Raw LLM Responses

Well... I guess it depends on your perspective. Look at the work of many artists, especially those who just draw OCs, I doubt that any of them approach the placement of elements of a composition necessarily with the idea that you will wonder why this leaf is in the background in this particular place - they just make a background for the sake of a background and that's it. You will never wonder why this witch character has a conical hat, because a conical hat is associated with a witch and there will be no other deep meaning. AI is good at creating exactly these cute characters at a tremendous speed, and the only big drawback is the lack of variability, because not all AIs are trained to create fantasy creatures well. A lot of people just draw cute pictures without meaning just for cutiness, and even more people just look for cute pictures without meaning to say "Oh! How cute!" and not look for the meaning of "why are the eyes exactly this shape and this color" and "why exactly this pose". And yes - it's hell fun trying to generate something from your imagination in seconds, even if it doesn't match what you have in your head - it's like a gacha game where you try to roll something cool. Or trying to procedurally process your old childish drawings to see how it looks in high detail and different styles - that's fun too. And not only childish - I'll run my own works through AI just to see how they look processed in different styles and how the AI ​​can change them based on a reference. It definitely has its own entertainment niche. For serious projects, though, yes, I don't see AI being competent enough yet. Especially considering that it can't create unique styles.