Being a musing on imagination, visualisation, aphantasia, and the magic of machine-learning art—aka. AI Art. Being also, the first of several (how many? I don’t know) pieces on aiart.
Featured image generated in a few minutes using Midjourney AI with the prompt “a glowing linden tree made of ice –v 4”. Text added by the author using Affinity Photo 2. Font is Abril Fatface.
Aphantasia
“The inability to form mental images of real or imaginary people, places, or things.” —Merriam-Webster
I wouldn’t say I can’t imagine things. But I have a hard time visualising images whole.
I am terrible at remembering things precisely. Words, events, places, things fall apart and free associate. Façades break down into floating fragments, facial features fly free. Lyrics break apart into moods and fragments.
I don’t know the words to a single song. I don’t remember the lines to a single play or film. I can’t clearly visualize any painting or building or scene.
Each time I recall something, it’s a little different. With effort, I can make it cohere, but I know by now that it’s probably different from the original. If I relax, my mind remixes and reinvents freely.
I can recall things in parts, like a blind monk feeling a loxodont. Ah, here’s the tail, here’s the leg like a column, the massive belly for fermenting vegetation, the great flappy ear, the powerful ivory tusk, the prehensile snuffly trunk.
But I can’t grasp it whole.
It’s part of why I love to draw and write, I suspect. Once I put ink to paper, characters to screen, what I imagine becomes real and concrete.
I scan a paragraph I’ve written and know how a house looks. I glance at a drawing I’ve made and know how the golem singing inside looks.
Creating art lets me see the things I imagine, it lets me remember.
AI Art
Recently, many artists and ancillary creatives have discovered the raw power of machine learning models.
I got my first taste Midjourney and Dall-E 2 earlier this year. Over the summer and into autumn the quality and creativity of their output has leapt and bound upwards.
One of the thing that is so amazing for me with AI art is that … I can imagine a sentence, and then I get to see it immediately.
It’s astonishing. It’s like magic.
Before, I would have to laboriously draw or write it out to see it fleshed out … probably getting the lighting wrong … search for references to capture what I imagined and wanted to see.
Now, a minute and boom.
There it is. A version or four of it.
Vivid and real and polychrome.
It makes me want to weep with joy. Suddenly the disparate fragments of my imagination are knit together and I can see a permutation of my imaginings.
I still love to draw. But now I don’t need to draw it to see a flash of imagination turned visible and tangible. I can draw it if I want to.
It’s liberation.
Next post, probably musings on the uses and misuses of AI art.
Going to point out that I’m still busily writing and drawing away over on my wizardthieffighter patreon. I think I’ll have something pretty to share soon: Another Golden Age for the Synthetic Dream Machine. More TBA.