Anyway.
You still need someone to generate the prompt. To pick up and refine something. And you need someone to appreciate the result.
So... in this light, I am mixed (speaking from my musician/composer point of view - where this will have an impact too already).
On one hand, it's pretty exciting.
It looks a bit like a big synthesizer/arranger, for pictures, where you have like, very specifiable presets you can choose from, and where you can filter/distort the result at will, with words rather than knobs. Only, you don't _know_ how the output reacts to the input exactly.
Ok, you can get ideas from this. Or quick rough works where you can crossover concepts and ideas to see how it could go.
On the other hand, it's meh.
If you consider an artwork as a thing/product/content only, it's a bit worrying. Because there will be a massive shift of "producers" and consumers moving toward these generated/optimized pieces content. Just look at how people adjust/abide to the rules of social networks to market their content (because it's become content at this point, it's not art anymore).
Remember that the whole mess where we are today comes down from the fact that we, as a species, try to quantify, then over-optimize every aspect of our lives, for profit: money wants even more money. And it became even only worse with the internet.
If you consider an artwork as a medium - between people, across generations. Then it's a very exciting additional tool in the palette.
I can’t give a great painting example, but in the world of music, should The Beatles get a cut from every single band they influenced, which is basically all of them.
It seems like of artists went down this road the AI developers could come up with a pretty strong defense.