Dear Specialist L.O.M.,
Here is my email response to your questions in the thread concerning “Dancing with No One.”
[THREAD ID:1-6J4PG5E]
Following your four items:
(1) Per your request two best editions of this work are en route to the copyright office, and are bundled with a paper copy of this response. I apologize for some confusion on my part here (FedEx Tracking ID 779253132776).
(2) On the phrasing of “book design” – the book was initiated as a catalog for works on exhibit, but the specific sequence is part of the work. This is not a random pile of pictures, but a deliberate poetic narrative. Should I replace “book design” with “sequencing” or “narrative sequence”?
(3) The creation of the illustrations in “Dancing with No One” included the assistance of what are now called AI tools, but are a part of my much longer artistic practice concerning computers. I have been an award-winning computer-based artist since graduating from art school in the 1980’s, including winning the best-computer-animation Clio for three consecutive years, have author/edited three best-selling computer graphics books, and was a supervising artist for groundbreaking computer-assisted animated films including “Toy Story” and “Final Fantasy:The Spirits Within.”
In the early 2010’s I began working more deeply with machine learning algorithms, since my background writing code and patents at NVIDIA had exposed me to them – I even managed to study with Geoffrey Hinton, the recent recipient of a Nobel prize for his work in that field. Already in those days, around 2013/2014, it was clear that what were later to be called “AI” systems were radically different in function from human minds – not just the hardware but the nature of the tools, which learn patterns in a very different way from “intelligence.”
As in “Dancing with No One,” much of my graphics work, long predating “AI,” related to the expressive possibilities of flowing hair and cloth: my talk at NVIDIA’s “nVISION” conference of 2007, titled “Beautiful Women from the Future are Coming to Your Home” described techniques for hair and cloth simulations in video games, featured work from multiple 3D artists including myself, and was presented to an overflowing hall. Similarly, the hair of Aki Ross from the film “Final Fantasy” was in fact based on the hair of one of my patient and carefully-studied friends.
In the past, I’ve usually had to create most of my tools, whether for 3D, 2D, and starting around 2014-2015 including ones for diffusion and generative techniques. No one was talking about “AI art” in 2015 when I created the illustration “String Theory” below, using what were later called “AI” but which were for me a collection of data matrices manipulated by my homemade tools on my own large PC. This was posted to Instagram at that time ( https://www.instagram.com/p/_VkZsgy7l8/ ). It shows a hint of where I was proceeding by combining hand-made tools with hand-crafted image databases to explore my broader interest in figures, flowing motion, and the possibilities of my long-time favored medium: pixels.
In all cases I have used these tools as tools, not as some sort of robotic collaborators. These tools have no agency, no sense of artistic intent. They execute what is asked when applied by an artist, like a belt sander.
As for a broader interest in “flow” to show a continuity in this work, compare to a photograph https://www.instagram.com/p/42-80KS7nm/ from my year-long photographic-blur project “Liquidity,” also made in 2015, and building on methods developed over the previous decade.
These themes of flying hair, cloth, ribbons, leaves, ash, and sparks have persisted across many tools, though the costs of database training started to exceed my budget. “Ghosts of Democracy, Hong Kong” ( https://www.botzilla.com/photorant/2023/04/02/Ghosts.html ) was made in 2021 using a home-brewed CLIP-based diffusion tool: the same technology underlying later commercial tools like Midjourney or Adobe Firefly. It was one of the last made with such handmade tools before I turned to commercial databases due to hugely increasing costs.
The fourth image was to posted to Instagram while creating “Dancing with No One” ( https://www.instagram.com/p/CsO6z0lJ3Pv/ ), then called “The Ribbons Project” – I hope the continuity and progression of these works is fairly clear.
My process for those specific pictures has been to use LLM prompts developed as long chains of images, often with minute changes one from the next. This is not unlike the process of photographing many views of a motif and selecting from a photographic proof sheet (of which I also have thousands). One might compare my process to the long-term management of sourdough bread: each a refinement of the previous. At times I have provided the tools with a simple set of hand-drawn shapes as a basis image, to steer large compositional forms. I don’t recall if these specific pictures used that approach: the process was definitely iterative with variations of variations until the desired images were achieved.
Unlike (as I understand) most users of tools like Midjourney (these were made using an early edition), I constrain myself to fairly precise descriptions of the formal characteristics of the image and the items to be seen in it, and never using vague “in the style of…” or “looking like…” verbiage. I’m interested in new images from my own ideas, not emulation.
In the image shown here, the deliberate and iterative prompt for this and various rejected variation was “cascades of wide black silk ribbons::5 whirling straight up into the sky on the wind, wrapping around a hidden dancing figure:10; dense fog; burning embers::15 flying ash, light background; zorn palette” and for printing there was subsequent scaling from the original small image to 4K and regional manipulation of the color and contrast to lead the viewers’ eye where I wanted. In some of the “Dancing with No One” images, a few extraneous details were removed via Photoshop.
The title of this collection, “Dancing with No One,” is an acknowledgement that these are not photographs – there is no one, only a suggestion of figures created by the illusions of motion and wind. The work is also of regional interest, here in Sonoma County, California: it comes on the heels of the repeated devastating forest fires that hit our homes in the years immediately preceding the time the pictures were made. Local audiences have been quick to catch this reference: one gallery visitor told me that she classifies the work as part of a genre I’d been working in without naming: “climate sorrow.” My own mental classification of the work was a parallel to the “Ghost Dance” movement among Native American groups in the late 1800’s as their people and cultures were being erased. Defiant and sad, celebrating what is and has been in the face of an incomprehensible future.
(4) Thanks for the clarification regarding previously-published works. All images in “Dancing with No One” were created by myself, Kevin Allen Bjorke. In three cases these images were shared online but I have not relinquished any copyright claim for them. All others have never been published, only offered as limited-edition art prints and in this book.
I hope this reply satisfies your query.
Thank you for your time and your careful review of my application!
Kevin Allen Bjorke