Why your art sucks and AI-art is better

https://praveen.io/posts/this-is-why-your-art-sucks-and-ai-art-is-better

1 9 1 year ago

What happens when we mistake bias for taste in a world where anyone can create?

WorldPeas• 1 year ago

guy creates bad art, doesn't want to practice
ai is able to skip right to the middle of the skill range, but original user has no capacity to modify the work and it doesn't have a distinct style

art is about doing less and making people think. What you did here "gets the job done" it is not, and never will be great. Maybe that's the point, there is a place for that.

praveeninpublic3• 1 year ago

“Doing less and making people think” is exactly what I did.

I did something quick and made the readers of my graphic novel think rather than worry whether my art was real or AI.

codingdave• 1 year ago

This doesn't account for the fact that most people do art for their own reasons. It is cool if it gets attention and people like it, but that doesn't make the results of people's own personal creative practice "suck". The number of upvotes and the comments on reddit is a terrible way to evaluate art.

tolerance2• 1 year ago

This doesn't account for the fact that most people do art for their own reasons. It is cool if it gets attention and people like it, but that doesn't make the results of people's own personal creative practice "suck".

I think that this is the point that the author is trying to make (see the section "The real risk"), just in favor of AI art.

tolerance2• 1 year ago

The author does a fair job at indicating how contrived people's interpretation of art and good quality becomes when they evaluate AI art against non-AI art and that the reason your art sucks and AI art is better is because—as indicated by another comment—"you" aren't putting in the time and effort to get better and produce high quality work and AI is programmed to be used to get to that quickly.

People who are willing to bridge the gap between their effort and AI's instantaneous (though often imperfect) productivity shouldn't be penalized in the court of public opinion.

rvzx• 1 year ago

I want, and I think you might agree, for you to be the best Praveen Kumar you can be. To produce meaningful, valuable work that represents your taste. Is your current perspective getting you there?

I think you might be wrong about why people are dismissive of AI art.

Creating effective prompts is a truly valuable skill. It's one that writers and leaders have had to learn for centuries. But it was different; we used to only prompt biological neural networks.

Generating quality AI art is a skill! But it's a writing skill and a critical thinking skill, not a visual art skill.

Previously those of us without skill but plenty of taste would prompt artists to produce what we couldn't.

I'm reading through your other posts and I think you might agree with me in principle and that you might argue that this doesn't apply to you.

As you say: "This isn't about replacing creativity. It's about finally being able to complete something that was trapped inside my head and the skills that I needed to have a polished outcome is partly non-existent."

I relate to this, and I think that's exactly how we should be using generative models. Augmentation that helps us achieve our goals more effectively than we could alone.

Previously we did this through things like apprenticeships, artist collectives, and other person-to-person collaboration. I don't think those things should go away. I do think we can get even better at producing art by including generative models in the collaboration process.

But I'm skeptical that you are truly achieving your goals.

"I don't outsource my work, not my creativity."

You did. You did outsource your work. To a tool.

Is this bad? I don't think so. It happens in all creative work.

Is it understandable to me that people would be dismissive? Absolutely.

Here's the challenge: to really get people's attention now we have to produce work that is at least as compelling as the very best art that can be generated by models. And really? We need our creations to exceed what current generative models can do.

We need to do this to be successful as artists. We also need to do this so that when the next batch of models are being trained – with or without our consent – they will produce higher quality work.

As I'm looking through the pages of Meditation at the End of the Universe I'm seeing some really nice imagery. I'm not seeing work that I'd deem impressive without using generative models. This is subjective! My opinion doesn't really matter! But I think this might be what other people are seeing as well.

Feeding a stack of art like this to a model in training is not likely to improve its abilities. There are layout and clarity issues. There are glaring color and typography issues. There is a feeling of heavy-handedness, of mismatch between concept and presentation, of gratuitousness. A lack of constraint or refinement.

This is not a bad work. It's an early work. And that's a really good thing.

But I need to say something very direct.

Your blog post is not good. It shows you have no idea what you're talking about.

This is why people are dismissive.

This blog post shows very clearly the thing that your artistic work hints at: you need more experience.

This isn't a bad thing. It's completely normal.

But a blog post like this? Counterproductive at best.

This work is still not matching the expectations people have for an impressive graphic novel with or without the use of generative models.

That's totally okay. It doesn't have to.

But you need to hear that generative models are providing a false sense of achievement, and that there are artistic skills that you still need to learn to produce really great generative art.

I want you to produce the best Praveen-style work that you can.

Bruh, this ain't it.

Keep going.

praveeninpublic3• 1 year ago

Thanks for taking the time to write such a thoughtful response. It’s rare to get this kind of engagement, especially from someone on the other side of the AI-art conversation.

You’re right about a lot: the work is early, it’s rough in places, and there’s still a lot I’m figuring out. But what looks like lack of polish to you feels like forward motion to me. The alternative isn’t mastery. It’s paralysis. I’d rather be making progress than waiting for perfect conditions that may never come.

You called the blog post counterproductive. I don’t think it was. Without it, I wouldn’t have sparked a discussion like this. The post wasn’t trying to make a universal claim, it was just me writing down what I noticed, and trying to understand the reactions I got. That’s been useful.

I get that the work doesn’t meet everyone's standards. Some call it lazy because I used AI. Others call it unfinished because I didn’t use AI enough. That tension is actually what I’m exploring, how to use these tools just enough to get the work closer to what I see in my head, without losing the part that feels like me.

I’m not building this for a gallery. I’m building a record of progress. A timeline. Something I can look back on and say, "I made that, and I kept going".

I might not agree with everything you said. But I’m paying attention. And I’m still drawing.

I am running a 90 days challenge to learn, explore, challenge my past beliefs and build systems that will help me accelerate. If you're interested, you can follow me here on X: https://x.com/PraveenInPublic/status/1917824031660859667

whobre• 1 year ago

My art sucks because I am not talented. Nothing to do with AI

Why your art sucks and AI-art is better

Related Stories