← All logs

LOG 023: High Quality Slop

I don't feel emotions from AI art so much anymore. Every tool is so good that all the work looks the same. Why I think we need to slow down, and what the next chapter is.

LOG 023: High Quality Slop

I don't feel emotions from AI art so much anymore. And it's tiring to keep up.

A new model comes out. Boom, now we can do this. A new workflow. A new MCP server. A new way of rendering images. A new way to edit video. All the time there are new things coming out, and we move so fast that it's hard to keep up. I'm getting overwhelmed with all these new tools.

And to add to that, all the tools are now so good in quality that everything looks the same. In many ways, all the work from all of us artists who use AI somehow has the same look. A little bit surreal. A little bit 2D. A little bit two and a half 3D.

I'm just re-rolling

It's so easy to get really cool images. The level went up so much that I feel, at least myself, I'm just re-rolling, sharing work, and it's just contributing to the slop.

I'm not really taking a second to slow down and think: okay, what is it that I want to do?

What I've been doing lately is trying to engage in a bit more ambitious work. Something that is going to force me to slow down. A project that makes me really think what I need to do.

For example, I need to find a hero for my adventures. I want to have something that people recognize. Because what I see a lot now is aesthetics. Amazing aesthetics. But there's no story, there's no hero, there's no world, there's no places that I see and recognize.

I think that's missing.

We're just testing

Right now, the current state of AI art is everybody sharing little tests, little aesthetics and little explorations. It's a lot of exploration. There's no edge, because all of us are doing the same amazing images.

At least on my end, the solution I'm proposing for myself is to have a really big project to focus and apply all my learnings. I want to really create a world. Something that's been in my head for a long time. Something very ambitious, where I actually need to make decisions.

I think what we're not doing right now is really thinking what we want to do. I think we're just testing. Oh, what if I mix one celebrity with medieval knights? Obviously that's cool, those crossovers and cross-pollinations. But it's just that.

What is really missing is starting to appear, recognisable characters with these little AI influencers. But still we're not drawn into the world. They're more like gimmicks. Having a bit of a foundation, a structure, a basic storyline, some character or characters that drive the audience from one place to another. I think that's going to be very helpful.

I don't fall in love with anyone

And that's another problem I find. I'm starting to look for these characters, and I get an amazing palette of character sheets.

But I don't fall in love with anyone.

I think that's something that takes time.

We need to slow down

I think what I'm trying to say is that we need to slow down.

The more we slow down, the more we counteract this slop epidemic of really high quality slop. We can create something that is a bit more memorable, I would say.

And I say this to myself:

> Take your time. Think. Have an idea first that you want to communicate. Then bring that to the world. Because now the aesthetics part is solved. It's easy.

Now it's up to us to bring all these worlds and characters, to create them and make them live in a world that is ours. That is yours. A world where we can really fall in love with characters. With a space. With a story, with some drama. Where we can really try to generate emotions.

The 95% is done

I don't feel emotions from AI art so much anymore.

Before, I was just in awe. Look at all these worlds we can explore. Look at all these aesthetics that the technology is giving us. But it's really hard. When I get excited about something, it's an excitement that lasts maybe a week or two.

When Midjourney was coming out with new improvements, at least I felt a lot of emotional connection with tool, the output and the collective exploration. But now I think we've hit a wall. We're just going to be able to increment in maybe little percentages. A little bit, a little bit, a little bit of improvement. But the 95% is done.

I think now a strong moat is going to be: can you generate strong emotions in your audience? Can you really make them feel something with AI art? Something that makes them forget that they are looking at AI art, and really connect?

I haven't seen anybody do that.

I'm not saying that I'm going to solve it. But I think that's the next chapter.

This log started as a rant on camera. I will post the full video this week on my YouTube channel.

Links and references mentioned

If you want to go deeper