I’m similarly bored with “AI won’t have any meaningful impact on arists” takes.
Think of a well-reviewed game like Slay the Spire (https://en.m.wikipedia.org/wiki/Slay_the_Spire). Excellent gameplay, amateurish art which they had still to commission for ~thousands of dollars.
Now they can get better quality art, for pennies, in minutes. Is this a boring example not worth discussing?
It took me a couple of hours to get 5 generated images that had a semi-consistent theme and look that I wanted -- for PowerPoint slides. I don't see how AI art in its current state could possibly be used for something like the art in Slay the Spire. I challenge the "now... better quality, for pennies, in minutes" assertion. Maybe in a few years.
Many people take longer than 2 hours to learn how to do simple things on a computer. Especially in the past, when UI wasn't as good, many people had serious trouble using computers.
I think we can both agree that computers were a worthwhile technology that was worth spending time learning to use.
The fact that playing with these models for 2 hours a week after they came out isn't enough, does not mean that other people, given time, can't use them for crazy things and/or build crazy things around them.
So the alternative to paying an artist is... paying a stable diffusion jockey*? I'd rather pay an artist, they tend to have unique and consistent styles that can be evaluated a priori.
If your google-fu isn’t good you can hire a researcher. In general if you really suck at using a tool then you need to hire someone, it’s not the tool’s problem.
And yeah, I’m not sure everyone would prefer paying a human for 100x the cost at 100x the delay.
AI won't have a meaningful impact on studio artists. The art market values artist labor and persistence over a long time more than "cool visuals". When you invest in art, you're betting that the artist will keep going, laboring in their studio and producing new hand-made material for years to come. Technology that promises to eliminate the labor out of art won't have any impact on this.
Stable Diffusion is to studio artists what motorcycles are to professional runners. It's besides the point.
That said, I think SD will impact commercial gaming, marketing, etc., where cutting out labor is part of the goal.
Think of a well-reviewed game like Slay the Spire (https://en.m.wikipedia.org/wiki/Slay_the_Spire). Excellent gameplay, amateurish art which they had still to commission for ~thousands of dollars.
Now they can get better quality art, for pennies, in minutes. Is this a boring example not worth discussing?