LOG 020: What People Get Wrong About AI Art

Since I joined the AI community back in 2024 as a daily contributor, I've seen the space grow, expand, launch new careers. And at the same time I've witnessed online controversies, death threats, and a lot of things that felt confusing and coordinated.

So this video is in two parts. The first is about what people get wrong about AI art. The second is about what we, the AI artists, need to understand.

There is a lot of hate for AI art, and the original claim used to be about the training datasets that were used for training the models. Which is a fair claim, but I think it's just misleading.

You see, the fair part is that the images were scraped. That happened, and I'm not going to pretend otherwise. The misleading part is the jump people make right after it: from "the model looked at it" to "the model stole it." Those are not the same thing. And almost all of the anger lives in that gap.

Steve Jobs once described what a computer actually is. He said it takes very, very simple-minded instructions (go fetch a number, add it to this number, put the result there) but executes them at a rate of a million per second. And at a million per second, the results appear to be magic.

That's the whole trick. It was never magic. It was something simple, running faster than you can follow.

Which is exactly what people are reacting to with AI art. Not that it's doing something new. That it's doing something familiar, too fast to watch.

Because art has always been about learning from others. That is the way to learn anything. First you study the fundamentals, you copy the masters, and you evolve, finding your own voice.

The computer just makes it faster and productizes it.

I have been in the creative industry for about 20 years, and drawing since I was a child, something every creative always claims. I was never great at drawing, and I couldn't make a career out of it, but mainly because I had a bigger passion, which was to make art with computers. From drawing in MS-Paint, to creating my own animations in LOGO, to Photoshop, Illustrator, 3D, and now naturally AI.

Generative AI, for me, feels like having a camera into my mind. Before, I had to learn a lot of techniques to generate 3D images, have a big GPU, and wait hours for renders.

Now I can do that in seconds. And for me, that is progress. And it's progress that is not going away. People underestimate the fact that these models will get better and more efficient with time, and we can see that happening in the span of months, sometimes weeks.

On one hand, artists that enjoy the craft will have more leverage. I believe that besides the haters online, people that don't listen to the noise and just focus on their craft will become more valuable.

In the concept art community there were also big debates about kitbashing. And then when the top figures came out as using kitbash, that made more scandal, and suddenly it became common practice in the industry. Because why not have a tool or a skill that makes you work faster and get better results. The scandal lasted about a year. The technique is still here.

On the other hand, AI art is enabling a lot of people to explore their ideas. I've seen and been in contact with amazing artists in the AI community with backgrounds in statistics, plumbing, print work, whatever, that are now exploring their creativity using these tools. They are learning, creating amazing content, videos, short films, contributing to the pool of art.

It is opening new opportunities.

When computer graphics came out in Hollywood, they didn't replace entire industries. They enabled a whole new industry of visual effects at scale. Can you imagine the Avengers movies done with special effects from the 50s, with clay motion?

We can all enjoy much better content because of this technology, and AI will open up a whole new world of creative output.

Think of one-man studios. Production companies of one. Fifteen-year-old kids creating the most spectacular movies. It will not be long. In the next five years, an AI director is going to emerge from the dark corners of the internet and have their own Toy Story moment.

And if you want to be that person, this next part is for you.

What AI Artists Need to Understand

The biggest problem now with AI generated images is that it creates too many options. The output is just too big and too fast. And that output is now spreading across the whole internet at lightning speed.

There is a lot of slop being peddled. Some of it is extremely entertaining.

There are a lot of artists that don't mention they use AI (although you can tell) and people don't care as long as it's not obvious. Quality and production matter a lot. And the trick is always: not making it look like AI art.

There are obvious tells. Too sharp an image. Too saturated colors. Characters that look like celebrities.

So the "AI artist" now comes with a whole new set of issues to tackle.

First of all, the landscape is changing constantly and it's hard to keep up. But the biggest MOAT you can build as an AI artist is, in my opinion, a combination of personal identity and brand with a signature style that people will recognize as your own.

Those two can live separately, or in some rare cases become one, where the art is so unique and so hard to vary that it's impossible to separate the author from the piece.

The "work" stops being individual pieces. It becomes a flow of art. If producing amazing images with Midjourney or Seedance or Nano Banana is that cheap, then the constant output, the stream, the timeline, becomes the art.

The curation becomes the edge. And the taste, the decision making you use to curate, is what people start to resonate with.

We are playing a game where we have to compete with millions of people across the entire planet, fight algorithms for attention, and navigate different platforms and audiences.

That is why I'm doubling down and promoting the concept of world building as a tool that can help this new generation of artists stand out and have a clear direction.

It is one thing to generate random images and change styles every day or every couple of months. It is another thing to create a unique place, a world, an aesthetic, that you can explore endlessly and share your findings with the whole world. You can learn more about that on my website, VVSVS. Link in the description, I go deep there.

Building worlds will also prepare you for what's coming. World models are already arriving. You can walk around inside a generated environment today. When they mature, the models will do most of the heavy lifting, but you will have the fundamental knowledge to create and expand these worlds, and the intellectual tools to extract more out of them.

AI art is not here to replace old artists. Its arrival actually brings more value to those who want to dedicate themselves to the craft of art, whether that's drawing, painting, or writing.

The human just got more valuable.

AI art is here to enable millions of people to explore their ideas and share them with the world. And I think that is a beautiful thing. I love seeing what people make when nobody ever tells them what art is supposed to look like.

The challenge now is how to market and present your work in ways that quickly differentiate it from the old way of sharing creative content. And that is a topic I'll explore in another video.

For now, just go create, play with the tools, and share your findings.