Midjourney can definitely lower concept art costs by 5% to 20% as it stands. Nevertheless, AI generated art currently can’t fully eliminate real concept artists.
Midjourney can definitely lower concept art costs by 5% to 20% as it stands. Nevertheless, AI generated art currently can’t fully eliminate real concept artists.
I'm not sure that's even true either. After playing with Midjourney myself, it feels like the most aesthetically pleasing outputs regularly fall within a certain range of styles/prompts/hues ("hyper-realistic", "sci-fi", saturated neon orange/blue/red, etc). Getting consistency across prompts is also a big problem.
If everyone's using Midjourney to generate concept art, everyone's concept art ends up looking the same. At that point, you're presumably back to human artists (though maybe using generated art for inspiration).
You can change it by turning down their style parameter and then forcing it off to another part of latent space with the right prompt words, and it kind of helps. The aspect ratio flexibility then makes it more interesting than other models currently out there.
Seems like you meant to reply to a comment that claimed otherwise.
https://www.youtube.com/watch?v=fAhyBfLFyNA
Would anyone in 1972, even Ed Catmull himself, be able to imagine the Youtube video you linked, broadcast globally, for free.
I think these tools are going to be big, but I feel the whole AI thing actually detracts from it. I'd rather think of it like tracing paper or photography than as a replacement for artists.