My first month with art generators was focused on creating art products, it was a success both of my test stores were approved.

The art I have generated the first week, was good, and cute, but there was something to it, it was too ordinary I guess.

The second and third week, I made my art more whimsical, it was wilder, and weirder, it had emotion, it was not ordinary.

The newest batch of products adds color, mystery, aliens and it focuses on eyes, those windows into the soul.

At first, I wanted to create upper body portraits, but because I kept telling the AI to enlarge eyes…

It was just generating faces with eyes, leaving out most of the hair, and neck.

I think this is the right way of presenting large eyes, and I ultimately agreed with the AI.

But if an instance like this ever arises, I will let these AI programs or their wrappers do what they need to do.

I won’t try to coerce the AI to show the upper body, and maybe some window in the background, as the case is with Mona Lisa.

AI is primitive, the programs written around it, are aimed at common use cases, and are thus limited.

Don’t argue with primitive software, the computer always wins this one.

There is a way, I haven’t tried yet, and that is to take whatever AI generates, bring several samples into Krita or GIMP, enlarge the eyes, pull up the chin.

And then give the AI 5 samples to remix them, merge them, re-build a non-deterministic union.

Beyond that, you can train an AI, but adding manual labor, kind of takes away from product generation.

It is better to stick to the text prompts, and take what you can get.

This way by rejecting manual labor, you can make 5 products in 20 minutes, and that includes up-scaling, generating product text, and uploading to some store.

I didn’t publish it, but I took one of my paintings, and transferred the design to a mouse pad, it looked great.

The one thing about these e-commerce platforms, is that you upload one image but they allow you to create many products.

My first store has 30 images, if I made notepads, magnets, mousepads, towels, paper plates, phone cases, luggage tags, aprons, stickers and holiday ornaments.

I’d have 300 products, and possibly my first sale, search for how to sell digital art online and you’ll find the same stores.

In my fourth and fifth weeks, I took a moment here and there to see if I could generate powerful art.

Something that really stand out on a wall, and I did go to the internet, and grabbed 5 images of interesting faces.

I chose widely spaces eyes, because I like pop-surrealism, and it worked, it is not something I would want to hang on my wall.

But it is something that I would expect to see in a museum, those are unusual and interesting faces to look at.

The world of artificial intelligence generated art, is not just about refrigerator magnets and cool aprons.

Soon we will see galleries, that will prefer unusual and interesting AI art, over what is manually painted.

You can make 30 unusual and captivating 60x60 inch paintings, to fill a spacious gallery in just about a week.

If gallery owners will reject AI art, then new galleries that embrace it, will make a lot of sales.

In closing, the AI can’t consistently generate interesting art, you have to figure out how to push it coerce it.

And the difficulty level is just about the same, as painting with the aid of a wall projector.

The configuration and the setup takes time, it is basically a form of art, that is very similar to collage.

AI is just another art medium, like stencils, or tagging, or slapping (where you slap stickers everywhere), everyone should know how to do it.

And the nicest way to learn, is to aim to print it out on canvas paper, frame it, and hang it up at a coffee shop.