Over the past few weeks, I’ve been playing with AI-generated artwork using a program (or bot) called Midjourney, and have discovered that, with the “right” prompts, it has the capacity to create some extraordinarily beautiful images. As I’ve explored what the program can do, I’ve been fascinated by the way in which it brings one’s imaginings into form, often more beautifully than one could previously have conceived.
For someone who is strongly visual, as I am, the sheer delight of being presented with four alternate images, which one can then further refine, is nothing short of magical.
The image above came from the following prompt:
Who are you? Beauty of lotus flower, spiritual teacher
The first versions that were delivered by the Midjourney AI bot were these:
I loved the bottom righthand image just as it first arrived, straight out of the hatch, as it were, and I’ve been using it ever since as my screensaver on my iPad. I’m sure there will be other uses for it too, which have yet to present themselves.
But I was also intrigued by the image at top left. So I then worked with multiple variations of that image until I arrived at the one that introduces this blog post.
For those who are intrigued about how the process works, each iteration offers four new images that are each variants on the theme of the selected image, and you can specify whether you want “strong” variation, or “subtle” variation. I first of all asked for strong variations on the image at top left, which bred this set of choices:
And from there I continued to refine the image through about 10 more iterations, both strong and subtle, until I arrived at the one at the top of this article.
It takes patience, but for me, it’s like what doing a crossword puzzle or playing solitaire (neither of which I’ve ever enjoyed much) is for others. I feel a kind of joy arising as the images get ever closer to something that I love.
For me, it’s all about beauty. For other Midjourney users, their objectives are clearly different, but it seems that whatever one has in one’s imagination springs to life in the most unexpected but delightful ways.
If you are interested in trying it out for yourself, you’ll need to create an account on Discord, and then sign up for Midjourney from there. Midjourney will walk you through the whole process by following the Get Started prompt.
I wish you as much fun as I am having. If you do take the plunge into this AI artwork universe, let me know what you discover.
And prepare to be amazed.
Note: Although I like Midjourney best for its graphics, it can be a bit more challenging to use than web-based AI tools, especially if you’re new to the AI world, so if you’d like an easier alternative to start with, a knowledgeable friend tells me you may want to try Leonardo.ai or Microsoft Bing’s Image Creator (powered by Dall-E). Both are free, web-based, generative AI tools for creating AI art, and will give you a good introduction to the field. One more to consider is Adobe Firefly, as it’s also easy to use, and very powerful.