AI images - Intention or Slot machine?

What is our role as the creative.

Author

Nejc Susec
October 31, 2023 • Estimated Reading Time: 6 minutes

Hello Marvelous Creator.

Thank you for being a subscriber. 💛

Let’s get to today's topic - AI images - Intention or slot machine?

My mind is excited.

This has been a wonderfully fun experiment to do. And my mind is excited about the outcome.

We’ve been having this conversation about whether we make the images or AI makes the images.

It’s complicated, as everything else in life, and doesn’t have a simple answer. This post won’t give you an answer. What it will give you, is a side by side look at creating with different AI platforms and the thought process behind it.

I set out to make some images. Coral cities, from the gallery later this week.

I’ll share my creation process with two platforms - Stable Diffusion and Musavir AI.

Stable Diffusion

I started by choosing a base model. This already set out with my intention on the style of my images.

I wrote a base prompt (+negative prompt) and got my first few images.

Incredible city made out of corals, a wide shot portraying the grandeur and magnificence

Musavir AI

I put the same prompt into Musavir AI and chose the “artistic” mode. Which I guess could be seen as a simpler version of a model?

Incredible city made out of corals, a wide shot portraying the grandeur and magnificence

Neither of these are what I want, so let’s go further.

Stable Diffusion

I chose to add two Loras to my prompt, as well as adjust the prompt.

First Lora is for thick outlines and the second is for watercolor architecture.

Incredible underwater city made out of corals, a wide shot portraying the grandeur and magnificence, organic architecture watercolor, artbytokiame

Musavir AI

I changed the mode to Realistic and changed the prompt.

Incredible underwater city made out of corals, spreading miles on the ocean floor, organic architecture


Now these are going in two different directions. I’m definitely feeling them more.

Stable Diffusion

I decided to add two more Loras. One for details and one for handpainted brush style.

I also adjusted the prompt

Incredible underwater city made out of corals, spreading miles on the ocean floor, organic architecture

Musavir AI

I started to be drawn to the style I was getting with Stable Diffusion, so I went for it.

I changed the mode to “Anime” and adjusted the prompt.


Incredible underwater city made out of corals, spreading miles on the ocean floor, organic architecture, thick outlines with handdrawn brushes.

At this point, I am freaking VIBING with my Stable Diffusion creation process. I’m cruising through the different styles, which fill me with inspiration. As a result I’m starting to get a strong feeling for the world I’ve started to build - the style, the shapes, colors, atmosphere.

On the other hand, I feel like the Musavir results are the same idea with small changes in execution. The first images I made, I felt - wow, that looks really good. By the third image I felt - Ah, ok. I guess that’s anime?

What draws me heavily to Stable Diffusion, time and time again, is the freedom of exploration. Whenever I add a new Lora, try a new model, or experiment with ControlNET, I feel like I opened a whole new box of goodies. A vast open field where you can go wherever you want. The results can sometimes be REALLY bad, and at other times brilliant like a stream of starlight.

With closed platforms like Musavir or Midjourney, we can only walk on the tracks they provide us with. The scenery is very beautiful while we walk on them and most of the time it will keep us away from the really bad stuff. As a result, we will only see the “beautiful” things over and over again.

One is not better than the other. They are different and both can have a role in our work.

Ultimately it’s up to you to choose what works for you.

Do you enjoy creating with freedom, exploring the vast lands and all the unknowns that come with them,

Or do you prefer to take the curated path and take photos of the beautiful scenery that has been presented to you.

I hope you enjoyed today’s post and walkthrough of my experiment.

If you’re itching for more workflows, you can find written guides here.

I’m also doing 1 on 1 live classes for Stable Diffusion and Visual storytelling with AI.

You can find the schedule and topics here.

If you have questions, reach out at any time. My messages are always open.

Keep creating and don't forget to have fun. ☀

I also want to thank everyone who takes the time to click the button on these polls and even more, leave a message.

I see you, and I appreciate you.

Love. 💛