I actually think the “it’s soulless… FOR NOW” panel is pretty important.
People who believe in the value of human creativity have been pretty casual about saying that AI generated work isn’t as good as work created by a person, but what happens if in another iteration or two it actually CAN produce “good” “art”? Like, what happens if it’s cranking out screenplays and paintings that DO pass muster? We’ve got to be prepared for that possibility, and try to act now to make sure that our world is structured around preserving human dignity on its own merits. The existence of a faster work-doing machine shouldn’t necessitate that all human workers must now starve.
A computer-generated “Van Gogh” is not art any more than a mass-produced coffee mug is artisanal, no matter how “realistic”.
This has all happened before. Take photography. People thought it was the end of visual art. If anyone can take a photograph, why would anyone spend years learning to paint?
Artists answered by pushing the medium beyond the limits of realism. Impressionism. This did not make photographs go away. But when I see a picture of someone’s cat, I don’t usually go “art!” – even though 200 years ago the mere existence of a photorealistic picture would have implied very impressive artistry.
The work that clankers are very quickly taking over is that which does not require art. Visual filler. Lorem ipsum. Corporate communications. Out with artisans, in with industrial machinery. This is the same story that has already happened to almost every artisanal trade, from scribery to pottery to smithing. Visual artists and writers thought themselves exempt from the industrial revolution; they aren’t. It will be a worsening socio-economic crisis. But it won’t “end” art. Clankers definitionally cannot, and will never do art. Not until they gain a conscience of their own.
That being said, I think AI Bro existentialism and singularity hype has a lot of people on particular edge, beyond what the camera and other past innovations triggered, since it’s pushed at such high levels of our world. But (speaking a fervent local ML tinkerer), the proof is not in their puddin’, as professional, foundational researchers would tell you as well. Not just because of technical limitations, but because corporate enshittification is already taking effect.
People on the more spiritual side of things thought that being photographed meant trapping a part of your soul into the camera. It was a more existential creation than we give it credit for. Before then, nobody had ever stopped time.
I think this idea misses the fundamental way that the transformer works on neural networks. The output can be useful, but the mechanism of arriving there is more about probability than creativity.
An LLM cannot create true art because it cannot experience feelings, it has no continuity of being. It can only replicate the artistic patterns it was trained on; those patterns can come from true art, and can be combined in unique ways, but the only real art is in the writing of the prompt and the data it was trained on.
It’s like how the patterns of a kaleidoscope can make beautiful images, but all the creativity is in it’s construction and how it’s used.
We could conceivably extend the transformer model to include other aspects of thought, possibly even a consciousness capable of artistic expression, but it will take a lot of new work, it’s not a place we can arrive by simply adding more power or additional training to our current models.
Almost all the algorithms used by modern AI were written decades ago, it’s only usable now because compute power has made such huge gains. It will likely take many decades more to create true artificial consciousness.
What you’re describing is Clarktech, technology sufficiently advanced to be indistinguishable from magic. We don’t know remotely how to create an AI artist that can actually create original works of art with their own perspective, critique, and soul. A system like any we know how to design has to create art from what is essentially the averaging of the work of many artists. Everything they make is a work by committee. Any individual perspective is washed out in the generating process.
We simply don’t have any idea how to create an AI that would exhibit the kind of individual perspective of a human artist. Until we at least have some plausible pathway for that, we might as well be arguing about what happens if it turns out magic is real.
what happens if it’s cranking out screenplays and paintings that DO pass muster?
It’s inevitable. Eventually we will be able to ask for, and then refine, the perfect TV show for our particular tastes. Want ‘Buffy’ but set in the Fallout universe with Dumbledore and Boromir? Give it a minute and you’ll have it.
Nope. Think about the massive amount of computational grunt going into all these LLMs now, they’re thrashing AI into every possible nook and cranny, desperate to find some place that makes actual profits. There’s also a tremendous issue with gigo - AI learning on AI slop is never going to produce masterpieces.
It is definitely possible to create that. The question is, will it ever be profitable, or cheap enough to be user made/controlled? I doubt it. Tech growth isn’t just limited by what’s possible, but also by what’s practical.
Then human interest will move on. Maybe to video games, or social VR spaces… I dunno, but it’s the same principle of, say, photography zapping the attention of photorealistic painting, and other things draining attention from photography, or TV zapping novels. See azer’s comment above, worded much better than I can.
On the flip side, I think the more likely scenario is these models will always run off the rails super easy, and need humans to guide them…
Think how neat that is. What if an individual writer (and an artist helper?) could make a TV show without a mega corporate budget and production studio, maybe even on their own computer for free? What if fans could make and share TV? Think of what that’s already doing to the video game space, and we are not that far from that with current tooling like Wan 2.2.
I actually think the “it’s soulless… FOR NOW” panel is pretty important.
People who believe in the value of human creativity have been pretty casual about saying that AI generated work isn’t as good as work created by a person, but what happens if in another iteration or two it actually CAN produce “good” “art”? Like, what happens if it’s cranking out screenplays and paintings that DO pass muster? We’ve got to be prepared for that possibility, and try to act now to make sure that our world is structured around preserving human dignity on its own merits. The existence of a faster work-doing machine shouldn’t necessitate that all human workers must now starve.
A computer-generated “Van Gogh” is not art any more than a mass-produced coffee mug is artisanal, no matter how “realistic”.
This has all happened before. Take photography. People thought it was the end of visual art. If anyone can take a photograph, why would anyone spend years learning to paint?
Artists answered by pushing the medium beyond the limits of realism. Impressionism. This did not make photographs go away. But when I see a picture of someone’s cat, I don’t usually go “art!” – even though 200 years ago the mere existence of a photorealistic picture would have implied very impressive artistry.
The work that clankers are very quickly taking over is that which does not require art. Visual filler. Lorem ipsum. Corporate communications. Out with artisans, in with industrial machinery. This is the same story that has already happened to almost every artisanal trade, from scribery to pottery to smithing. Visual artists and writers thought themselves exempt from the industrial revolution; they aren’t. It will be a worsening socio-economic crisis. But it won’t “end” art. Clankers definitionally cannot, and will never do art. Not until they gain a conscience of their own.
+100. I wish I could pin this.
That being said, I think AI Bro existentialism and singularity hype has a lot of people on particular edge, beyond what the camera and other past innovations triggered, since it’s pushed at such high levels of our world. But (speaking a fervent local ML tinkerer), the proof is not in their puddin’, as professional, foundational researchers would tell you as well. Not just because of technical limitations, but because corporate enshittification is already taking effect.
People on the more spiritual side of things thought that being photographed meant trapping a part of your soul into the camera. It was a more existential creation than we give it credit for. Before then, nobody had ever stopped time.
I think this idea misses the fundamental way that the transformer works on neural networks. The output can be useful, but the mechanism of arriving there is more about probability than creativity.
An LLM cannot create true art because it cannot experience feelings, it has no continuity of being. It can only replicate the artistic patterns it was trained on; those patterns can come from true art, and can be combined in unique ways, but the only real art is in the writing of the prompt and the data it was trained on.
It’s like how the patterns of a kaleidoscope can make beautiful images, but all the creativity is in it’s construction and how it’s used.
We could conceivably extend the transformer model to include other aspects of thought, possibly even a consciousness capable of artistic expression, but it will take a lot of new work, it’s not a place we can arrive by simply adding more power or additional training to our current models.
Almost all the algorithms used by modern AI were written decades ago, it’s only usable now because compute power has made such huge gains. It will likely take many decades more to create true artificial consciousness.
What you’re describing is Clarktech, technology sufficiently advanced to be indistinguishable from magic. We don’t know remotely how to create an AI artist that can actually create original works of art with their own perspective, critique, and soul. A system like any we know how to design has to create art from what is essentially the averaging of the work of many artists. Everything they make is a work by committee. Any individual perspective is washed out in the generating process.
We simply don’t have any idea how to create an AI that would exhibit the kind of individual perspective of a human artist. Until we at least have some plausible pathway for that, we might as well be arguing about what happens if it turns out magic is real.
It’s inevitable. Eventually we will be able to ask for, and then refine, the perfect TV show for our particular tastes. Want ‘Buffy’ but set in the Fallout universe with Dumbledore and Boromir? Give it a minute and you’ll have it.
Nope. Think about the massive amount of computational grunt going into all these LLMs now, they’re thrashing AI into every possible nook and cranny, desperate to find some place that makes actual profits. There’s also a tremendous issue with gigo - AI learning on AI slop is never going to produce masterpieces.
Firmly in the dubious category here.
It is definitely possible to create that. The question is, will it ever be profitable, or cheap enough to be user made/controlled? I doubt it. Tech growth isn’t just limited by what’s possible, but also by what’s practical.
Then human interest will move on. Maybe to video games, or social VR spaces… I dunno, but it’s the same principle of, say, photography zapping the attention of photorealistic painting, and other things draining attention from photography, or TV zapping novels. See azer’s comment above, worded much better than I can.
On the flip side, I think the more likely scenario is these models will always run off the rails super easy, and need humans to guide them…
Think how neat that is. What if an individual writer (and an artist helper?) could make a TV show without a mega corporate budget and production studio, maybe even on their own computer for free? What if fans could make and share TV? Think of what that’s already doing to the video game space, and we are not that far from that with current tooling like Wan 2.2.