsequent imitators) is required for the genralisation (genre) to form and be visible from a cultural perspective. Typically a cultural funnel is used in the delivery system (stations/networks/critics etc). This new content to taste ala carte paradigm that AI, algorithms and modern delivery systems offer somewhat flies in the face of the idea of a genre/following or subcultures forming. It is likely going to eventually be about hyper individuality. The initial wave of course is nostalgia regurgitation but eventually most people will be in their own private idaho with ultra personal content that no one else but the individual experiencing it cares about. Much like dreams I guess.

Call me crazy, but I would instantly subscribe to a service that would give me all new seasons (with all of the original cast) of the original Star Trek set back in the time period in which the series was first created, on demand. That would be pretty cool.

And I could see the surviving members (or families of original cast members) being completely OK with getting new residuals checks for roles those actors could no longer conceivably play.

Absolutely, and there will be a point where individuals will have the resources to have this generated for them, without need of distribution or even sharing with others. I mean, what the hey, a 2D holodeck still seems sexy af to me. Sign me up for arnie in a conan tv show made in the style of a 1960s cinemascope classic.

knewspeak

@simonnowis said:
did previous 'artists' not at least take a little time to choose, digest, and regurgitate their 'influences' ? ... the machine has swallowed the lot

The dataset the LLM’s have been trained on so far has been a ‘free lunch’ in future that won’t be so, the amount of resources needed for this phase increases exponentially, some seem to believe A.I. will be able to resolve this problem, if not we could always live on bugs.

[Deleted User]

The user and all related content has been deleted.

JanKun

@NeuM said:

@JanKun said:

@NeuM said:

@JanKun said:

@JanKun said:

@AudioGus said:

@NeuM said:

@JanKun said:

@NeuM said:
“Good artists copy, great artists steal.”
—Steve Jobs

Is the fact that this quote is originally from Picasso your whole point ?

Yes, the phrase has a number of variants. There are few original ideas. That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

yup, that is why simply trying to call it 'theft' will fall flat for most people.

"Bach ? Coltrane? Nah, all the same stuff. they use that same boring diatonic shit"

I understand your argument. Let's get rid of intellectual property and copyrights. Every idea, every novel, every movie or image belongs to me, even if I don't have a clue how it was made or what it actually means. it's basically more or less the same "stuff" so it doesn't matter, I can rip it off.
Today, thanks to you, I can at last declare without any shame that I am Stanley Kubrick. I am Charlie Parker. I am Omar Khayyam and Frida Kahlo. Everyone, all at once. Yes !!!

I'm not making any arguments against intellectual property rights. I favor all legal protections for authors, artists, inventors, etc. But ideas cannot be protected. And making something new or different based on someone else's ideas is fair game for humans and machine learning systems.

We agree on that point of transmission imitation and inspiration. We differ on one major crucial point, you want to give the same artistic creation rights to machines as humans. I don't. Because they machines. But you probably don't consider artistic creation as craftsmanship. Mostly when I read your comment, it seems to be > @simonnowis said:

@neum ... but when the machines create everything for us ... ideas are all we may have to create?

@NeuM said:

@JanKun said:

@JanKun said:

@AudioGus said:

@NeuM said:

@JanKun said:

@NeuM said:
“Good artists copy, great artists steal.”
—Steve Jobs

Is the fact that this quote is originally from Picasso your whole point ?

Yes, the phrase has a number of variants. There are few original ideas. That "A.I." systems scrape publicly available data makes them arguably no different from other artists and musicians, who also base their art on what has come before them.

yup, that is why simply trying to call it 'theft' will fall flat for most people.

"Bach ? Coltrane? Nah, all the same stuff. they use that same boring diatonic shit"

I understand your argument. Let's get rid of intellectual property and copyrights. Every idea, every novel, every movie or image belongs to me, even if I don't have a clue how it was made or what it actually means. it's basically more or less the same "stuff" so it doesn't matter, I can rip it off.
Today, thanks to you, I can at last declare without any shame that I am Stanley Kubrick. I am Charlie Parker. I am Omar Khayyam and Frida Kahlo. Everyone, all at once. Yes !!!

I'm not making any arguments against intellectual property rights. I favor all legal protections for authors, artists, inventors, etc. But ideas cannot be protected. And making something new or different based on someone else's ideas is fair game for humans and machine learning systems.

You know as much as me, that those learning systems have stolen for a few years and in total impunity, Copyrighted material from real artists made of flesh and bones. The most flagrant exemple is how artists' voices have been stolen without their consent. That's a clear infringement of intellectual property.

This is simply untrue. Anything original (or even derivative) created by a person can be copyrighted. And all of these works can in turn become the subject of new or derivative works by others so long as they do not run afoul of copyright laws.

Your argument to create new content inspired by older material is indeed nothing new and a perfectly valid one. We've been doing this for centuries. The thing is that when an individual take interest in a very specific intellectual concept and decide to incorporate it or use it as a starting point to develop upon, excluding content piracy, to copy or use this idea, the aspiring creator had to somehow give a retribution to the original owner. buy a book to gain more in-depth knowledge about a specific topic. Buy a cd, vinyl or subscribe to a streaming platform to watch or listen to content to be inspired and then learn from it. There s a clear questioning about Udio policies on that matter

I am also very much looking forward to the udio lawsuit. The main question being what exact material was used to train their AI and where did they get it from. Contrary to Stability AI who clearly licensed a dataset of music, Udio has been pretty vague about this matter. Can court force them to comply and provide a detail of all the learned material and its origin. To what extent their claim can be falsified: that's just my guess, but I bet it should be pretty easy for them to hide the usage of compromising copyrighted data's during the AI learning process. I am happy to be proved wrong if anyone has some insights. Afterall, Everyone is innocent until proved guilty. But knowing where Sunio start up team core comes from, I honestly have strong doubts...

I slightly disagree when you say that we cannot own an idea. You can't own, for exemple, the idea of Heartache even if you have experienced it in your life.
But there are ideas that, when articulated in very specific way, can be owned. That's what patents and copyright were made for in the first place. You probably already heard about this but if not check Steely and Clevie "Fish Market" reggaeton copyright infringment lawsuit. These 2 producers consider that the infamous reggaeton pattern is theirs.
Anyway I think I am gonna stop interacting on this thread and in general, any AI discussions. I have better things to do with this time. I have a very clear idea on the AI topic though and this thread comforted me in my opinion, which eventually is the only thing that matters to me. I don't hate AI per see, as a narrow AI working as an assistant with very very little limited scope in the creation process I find this perfectly acceptable. In the current context, even with good prompt engineering, there is still too much freedom given to the AI in the generative process to call it a full creation of the human who wrote the prompt. At least that's the conclusion I came to after digging a bit about prompt engineering. I couldn't consider any of the generation as truly mine and after almost 2 years of use I eventually decided to get rid of them and I am pretty happy I did. It was nice interacting with you.

NeuM

I find your comments and concerns completely reasonable, @JanKun. Whatever results are found in multiple court cases will be interesting and I have almost no doubt that the judgements will damage US conpetiveness, while other countries will continue to make impressive advances without regulations.

knewspeak

It’s not only going to be tested in regards to Intellectual Property rights but more widely in regards to free speech and freedom of expression, legislation exists in many jurisdictions relating to human rights but nothing regarding A.I.

wim

Sort of OT but ...

https://www.semafor.com/article/10/04/2024/berlin-club-scene-divided-over-new-ai-dj-tech

This will be a good opportunity for intro level robotics too. It should be pretty easy to make a robot that bobs it's head, pumps it's fist once in awhile, pretends to tweak knobs and faders just-so, and studiously pulls on and off some headphones. Interchangeable mouse and other heads sold separately...

Blipsford_Baubie

@wim said:
Sort of OT but ...

https://www.semafor.com/article/10/04/2024/berlin-club-scene-divided-over-new-ai-dj-tech

This will be a good opportunity for intro level robotics too. It should be pretty easy to make a robot that bobs it's head, pumps it's fist once in awhile, pretends to tweak knobs and faders just-so, and studiously pulls on and off some headphones. Interchangeable mouse and other heads sold separately...

Can’t do cocaine with a virtual DJ.

Blipsford_Baubie

But i do enjoy using AI mix for my own personal use @wim . I actually use the one by Algoriddim <sic?>, the one specifically mentioned in the article which you linked to above. Here is a specific example of where it comes in very handy for my personal enjoyment:
I have a private playlist that i made on Soundcloud curating songs or mixes posted from Deep Jungle Records SC. However one of their playlists are only snippets of their early to midi-nineties vinyl catalogs. So listening to the playlist has abrupt changes. DJay can link to SC and i use the AI for a smoother listening experience. The type and length of the transitions are customizable.

MrStochastic

@wim said:
Sort of OT but ...

https://www.semafor.com/article/10/04/2024/berlin-club-scene-divided-over-new-ai-dj-tech

This will be a good opportunity for intro level robotics too. It should be pretty easy to make a robot that bobs it's head, pumps it's fist once in awhile, pretends to tweak knobs and faders just-so, and studiously pulls on and off some headphones. Interchangeable mouse and other heads sold separately...

😆

JeffChasteen

@Blipsford_Baubie said:

@wim said:
Sort of OT but ...

https://www.semafor.com/article/10/04/2024/berlin-club-scene-divided-over-new-ai-dj-tech

This will be a good opportunity for intro level robotics too. It should be pretty easy to make a robot that bobs it's head, pumps it's fist once in awhile, pretends to tweak knobs and faders just-so, and studiously pulls on and off some headphones. Interchangeable mouse and other heads sold separately...

Can’t do cocaine with a virtual DJ.

“Good, more for me!”

Redd Fox

knewspeak

Critical thought.

https://appleinsider.com/articles/24/10/12/apples-study-proves-that-llm-based-ai-models-are-flawed-because-they-cannot-reason