Post by DREAM_SEARCH_REPEAT in if you use ai art your lazy

Viewing post in if you use ai art your lazy

One of the problems is the idea that AI do everything with a prompt. While at first it may look like that, a prompt is not enough to make a useful (not just pretty) but useful asset.

I, for example when I need it, I like to draw what I want in Krita and then use the plug-in 'Krita AI diffusion' to do a refinement, (Imagine is like a filter with steroids). That way I can keep not just consistency, but control in I want to make.

Creating an actual good piece of art using AI (both pretty and useful) also takes a lot of effort. Perhaps not the same as fully drawing from scratch. But what you gain is speed, quick modifications (good for brainstorming), and drawing in a style that's not actually possible by hand. For example, imagine drawing a photo-realistic sprite by hand in a couple of minutes.

There are also tools that just limit to specific stuff, like auto colouring, up-scaling, background removal, face modifications, etc. Not fully generative, but assisting but any form of AI in the end.

It's too early to say if anybody is good or bad with AI. Perhaps, if you tried to experiment with it the problem is you tried with the wrong one? I mean, one that didn't went well with what you tried to do?

I made a post in the other section talking briefly about tools: https://itch.io/t/6524872/sharing-some-ai-tools-for-artistlearners-or-just-speed...

I'd like to invite you (and any other that read this) to take a look. Is not great, but perhaps it may help a good starting point to face this new world we may be all facing.

Sorry if the response is too long. haha. I need to learn when to shut up.

Like Reply

Dickmilch8 days ago(+1)

Not at all.

Quite insightful.

I use ComfyUI myself and its so far great. But the expectation where higher than what i get as a result.
There were a lot of things i learned. Like that a Lora doesn't necessarily solve problem but maybe add problems.
There are a ton of assets that one can use and i assume most do NSFW stuff with it.

Anyway, to be really artistic or create something artistic, it takes effort and patience.

For myself, i rather draw something on a paper. It feels more real and my ability is the only thing that makes it worst or better. What i give to AI is, that it can create things that almost look real. There are of course people that can do amazing things. I try to see AI as a tool. It needs input to imagine things.

I am sure i will never be good with AI. I am too old for that.

Before, i was working with DAZ. Pretty, but expensive if you buy all the assets (which i did).
But i never found scenes (environments) that i really liked. Because like Comfy, DAZ represents are reality that is fake. Things, people would like to be. Beauty, desire, luxury.

Gee, its been a long time that i used Blender. Nice software but takes time to understand. I only briefly used it. Mostly to make holes or cut other assets.

Like Reply

DREAM_SEARCH_REPEAT8 days ago

I can tell you draw. At least, i suppose your profile picture is your creation. I like it. :-) I also use to, not often or anything, but when i need to have a more controlled sketch.

I tried Comfy too, but i found a hard time to understand it's nodes and all. Too techie for my taste. (And i am more programmer than artist haha)

I don't get what you tried to say with the Lora stuff. A Lora is not for trouble shoot. But imagine is like a plugin for a checkpoint.

Something like: you have a checkpoint that only generate anime. A Lora can be a character Lora, for example... a Goku Lora. That make all the characters in the generating picture to be Goku or Goku-resembling.

It can be troublesome if you try to use an Style-Lora in a 'All styles checkpoint' I have at least one checkpoint per style. (one photographic, one anime, one western cartoon, and so on.)

AI in general doesn't imagine things. It's more like a weird statistic calculator that returns that most probable output for a given input. (To put it in a shorter of shortes way.)

Maybe you are not bat with AI, like i said. It just you didn't find the right tool for you. Or perhaps you don't need it at all? IDK.

I don't want to sound obnoxious or like i am doing marketing. But I like 'Krita AI Diffusion' because i can draw like i normally do, select a checkpoint and control exactly how i want it to change my image, and then use my normal 'pre-ai' image edit in skills to go on. If you are a illustrator. cartoonist, designer, somebody with actual art skill, etc. I think that tool may be more useful to you.

Anyway, wish you luck.

:-)

Like Reply