[Bakary]

This does not take into account the years of practice and the natural talent that led to these composers being able to do this. In comparison, any schmuck can learn how to find prompts within a few hours, and can get pretty good at it within a few weeks.

[bottlepalm]

Like I said, people trivialize and discount how hard it is to create good AI art. It’s a lot tougher since everyone now can create something passable, it comes down to vision. Is your vision great enough to create award winning AI art? Mine isn’t, yours probably isn’t either.

[Bakary]

The current type of AI art is less than a year old, so no person can have more than a year of experience, at most. Most artists spend years and years developing their craft, and it involves daily deliberate multi-sensory practice. By comparison, all a prompter has to do is play around with types of words and iterate over a few weeks. Yes, they might need some eye for detail or composition or some intuition with the sort of words that could work, but since the AI is a blackbox anyway it cannot logically be much more but throwing mud and finding what sticks.

[bottlepalm]

The real limit with AI art is imagination. The bottleneck of physical ability has been removed. The new bottleneck is imagination itself.

The more you have, the better your AI art will be. You should try sitting at a prompt and making something compelling.

[Bakary]

Don't get me wrong, I am in fact excited by this technology. The applications are clear in terms of generating new art, in particular in hybrid contexts where a person talented in one sphere only could finally make it work e.g. graphic novels

I suppose what bothers me is the idea of a prompt user seeing themselves as an artist, not because I want to gatekeep people's minds but because that misunderstands the nature of the technology. A visitor to an AI art gallery (I do mean material that is 100% the direct result of prompt generation) could be able to generate maybe not the exact same art but something of equivalent worth upon arriving at the door. The contribution of the human being or to be more precise the specific individual is almost immaterial. Sure, they might have come up with a prompt with many words but you could come up with any number of equivalent artworks in the same breath. The artist is wholly interchangeable in a way that did not exist before.

I'll admit that it does require cultural understanding of art and some understanding of how prompts work because you need to know the names of artists and styles to fit into the prompt. But beyond having a good vocabulary there's not that much more involved. By comparison, if we're talking about artwork that uses AI as a tool as opposed to being the entire thing, then it becomes much more interesting and is what I am excited about because there will still be meaningful human input in there.