Steampunk Balloon Journey

Midjourney AI generative digital art

Steampunk Balloon, generative AI digital art by Midjourney/Simmy

Amazing art within seconds. How should the images be captioned? I’m hesitant to say that the image is by me. The caption should probably include the application, so as written it is by Midjourney. Although, it does take some creativity and time to come up with a theme and prompts that will end up looking good, the execution is all AI.

This is great for people like me who have a big imagination and wild dreams, but execution is another story. While I’ve tried my hand at painting and drawing as well as digital art and I enjoy it, I’m probably more of a crafty kind of person. However, I’ve always had an appreciation for art and artists and I know that it takes a lot of patience and talent.

So far, I’m having fun with DreamStudio, but I was getting frustrated with it generating images of people with cut off heads, or too many arms and hands. Sometimes the facial features are distorted and wonky or the eyes look spooky. However, it’s easy to generate 100s of images at once in a couple of minutes, and then it just takes a while to wade through the generated images and pick the right ones. There’s a way that you can take an image that you like and edit out parts you don’t like, but why do that when you can just generate more within seconds and choose among those? Also, I think DreamStudio is improving upon these issues as I type, since it is in Beta.

My son recommended Midjourney to me, saying “you’re almost always going to get images that look good”. Being the tech savvy guy that he is… be bop and bibbity boo, he showed me how to get the Midjourney bot into a discord server for me so I could play around with it. With only a few basic prompts, I got some amazing art:

Success on the first try! And wow, so many intricate details! Since “steampunk” was such a success, I decided to continue with that theme:

Steampunk fairy village, generative AI digital art, Midjourney/Simmy

I had some other ‘failures’ (things that did not look good or meet my expectations) and some more successes with a few other themes, such as portraits (prompts: “a dream of a beautiful brunette, woman, wearing a long white dress, blue sash, pastoral setting, Edgar Degas, concept art, matte painting, HQ, 4K, realistic, surreal, flowers). Where the “surreal” or “Degas” inspiration is in this, I’m not quite sure, and the blue sash is missing, but hey, I can’t really complain.

Execution did not match expectations of my imagination in my last attempt. I was imagining a place I used to take my kids to play. It was actually a pumpkin patch, but they had a playground, a children’s garden, games and fun things to do. I was hoping to capture my memories, but I just got a mishmash of nonsensical things thrown in together into the image. I also did not like the style. Too many prompts, I guess:

I fine tuned my idea to simplify things, and kept taking items out of the prompt until I was left with this image… it’s not quite what I had in mind, but much better than the first try and it did capture an essence of the memories that I had in mind:

I had generated quite a few images, but after this image my free trial ran out. The discord server for Midjourney is quite inspirational too, with amazing art by much more talented and experienced digital artists than myself. But still, the fact that high quality original images like this can be generated with only a few prompts by the average person is quite amazing. Probably these more experienced artists know how to more precisely coax the AI into generating the images that more closely matches their imagination.

This experience left me with a few questions about the future of art in general. With just about anybody being able to produce high quality art all by themselves, the market for digital artists is probably going to be flooded with competition. On the other hand, even though producing images is faster than ever, one still has to take the time to select the right image. People who spend their days being productive in other ways, don’t have a lot of time on their hands to sift through a bunch of images to find what they are looking for, it does take time— and time is money as they say. They might end up needing a curator, or decide to find a designer with a particular style that they like. Or, I’m just imagining a possible scenario of a client that says: “But I asked you for a blue sash, that’s the whole point because our clothing line is called- The Blue Sash.” Or other things that a client might say if art doesn’t meet their expectations. So I think good digital artists will still be in high demand.

As for my comment at the beginning of the article: “this is great for people like me who have a big imagination and wild dreams, but execution is another story.” I think I have to revise that statement, because execution is still another story. AI generative art is amazing, but to actually get what you have in your head in an image is hit and miss. If you get lucky, you’ll get a great work of art even if it doesn’t quite match your expectations… or you’ll get something that exceeds them. If someone still needs a very specific thing to be executed, that is still going to require some skill and talent.

Where does this leave actual artists that painstakingly and patiently use their skill and talent with an actual brush and canvas? I personally think their work might actually become even more appreciated, admired… and valuable. Future kids might say about grandma’s paintings: “wow, your mom actually painted that with an actual brush and real paint?”