When you first see what image-generating AIs can do, you feel bad for real artists; after playing around with it, you envy them.

Spot on here! I've only just been using stable Diffusion for a couple of weeks now to help me visualise characters and locations in my world building. It's such a great tool when you really have no artistic skill. But the limitations soon become apparent and a lot of problem solving goes into trying to regenerate the simplest things.

Can generate a near perfect image in a minute if the prompts are right and you get lucky. But the details take hours, where an artist would be able to simply visualise and draw it in.

I think the key is to develop basic skills to draw a really shit mockup of what you want, then img2img it from that... I'll get there, maybe.

Exactly. At first glance it can spit out some really impressive stuff, but turning that content into a coherent piece of artwork still takes imagination and skill.

I can't draw very well, but I've gotten good at compositing and image manipulation over the years. SD is amazing, but it doesn't mean that I don't spend hours and hours piecing together an image to be the way I wanted it.

I've always explained it like this:

Every time you press that button, you'll get an image. Maybe even a really good image. But it will never be the image you had in mind.

But sometimes you're restricted by your own imagination and the image produced reflects your desire stronger than your vision. The ability to consistently generate meaning and purposeful images with generative ai is an art in itself. It's just a different medium for artists to use.
- I've only found that true for days I am prompting without a specific image in mind. The second something specific gets in my head, and even hours of fiddling and tweaking prompts won't get the result that satisfies me (though I might get a bunch of cool tangential output along the way).

AI tools are support for professionals. For non professionals they are a toy, a magic box

The question is, could one million people reaching into the magic box at once pull out at least one thing better than a professional could make in the same time?
- It is the same as science. Many discoveries are serendipity, they happen by chance. You need a professional to understand that that is a new discovery.
  
  Anyone who manages to make something significant with AI is a professional.
- In the same time? Yeah absolutely, I've put together some really cool pieces of AI art in like 5-10 mins. A "real" artists just couldn't match that pace.

I don't understand what this means.

I believe they are saying that the skill it takes a real artist to produce art is something to envy.
- yea, specifically I envy the ability to form an original idea and then using your skills to create something exactly how you want it to be. Must be nice. Using those AIs is just unsatisfying, it's never your thing, never your style and never your idea, never your vision, never your art – even if it might look kinda cool sometimes.
  
  I tried really hard to "create" an older Version of Daria (from the MTV cartoon, so not even an original idea cuz I cannot form those) standing in front of a billboard in a slightly dystopian environment. Despite numerous attempts of prompting, the best I got was this hyper-attractive girl in a green jacket.
- AAAAH!! Yes, that makes sense, thank you~

For data workers, AI is a lever, and a damn good one.

Whenever people call AI the next segway it's an immediate flag that they don't work in a field that uses it.

I'm not sure what you are trying to say. Are you talking about a segue or a segway? A leaver is much more useful than a Segway but they are both completely different things. Why are you comparing a machine to a leaver? Also, the Segway as a product wasn't nearly as popular as the inventor hoped. Are you saying people who think AI will flop like the Segway did are wrong or are you saying that the Segway is some super tool and AI is not?
- I honestly can't tell if you're trolling, idk how you could even come to the conclusions you did with my response.
  
  AI isn't going to replace us anytime soon. Data entry will be first, in the next 5-10 years most likely. When I say it's a "lever" what I mean is that for people like myself who work in IT, AI tools have multiplied our efforts and allowed us to spend more time doing and less time researching how to do. Even when it's wrong, it's usually good enough to get me on the right track and save me time.
  
  The Segway analogy was intended to illustrate how the average person who doesn't work on or around computers sees the invention. They think it's the next Segway (as in, a useless item that is supposed to change everything but disappears in months) because they don't have an actual use case for it and spend most of their time trying to make it write smut or whatever else they're dicking around with.

Also, all of the time that people spend tweaking their prompts could just as easily be time spent learning to create their own art. The distance between absolute beginner and competent artist isn't as vast as some people would like to believe. It's just intimidating to a lot of people to get up the will to even try.

Nah dude, I can draw somewhat competently, but it takes a lot of time and dedication to get there. Tweaking prompts is orders of magnitudes easier and quicker than actually drawing something neverming learning to draw something.
- I was awful at drawing and painting and now I'm not. I've walked that path, and I'm glad I spend the time learning those skills. I much prefer creating whatever I want to as opposed to pecking a word at a time, trying to coax something vaguely like it from an algorithm. That seems sad to me.
I've heard a prediction that in the near future there will be courses on creating AI prompts and that it'll eventually become a job.

I find that even with the best diffusion models, replication and character sets are virtually impossible. Also text does not work.