If you can make a compelling thing to read in a couple sentences, that’s great.
If you use it as a prompt to write a book, then I rather read the original 2 sentences than waste time with a book that can be summarized in 2 sentences.
Yes, poetry can be art too, but that’s because the use of the words creates meaningful associations to something that expresses more than the words themselves describe.
AI reads the prompt the same way that we read poetry and then delivers its attempts to describe the associations that it learned. At best, AI could recreate the role of a really obnoxious poetry critic, but it will never be the artist.
Yeah, LLMs cannot learn and that is fundamental to their base algorithms of diffusion and backpropagation. Instead, they can only be trained to asymptotically approach human communication thru the use of latent space, a method of algorithmically abstracting concepts.
They can only ever combine concepts they have been trained on, but cannot continue to build upon those compositions like actual artists do.
I reckon, the medium of literature is probably more like the author or reader’s (shared?) imagination, the words are just the chisel used to shape what you imagine, and if it works you get something like the experience they wanted.
The ai can copy the tools, but does it think about how it is shaping your imagination with them? I doubt, probably not current LLMs, only in as far as they might choose to plagiarise work of someone who had some vision, but how can we know?
I mean, that’s basically what Shakespeare did with his monkey farm, isn’t it? Or did he sell that to Jorge Luis Borges to fill the worlds most useless library?
Same is probably true of an art photograph, you’d think flock, or google street cams , or anpr cams would create loads of art photos. But they’re not aiming to captures something extra that isn’t technically in the image (feelings, mood, , culture ) but that the audience can read into it, based on style, lighting , colouring or context.
Ideally, art should express something that words can not.
Writing a prompt using words to describe what art to generate is missing the point.
would consider text a form of art. literature.
If you can make a compelling thing to read in a couple sentences, that’s great.
If you use it as a prompt to write a book, then I rather read the original 2 sentences than waste time with a book that can be summarized in 2 sentences.
Yes, poetry can be art too, but that’s because the use of the words creates meaningful associations to something that expresses more than the words themselves describe.
AI reads the prompt the same way that we read poetry and then delivers its attempts to describe the associations that it learned. At best, AI could recreate the role of a really obnoxious poetry critic, but it will never be the artist.
Yeah, LLMs cannot learn and that is fundamental to their base algorithms of diffusion and backpropagation. Instead, they can only be trained to asymptotically approach human communication thru the use of latent space, a method of algorithmically abstracting concepts.
They can only ever combine concepts they have been trained on, but cannot continue to build upon those compositions like actual artists do.
I reckon, the medium of literature is probably more like the author or reader’s (shared?) imagination, the words are just the chisel used to shape what you imagine, and if it works you get something like the experience they wanted.
The ai can copy the tools, but does it think about how it is shaping your imagination with them? I doubt, probably not current LLMs, only in as far as they might choose to plagiarise work of someone who had some vision, but how can we know?
I mean, that’s basically what Shakespeare did with his monkey farm, isn’t it? Or did he sell that to Jorge Luis Borges to fill the worlds most useless library?
Same is probably true of an art photograph, you’d think flock, or google street cams , or anpr cams would create loads of art photos. But they’re not aiming to captures something extra that isn’t technically in the image (feelings, mood, , culture ) but that the audience can read into it, based on style, lighting , colouring or context.