[Stable Diffusion] AI Image Generation: A Soberly Viewable Psychedelic Space [with various images]

The AI-generated Coca-Cola commercial is one example, and even looking at the AI-generated video of BAD APPLE, I felt that it has a high affinity with psychedelic spaces. If you're going to leverage AI, psychedelic spaces are perfect.

Stable Diffusion is amazing.
It seems to be quite useful, especially in the CG industry.

So, I'll post some psychedelic images, and then I'll post AI-generated humans and AI-generated art.

Jimi Hendrix specified.

Bob Dylan

Bob Dylan

I didn't specify anything, but it turned out like Paul McCartney.

There's a sense that past art techniques are becoming out-of-place artifacts (OOPArt), so I want to revive them with the power of AI.

Now, what about realistic humans?

Until recently, the finger modeling was a bit crude, but the quality is quite high now.

Now, what about paintings/art?

This is the first image I generated after starting to use Stable Diffusion, and I created it thinking that cyberpunk-style art has a high affinity with AI. It's a bit unrefined, though.

So, let's try generating Teto Kasane, who has a strong open-source aspect.

What happens with a different art style?

It turned out with a vibe from around the 2000s.

Thoughts and Purpose

Originally, I wanted to generate product images of things I hadn't dealt with before as concept art, but the text generation was particularly poor, and there's a lot of information targeting human figures, so AI for "things" is still largely unexplored territory.
I wonder if text issues could be solved with Photoshop or similar tools.

This time, I tried it out because prompt engineering and whether AI generation could really become a profession were hot topics, and I'm convinced. By the way, if you want to create a psychedelic AI free space without specifying a character (like Teto Kasane, which might involve around 147 phrases to describe), fewer prompts are better.

It's probably instructive for people in music-related professions or CG creators, like those at Pastime.
Pastime Factory

Tracing the history from the domain, it seems the site itself has been doing CG design for characters since around the early 2000s, so their work is probably in that field too.

Eventually, though, since Amazon and Google provide cloud GPUs (which are quite expensive), I thought it might be possible to create a next-generation search engine using AI and cloud GPU power. After all, as long as someone isn't an otaku (obsessive fan), regardless of who wrote it, they just need to get the text, images, etc., that contain the information they're looking for.

If everyone were an otaku, they would seek human-like qualities and things with taste, but if not, the future results might make us wonder if humans are really necessary.
In a way, it's like, "Humans, try harder!" kind of vibe, isn't it?
It reminded me of The Clash's "White Riot," which was used in the context of racial issues at the time. (The author does not hold discriminatory views towards any subject.)

I'm a supporter of historical things like photographers and artists, but I wonder if by actively engaging with this kind of AI, we can discover the inherent value of humans.

...Having said that, the things I generated are too cute.