The 11-20 posts

 11

Now long ago when I was playing with Ai. I discovered that it cannot even even give a Cheeto. That means that it does not know what a Cheeto is. It has yet to steal an image of a well painted Cheeto to make it so that you can have a character with red haired black or orange hair ride on a Cheeto and there be a volcano in the background. Like everything else; ai cannot fly at all and in fact sucks. So no woman riding on a Cheeto away from an exploding volcano. 

Let me tell you that ai is so robotic that you cannot do anything with wonky or funny images and so I will say that I will make this woman naked out of pure rebellion of the constraints of ai. I am to tell you that ai does not work to make engaging images but lackluster things that want to be the such but cannot.

Instead of the masterpiece that should have been: you get a girl riding a bicycle! A girl riding a bicycle and it's just so fucked up by the Liyu being there that they legs on the girl don't even show. Fucking f for the win this turn. How hard is it to do these simple ass things. Make them on sky flying on a Cheeto with a volcano in the background but ai refuses because there is nothing it's trained to produce these results.

12

You remember that AI prompt before where the girl wears a tank top and black jeans with a unicorn on the side of her pants and then is in a castle courtyard well the AI delivered in part. For this I will give it a B. What is special about this one is that it uses a Jello  world but type lora aka geltech Lora or something along those lines. Yes it gives somethings in the background a jelly appearance. 

The model for this one contributed towards the idea being followed even though it's not exactly what I wanted. The courtyards are good but the unicorn is missing and so basically ai cannot do Cheetos and nor can it make unicorn as you would want to include the . They decide the unicorn should be some odd character on the ground made out of jello or something. Like what? 

The whole thing of the jello look is just terrible and tacky and foremost you would think I was praising the look but no. It just sucks because I really wanted that unicorn on there and the ai decides to do something else instead that is not what I asked for and thinks that is okay.

13

Used a creator owned Laura on this AI set of artworks and for which pixai has been allowing people to train their own laws which means they can now make these microwavable packages to heat up in the microwave essentially. It's like you being able to have the the resources to make your own brand of popcorn. This was a step up in the later July and early August.

This Lora was made by someone of the Tumblr era and really has this younger trendy outlook that is rare. I never tried it realistically but it turned the complex theme and prompt into one where the woman looks like a unicorn pirate. Yes the black tank top prompt previously is in this mega prompt and it showed greatly.

I liked the look of things but what is wrong is that the bed she is supposed to be on is missing and the movie lobby is very crappy indeed if not bland of any identifiers .

My adjustments show everything should be but the pirate theme is welcomed though I did not continue it in the draft.

14 

This was another test run of the Milo format AKA basically you put a woman in a room with snow at a window looking out somewhere and then that's the problem that gives the images. For what it's worth that in the AI it comes up with different results when you change it a bit and frankly this time I try to make it a pizza party with anime in the background and they will be drinking 7UP but like Cheetos and unicorns and I does not understand what 7Up is at all and so basically you get some very bad errors in this one that look kind of abstract and artistic but are just accidents by the ai. 

Personally again, the ai just misses a lot and it kinda of comical. Even even though it tries to catch up it can never and really much while this mix benefits from the strength of Liyu lora's backgrounds and the piercing sharp eyes grunge look to the Milo character; it ultimately does not work to even produce what is a 7 up led party. Where is the table with the 7 up and treats!? Where is it? 

I could just draw it all but for the corrections I did something else indeed.

15 this was where I perfected the ultimate Milo combination using the original post prompt from long ago but using a newer pixai model not available from the start with the realistic, Liyu and 3d model loras created the ultimate Milo in her room which would be lended to other characters and adaptions of the same theme but differently.

Rather this one allows Milo to appear in any prompt and basically is consistent across many different prompts as long as the quality is high you get the same face and even can change the hairstyle a bit.

In sum this is the theme where she is in the snowy room with the shirt and I am to say out of all the versions this is the best mix that sorts all loras involved well to create the best image but my compliant is that there are no legs and basically they can only show one emotion. My correction below shows what I would improve and greatly it would be the face to something more naturalistic.

16@

Hmm, a continued study of the Liyu Lora in pertaining to how different models follow the basic prompt that created the Milo original. Like bruh, I really like the Milo idea and it turns out nicely amongst many of the different mixtures of 3 lora's that all contain the Liyu model or not. Well something nightmarish happened and well what? It was like the exorcist came to town.

I made the girl at the party with blue hair this time and it was interpreted as greenish turquoise to which in the minds eye the whole battle was that this hair looked very sparsely and whilst it seemed to deliver on the advent of a party with pastries and drinks only did the pastries show. It made the girl naked instead of giving her a tank top under her black jacket of course.

The corrections are many and as we approach the end of the series more will be made.

17 in the super mega compilation of images done by AI I do say I got to like these images. I did a lot of experimentation with these images and really much that use all the different features you can find on pixai but this is not an advertisement for pixai. Nope,it is a place of improvement I come for this one episode. You got the complex themed prompt ones that made the girl on her bed in the movie lobby or was supposed to but ai failed at that. You got the Milo in a snowy room but attributed to another ai creation. It came okay. Then you have the showing of how ai fucks up any image that comes from the outside to which you assembled yourself. Next you got the Milo in the jean jacket from the flying over the yellow tubes theme to which makes quite a mega.

The suggestions I have made basically imply that really much they are all inferior versions of things. Like whilst they can get somewhat neareth the things I want to see- in actual art that is not even anything good for what you see yourself your drawing many times is just the start before you discover something good. So basically are you even beginning when you cannot even to your main concept at a 100%. The answer is no. You got to have all the stuff even drafted before you start and the ai lots of times gets ahead of itself and has no program that understands this fundamental of art. The ai cannot even get the basics and essentially wastes what is promising. Surely in real art you should draft everything you want in the picture because you literally won't be able to add it later unless with Photoshop and that is to say it won't be easy.

Art concepts are not founded in ai and it's funny that from it's programming all that it is made of is grift and I do not like that and fundamentally is flawed because that is not what art is. Art is finding and discovering and discipline and skills to make ideas that appear in the dreams or the head. By the way ai could never mao human dreams but a human can do it simply.

18 ever wonder if the AI could take the concept of a book fair and make the most generic unrealistic nonsense ever well then these are the images for you.

I was original inspired by a jojolands setup which hand cutouts for characters and manga of the work on the table. This is what I was aiming for but what you get is something that does not resemble a book fair nor has the life nor the energy. There is nothing much to say but it is lazy effort by the ai.

Now what I have planned is just like that jojolands setup but for something or another. The thing is that AI just does not get Cheetos, unicorns, 7 up and now book fairs and book signings. It's just so very hard to have these things and the like.

19

In this version a woman basically follows the prompt of the woman in the tank top with black jeans with an unicorn attached. Due to the Cartoonish model and a different combination with the Liyu model you basically get a more colorful version of things but still no unicorn. This would lead into a realistic version of things that actually allowed things to be in a courtyard. Revised as the following down below.

Well I am to say what reclaimed was the original idea and basically made everything more detailed and different than what an ai can even do and basically that entails patterns wherever and things wherever.

2o

In the world of Liyu Lora; the great thing is that these things eventually with low funds can become disasters and such is what happened. Basically everything becomes corrupted and what you are looking at is that result. I would say that everything is so low quality nothing filtered out right causing these abominations. The correction literally comes from adding more points and using the error image as the base which gives insight to how it would look basically uncorrupted. 

From that I have not create the adjustments but went off my head and what these shapes could have offered and so basically I did draft three of them as one comes up later on in the series but basically  this would have never happened on me doing it by hand and the result would have been better.

Truly the limits of ai is the point system many 3rd parties engage in and ultimately you can be so hundred by the ai especially when you need to get an idea out instead you get this nonsense which itself is ridiculous.