> they don't actually have a serious and nuanced appreciation of art and they don't have a good understanding of all the types of work that artists do.
I would extend this lack of nuance and understanding to the deep learning implementation side also. A lot of people seem to have some very foundational misconceptions about what deep learning is and what it does. In the case of generative art: these models are “simply” sampling from the frozen statistical structure they have learned from web images and their captions. They don’t understand the relationship between objects in space, they have no ideas or feelings to express, and they communicate nothing. That’s why the even the best output of these models tends to have a perceptible hollowness: you can detect the lack of a coherent authorial intent in it.
I would extend this lack of nuance and understanding to the deep learning implementation side also. A lot of people seem to have some very foundational misconceptions about what deep learning is and what it does. In the case of generative art: these models are “simply” sampling from the frozen statistical structure they have learned from web images and their captions. They don’t understand the relationship between objects in space, they have no ideas or feelings to express, and they communicate nothing. That’s why the even the best output of these models tends to have a perceptible hollowness: you can detect the lack of a coherent authorial intent in it.