zlacker

I've played around a bit with Stable Diffusion and as far as I can tell, it's just a new tool, like a much better paintbrush.

It still needs a human to tell it what to paint, and the best outputs generally require hours of refinement and then possibly touch-up in photoshop. It's not generating art on its own.

Artists still have a job in deciding what to make and using their taste to make it look good, that hasn't changed. Maybe the fine-motor skills and hand-eye coordination are not as necessary as they were, but that's it.

>>imgabe+(OP)
> require hours of refinement

Not disagreeing with your comment but this is not the case with Midjourney. Very little is needed to produce stunning images. But afaik they modify/enhance the prompts behind the screen

>>kecupo+T1
There’s a big difference though between “a stunning image” and “the stunning image you wanted”.

>>mtrowe+O2
That's very true, I stand corrected. I see people tuning their prompts for hours on public MJ channels

>>kecupo+D3
A key difference is someone with some prompt-writing skills and a tiny amount of aesthetic taste can now compete with trained artists who actually know how to create such images from scratch. Sally in Sales and Tom in Accounting can also do art as part of their job, whenever it calls for art. And copy-writing, et c. Or will be able to in the near future. Fewer dedicated artists, fewer dedicated writers, and so on. One artist can do the work of ten, and almost anyone in the office can pinch-hit to do a little art or writing here and there (by which I mean, tell a computer to make some art, then select which of that art is best).