MZODERN Vs AI 2

 The whole thing about ai and these ai images is that they can be good but they are really underwhelming to what I really wish and want. The thing about the ai works in this selection is that they are kinda of lackluster but are still good. They have these unique elements aka the girl with the sparks in her hair though her clothes colorful kinda make her look homely. Unexpected of a prompt it definitely refused to follow and refused to put the woman on a bed flying through the air as smoke came from some yellow circular tubes. The great thing about ai is that it literally dumbs everything down but gives such interesting artifacts so you can build with that. Its possible with photoshop to make the image you want.

I really like the metro setting of the one girl in the jumpsuit but literally the design is that she is supposed to have pink horns which do not even appear. So basically ai has a lot to work on really. The horns that are pink where supposed to be apart of the design but it just decides it knows better based on the elements of what it is trained on and really cannot match what is even in the prompt. A grand dissappointment; the horns appear as hair oranments or worse these ugly things that do not even qualify as horns. 

Another grand failure was the advent that all women look the same and aka you literally can get a black person out of the Liyu lora which I commonly use as my Pikachu of sorts in the combination of the 3 loras you to get to mix on Pixai but you have to use another lora for black people. Which I am to say is never guaranteed to work. I love that ai cannot make the most attractive women but the most generic ones. It did take some experimentation to make the common modern look of the character I name Milo. Though if you try to make this character do something complex this quality goes away and they become in less quality. In the sum of ai having a good ratio of distance from the viewer as well not being so large into the picture gives it a good quality. Mostly the face is best when it not far away as that is what ai has trouble with and you ultimately have to refine this image to get it to appear any type of way.

My great complaint is that the boobs on the women aren't even big and well I am to guess that Liyu lora is to blame as it seems to be based on more a woman that does not have breasts and aka can be confused for a younger woman. The thing is you literally are sacrificing so much of your vision here but I like the somewhat abnormal realism of the Liyu Lora and it makes it worthwhile.

I have drawn corrections for what I ultimately want to see and that basically is to say that for all of these the corrections are greatly summed up in this one image below.