Learning by doing

General Forums off-topic talks
1

Is how most of us learn to make. . art.

For me, it took so much time just to make a little bit of improvement. . . there’s around 3 to 4 years gap between each of the pic i post below. . . And I. . started around 2014. Almost 10 years ago.

image
image1340×3130 151 KB
image
image1340×3130 171 KB
image
image1340×3130 178 KB

Though, I only draw one or two pic per month between my regular job so it’s not a surprise.

But then, we are in 2023.

We have control-net and stable diffusion. Anyone who want to go into artmaki. . .artgenerating no longer even need to learn what we called ‘the fundamental’. Taking away the pain and struggling of making art from the equation.

This honestly made feel a bit conflicted.

It’s only a matter of time until art too joining the I don't know how this work, but i don't care and i'm happy with it rank. Like how most people treat car, their smartphone, their computer, their appliance, and even this site.

They don’t care. . we! don’t! care, how it works. Only that it works, and we can use it. And maybe, we shouldn’t care about it in the first place.

But, art is not just about the result. But the process too. Is how I feel.

BUT AGAIN

If I started making art this year. I too probably wouldn’t want to learn the hard way when there’s an easy way. Because all those years ago, I too. Thinking a lot about how to get some shortcut into art making. If at that time there’s already such a sophisticated tool I wouldn’t even want bother to learn to actually draw.

I feel like we are a horse at the start of a Car age. It’s not if those generator would go mainstream or not, but when.

And when that time art has become commodity where everyone can just generate their own. Where it’s no longer something special. Would you still able to hold into your passion? making art the old way?.

Like how some mad people make their own car, even though car from common manufacturer looks better, run better, and more reliable than whatever you can build.

Me? I honestly not sure.

#yeah, i know. Another topic that ponders about nothing, filled with doom and gloom, and doesn't offer any solution to any problem. But, i just don't know where to put this restless feeling i have except here

6 Likes
2

I think I understand. Just like you, it’s the process I enjoy. Even really slow progress feels good.

So after all these image-generators take over the art world maybe we can get jobs in the circus demonstrating how people created art in the olden days – using their hands.

5 Likes
3

Think about for whom do you make it? I don’t think of using this for my paintings, I paint the way I do because I exactly want that in the way I do it. And I’m doing it first and foremost for me!
And don’t call this generated pixel-accumulations art, it is nothing more than the result of calculation/computation, thus has nothing to do with art.

Michelist

1 Like
4

So I’m coming back to art after a long time away from it. I’m also coming back with the goal of making it my new career, despite the looming AI. I’ve been spending a lot of time trying to find as much as I can about AI and generated images and to be honest, I’m not that worried.

As artists in the future will we be using AI, yes, will we still need the fundamentals we’ve developed?
100% yes!

I’ve seen AI art generated by people with artistic understanding and AI art generated by those who’ve maybe taken a graphics course in high school at best. Those who’ve really put effort into their fundamentals still shows through in their AI work. Along the way as artists, we developed a vocabulary to better describe art, and analyse art. This isn’t even to mention the areas AI naturally struggles with.

As it stands I’ve noticed AI’s biggest issue is correctly dealing with 3d space in a 2d image. For images that hold a single person that isn’t interacting with the environment, AI can produce amazing images. When that person has to sit on or hold something complex the AI can’t rationalise what’s happening in a dynamic way.

Manually editing AI-generated images is going to be a big part of the workflow in the future. If you can’t draw or have a good understanding you won’t be able to do this. The alternative is playing slot machines until AI generates a perfect image for you. There are other aspects of generated images that are obviously problematic but aren’t related to the AI part of the generation. Mainly the clean-up of the image in a very filter style way.

Something I hope they solve in the future is to make it easier for an artist to intervene in different aspects of the generated piece. At the moment AI is just a black box with input and output. I doubt this will happen with this current generation of AI, with the way it works.

I do hope the world nations act quickly to bring AI laws into place, especially around copyright. Not enough is being done to protect art. I also would like to see a confirmation on AI art not being able to have copyright applied to the generated image.

To end this rambling, I am excited about AI, but I don’t have a body of work that has been attacked by AI. Anything I produce now I do so with AI in mind. AI is pushing me to make my art as dynamic as possible, reaching areas AI can’t. If i fail in the digital sphere, traditional art will always be an option. People naturally want to see art by other humans.

4 Likes
5

How I see it people like us will still exist.
Also, humans always will want to connect.

What do I mean?

“We” (most people on this forum) see a wall of masterpieces on some site made by 12 year old Asian prodigy. Yet, we still want to learn how to make it ourselves because it’s fun. It’s not important that somebody is faster or better than we are, we do OUR thing OUR way just for fun. This is OUR adventure.

Same with people that commission or consume art.
Yeah some of them only want a picture, but most of them wants to connect with the artist too. They don’t want only the object they want to be part of somebody’s journey and have side characters in their own story.

There are sociopaths that don’t care about community but vast majority do care.

Only thing that I don’t know how will turn out is corporate business market.
I think there blood might be spilled…
I’m not sure how it will affect me personally. I hate how companies treat art and artists and I want to move to real art market either way, start Patreon, publish my own comic and stuff. Problem is majority of artists live from corporate money (me too for now…) it will hit them the most.
It’s already toxic and stressful environments I’m afraid that when they will be forced to become prompt typers and output arrangers for designers, majority of them will be forced to switch career… which is super sad…

6 Likes
6

Ahh yes. . Instead AI helping human. . it will be human helping AI finishing their commission. This remind me of this

IMG_20230125_224047
IMG_20230125_224047952×720 314 KB

5 Likes
7

Anndd that is one of my concern. . When something is flooding a market, it would ended up become consumable goods. Making it just like a snack in grocery shop.

You may have preference, it doesn’t matter who made them. Only the fact that it exist. And if you don’t like it, there’s still literally countless number of the same thing

1 Like
8

I don’t think AI can take the creativity aspect from us. What creativity AI has is taken from humans, and is still a poor imitation. I hope we can get some laws to protect this aspect of what artists do more. The sad truth is artists have it better than some industries, we just got hit in the first wave.

People didn’t like digital art because it took away from traditional skills. Yet digital art allowed us to create things that weren’t possible with traditional means, and traditional mediums still hold some great advantages over digital. Personally, AI is just another tool I get to use and explore.

If I really had to boil it down, AI art is a bad version of photography. At least photography has consistency. I don’t know if you’ve tried making images with AI, it’s not as simple as it sounds. Getting one character to be created over and over is a pain and it doesn’t really have consistency.

Photography, 3D Modeling, and Photoshop were all meant to make artists less useful. Artists just made them into tools to accomplish greater things.

1 Like
9

Problem with AI art for me is that it is extremely boring to do.

1 Like
10

There are much more advanced ways to create art, yet here we are, artists like myself are interested in traditional look and feel in 2023 and are looking for ways to replicate that in Krita. You would think that with advancements in brush engines that would kill off pixel art, right? No. Teenage Mutant Ninja Turtles: Shredder’s Revenge did pretty well! There are artists who still enjoy and creating pixel arts. AI can beat human chess players, but that doesn’t stop people from learning how to play Chess. Apps such as Poser makes the creation of human model at a click of a button, but that doesn’t stop artists from learning human anatomy and modeling or sculpting it from scratch.

AI is here to stay and it’s a tool that can help us to make better arts if used rightly. But I wouldn’t worry too much about it. When something is new people react to it in various ways, and right now I believe people are in a reactional phase and are trying to make sense of it. Once the dust settles people will just go back to life like AI is nothing. AI’s art has a distinctive look and people will get bored with it eventually. Anyway, I’m not a prophet so I can’t really say what will happen years from now. But as for me, I’m not worried. We have 3d printer and people still sculpt with clay. If anything, AI might even push people back to traditional mediums. Think of social media and the internet, now many just want to get away from it and back to the old ways of life because they’re tired of it.

2 Likes
11

From a programmers prospective, the art is in the code and creating the prompts to get a result, as much as the resulting image. AI is a tool that is only as “creative” as the people who code it and use it can make it. Computers and code don’t do creative on their own, kinda like brushes and paint :slight_smile:

For now at least, AI is a pretty blunt instrument for making fine art, that will change over time and people will adapt to the idea. I have a lot of musician friends, back in the 80’s they were all up in arms about how synthesizers were going to put musicians out of work. Musicians didn’t go away and using synthesized strings and orchestras in recordings is common and most musicians don’t take issue with it these days.

In the past the masters didn’t have AI or krita to increase productivity, they had ateliers and apprentices who worked under their guidance to produce the bulk of a work. At some point the master completes the work by adding the “master strokes”, like doing the post work touch ups on something computer generated, so it looks right to both the untrained and trained eye, in the style of that artist.

There are going to be staunch traditionalists in any type of activity, but most folks will warm up to new tools and techniques over time. I’ll bet that, if offered an air chisel, Michael Angelo would have snatched that thing up in a heartbeat :stuck_out_tongue:

1 Like
12

if the brush don’t think on it’s own yeh.
synthesizers doesn’t think on its own too. You dont ask it to make music and it spout music.

2 Likes
13

when you commission an art to someone. . . who need to get more creative. . . the artist or the client?

Now. . . when people commissioned art to a machine. . . is it human or the machine that decide most of the things.

We can tell the person we commissioned or the machine we used to do this or that, to set this this amount this or set that thing that amount. But then, it was up to them how to implement it.

Just like how you can’t a client an artist even though they can edit your janks, and do a better job at finishing it than the actual artist who moves the brush.

If i just offload everything to AI, i feel i can’t call myself artist anymore. After all, I’m just a client of its service. Yes, even if i do amazing job at doing finishing touch. In the end im just his client doing some correction.

2 Likes
14

I have so many ai generator on pixiv that used to post literally everyday, stopped a few months after they eventually got bored.

3 Likes
15

Honestly, try to use AI to generate a full person in a specific pose. I did this a few days ago, a woman looking over her shoulder holding a sword. First AI struggled to put anything in her hand, then I noticed the same silhouettes popping up over and over. One of the poses I recognised was Katarina from league of legends. There were about 4 images that were the base of the generated images, given enough time i might even be able to identify the other 3. If I want unique designs and art, AI cant generate that for me.

This is also why AI looks so good in portrait or half body, it’s mostly the same silhouette with the image being predictable. AI doesn’t have creativity at all, it’s a collage system based on what is given as a prompt.

AI fails currently in:
-Being Unique
-Being specific to the design(it’s very hard to generate a character with the same clothes and features consistently.)
-Dealing with 3d interactions(holding an object or continuing objects behind other objects)
-General Perspective (objects will often have different perspectives)

(A disclaimer i don’t promote using AI in its current form for art, i use it to keep in the know and have the skills if i need them in the future. I honestly feel AI currently is stealing designs and such from artists, and to use the art in my art would be doing the same.)

3 Likes
16

I tried creating a logo for someone based on line art of uncle sam from ww2 era. I prompted for the picture to be full perspective and for him to be holding a pipe wrench. Over half the time I got upper body or head shots with no wrench, when I did get something in his hand it went all Salvador Dali on me with twisted perspectives. Even when I got semi consistent results they were still not usable for anything and I gave up on it after a couple hours :face_with_diagonal_mouth:

I’ll have another go at it in a couple months :wink:

1 Like
17

At my work, me and my colleague took some AI generated ones (or “Artificial Ignorance” ) :D. They were posters of lions and the images had to be edited. Reason: some of the animals had extra toes. In one of the images there were at least 14 fingers on one paw!

On the poster I treated there was an image of a lion and four cubs. I tried to remove the extra fingers from the lion’s paw, I couldn’t. The solution was to duplicate one of the puppies and adjust it so that it covered the paw and without harming the rest of the image. True, it wasn’t an elegant solution…but at least the owner liked it. :smiley:

Personally, I didn’t like any of the posters, they had a lot of proportion errors, two of the puppies had heads that were out of proportion to their bodies. In the poster my colleague treated, two lionesses had leopard spots (!) on their heads and the lion appeared to have a slit down the middle of its face.

3 Likes
18

All those are technicall problem. . Which as we know. Can be fixed in a matter of months by stuffing more data into the black box.

If you haven’t heard of it. . Control-net already solved many of those janks

19

Hmmm…I don’t think just entering more information will do the trick. In my opinion, there is no “Artificial Intelligence”, as the advertising on those sites would have us believe: there is no one there making judgments and criteria, everything is based on programming lines. See this example:

Click to view example - Pinup style illustration

3-584d5408
3-584d54081920×1920 237 KB

The drawing itself has some beauty…but (I’m not a shading expert :slight_smile: ), I suppose the shading of this illustration is wrong. Or not? There appear to be multiple sources of light…

20

Which would not be unusual for a catwalk-situation/lighting.

Michelist