tl;dr - “art” generated by LLMs is ultimately lame and uninspiring. It’s probably never going to inspire people very much. It’s a parlor trick and everyone intrinsically recognizes it. Don’t expect to be taken seriously as a creator if this is your primary tool.
It’s generative AI though, not creative. It can literally only create what it’s seen before. It’s incapable of being original. Not that there’s anything wrong with that. Andy Warhol painted soup cans. But anyone who expects inspiration and creativity from generative AI doesn’t understand the technology as it’s applied…
I am skeptical about “never”, but right now I agree that’s true. I expect it to be true for many years to come. That being said, we have seen a lot of improvement (over even the last few months) in AI image quality, composition, and prompt adherence.
In order for an art piece to exist, an artist have to have something to say by said art. Fancy autocomplete is not an entity, it’s an algorithm to generate something looking like something else, and even if it crawls out of the uncanny valley at some point (which I’m not sure is possible), the best case scenario is that it will generate something that looks like some people did at some point. It’s not what art is, and it’s not what people look for in art. This will never change, this is the never in said never.
AGI will create art, but at this point we’re further away from it than we were 10 years ago, or even 50 years ago (and I would argue it’s a goos thing)
That’s not what I’m saying. What I’m saying is that it’s going to be increasingly difficult (for the layperson) to tell if a work is by a human or computer. You and I may think there’s some sort of moral superiority in human art, but the average TikTok user doesn’t give a fuck… and they outnumber us greatly.
tl;dr - “art” generated by LLMs is ultimately lame and uninspiring. It’s probably never going to inspire people very much. It’s a parlor trick and everyone intrinsically recognizes it. Don’t expect to be taken seriously as a creator if this is your primary tool.
It’s generative AI though, not creative. It can literally only create what it’s seen before. It’s incapable of being original. Not that there’s anything wrong with that. Andy Warhol painted soup cans. But anyone who expects inspiration and creativity from generative AI doesn’t understand the technology as it’s applied…
It generates new content based on what is trained on. not just what is trained on
I am skeptical about “never”, but right now I agree that’s true. I expect it to be true for many years to come. That being said, we have seen a lot of improvement (over even the last few months) in AI image quality, composition, and prompt adherence.
In order for an art piece to exist, an artist have to have something to say by said art. Fancy autocomplete is not an entity, it’s an algorithm to generate something looking like something else, and even if it crawls out of the uncanny valley at some point (which I’m not sure is possible), the best case scenario is that it will generate something that looks like some people did at some point. It’s not what art is, and it’s not what people look for in art. This will never change, this is the never in said never.
AGI will create art, but at this point we’re further away from it than we were 10 years ago, or even 50 years ago (and I would argue it’s a goos thing)
That’s not what I’m saying. What I’m saying is that it’s going to be increasingly difficult (for the layperson) to tell if a work is by a human or computer. You and I may think there’s some sort of moral superiority in human art, but the average TikTok user doesn’t give a fuck… and they outnumber us greatly.