Artificial, yes. Intelligent, not so much.
1top of clouds seen from a mountain top paintee gold by the morning sun on the horizen.
2top of clouds seen from a mountain top paintee gold by the morning sun on the horizen. make the clouds flatter so they look more like an ocen. Mountains are not supposed to be gold. clouds are supposed to be gold on the sunward side.
3top of clouds seen from a mountain top paintee gold by the morning sun on the horizen. Clouds should look like an ocean without being piled up in places. Mountains should have only peaks rising from the clouds like islands in a white sea with gold tinge.
Ok, so the bot wasn't there. But the third render is almost identicle to the second. How hard is it to get that the clouds look like a sea?
Comments
prompt to further experiment with:
panoramic view of three mountain peaks poking through a thick cloud layer, view from above, sunrise,
or replace clouds with fog:
panoramic view of three mountain peaks poking through a thick fog layer, view from above, sunrise,
Daz AI Studio: Prompt Engineering Guide - Daz 3D Blog
Yielding the exact number of some subjects may prove challenging because the AI may not have been trained with numbers for those subjects.
SDXL does understand
solo, 1woman, 2men, 3women...
###
panoramic view of 3girls poking through a thick cloud layer, view from above, sunrise,
AI hates fingers.
They should teach the programme to pay closer attention to the details of the prompts. Yesterday I had a picture on the subject of 'buffalo hunting' showing Native Americans riding along comfortably on buffalo instead of horses... LOLs
apparently possible
The question is, how did those buffalos get to north america from africa?
They're good swimmers !!
I also did not know that, so I have used a prompt:
buffalos swimming from africa to north america
and got some clues...
Botticelli?
This one comes closest. However I was on the closest mountain top, there were no dragons in San Francisco that day, the clouds are still way too fluffy and when I say gold, iI mean those clouds were painted with 24 karat.
Odd that AI came up with 1 bisonamoung the displaced buffalo.
Bison, buffalo - it's all the same thing ;-D
But I like your dragon picture. I've just discovered my interest in dragons. So if AI can't do anything with buffalo, the programme can do dragons. I've just tried it out extensively.
Artificial and intelligence are just "oxy-morons". Sure, real artists will use this new tool... someday.
The state-of-the-art, at this moment, is by essence copycatting. LLM are trained on existing corpus, it's probably useful for biochemistry, physics, boring paperwork, but as for CREATING it's just plain useless. As humans, we share many intellectual traits, like any animals of the evolved primate species. A genius is someone who is able to transcend our experience. There are very few of them, and many paid a heavy toll for this. Think about Cantor ! (https://en.wikipedia.org/wiki/Georg_Cantor), or Oppenheimer for a more recent example.
These LLM are certainly able to create as many Botticelli, Da Vinci pictures, just as they are able to create new Bach cantatas or Philip Glas ad libitum, or ad nauseam, they probably never be able to INVENT new stuff.
At this moment, I'm reading a biography of Louis Ferdinand Celine a very controversial French writer, his novels without his REAL experience of life would never have happened. His medical thesis was about Semmelweis (https://en.wikipedia.org/wiki/Ignaz_Semmelweis) who was ridiculed by all the medical society at his time...
On the contrary, the AI loves fingers and enjoys finding new and creative ways to mangle them!