If someone gives me art from their kid I will inherently overlook the flaws because I am happy to see them drawing. If I learn it’s from a billion dollar studio, it will drastically change how I view it. If someone asked me to objectively say how good the drawing is I would say it’s terrible. But I’m thinking about the artist so I don’t notice the flaws. If a billion dollar studio made the art I would say “I don’t know how anyone could ever enjoy this” while actively enjoying the 5 year old’s identical art. Just replace the 5 year old with “actual artists” and it’s the same situation. My enjoyment is partially from the person behind it improving and seeing feedback.
I would never say there is no “good” ai art, but I would also say 99.999% of AI art feels very generic. If I see a regular artist draw “good” art that is generic, I will say good job because they drew good art even if I find it generic. If there’s no artist… well it’s just generic “good” art so why would I not just look at an actual artists work for diverse “good” art, while not supporting the energy/water leech that is ai images?
while not supporting the energy/water leech that is ai images?
Not as much of one as you think, these days. If you have something resembling a gaming PC from within the last 5 years or so you can probably run AI image generation locally, with how large a model and how complicated a workflow depending on your specs. Less stressful for the hardware than running a current AAA game even, depending on the model (possibly demanding more GPU RAM though).
Not false but I would say most people are not doing that, and you’ll be using models that are not as advanced and don’t have as large a dataset as the big companies. If someone wants to generate them locally and enjoys them then more power to them (though still arguable they’re using ill-begotten data)
If someone gives me art from their kid I will inherently overlook the flaws because I am happy to see them drawing. If I learn it’s from a billion dollar studio, it will drastically change how I view it. If someone asked me to objectively say how good the drawing is I would say it’s terrible. But I’m thinking about the artist so I don’t notice the flaws. If a billion dollar studio made the art I would say “I don’t know how anyone could ever enjoy this” while actively enjoying the 5 year old’s identical art. Just replace the 5 year old with “actual artists” and it’s the same situation. My enjoyment is partially from the person behind it improving and seeing feedback.
I would never say there is no “good” ai art, but I would also say 99.999% of AI art feels very generic. If I see a regular artist draw “good” art that is generic, I will say good job because they drew good art even if I find it generic. If there’s no artist… well it’s just generic “good” art so why would I not just look at an actual artists work for diverse “good” art, while not supporting the energy/water leech that is ai images?
Not as much of one as you think, these days. If you have something resembling a gaming PC from within the last 5 years or so you can probably run AI image generation locally, with how large a model and how complicated a workflow depending on your specs. Less stressful for the hardware than running a current AAA game even, depending on the model (possibly demanding more GPU RAM though).
Not false but I would say most people are not doing that, and you’ll be using models that are not as advanced and don’t have as large a dataset as the big companies. If someone wants to generate them locally and enjoys them then more power to them (though still arguable they’re using ill-begotten data)