Blood on the Sheets: You’ll Know It When You See It…
Spinning up the AI image tooling PR machine
I’m weak with visual arts. While I’m never short of ideas or ways to express them, being concise with the written word has its pitfalls as well. I never liked drawing all that much but I’m working on a project that requires me to express concepts through imagery.
I’m doing it.
The exercise is good and I think I can improve but that’s not as important to me as working through the ideas for further development in other formats.
It makes me feel like the earliest days of when I was learning how to play music, fumbling around stringed instruments finding little melodies and connecting dots about the relationships of intervals.
Currently, I don't have any visual scales or arpeggios to draw but maybe I'll make a few in time.
Cracking Our Knuckles
However, I have been drawing pictures based on sketches for a storyboard set I’ve been working on over the last few weeks. It’s slow going, no rush.
The first iterations were pencil, which I have been scaling up by redrawing with pen. Going from a quarter of an 8.5”x11” sheet of paper to a single piece per drawing while also giving up the forgiveness of pencil for the permanence of ink is new to me. I’ve made some errors. Not a big deal, there’s more paper.
After mimicking the general scene I began to refine the detailing to improve the life of the image, but this is when the process really transitions because I’m not fully relying on a drawing utensil to pull it off.
Let me ask you: if I take the line drawing & push it through a diffuser with an editorial prompt to add additional texturing to the image, is it cheating the craft of illustrating?
Is that my creation if it’s infused with alterations via algorithms?
Really not concerned with the answers, I’m doing it.
This is where I see people raising a concern. Honestly, I’m over the purity test. The same crying occurred when Photoshop and Illustrator came on the scene. “It’s not real art!” Or “That’s cheating!” Ok. Boring…
An event near where I’m from recently announced they would ban all AI art from their floor and remove artists selling the work. Ultra lame and shortsighted, as well as hypocritical. I’m all for the Luddites having an event if they want, but they will be displaced by these new tools in the same way that Disney moved on from traditional animators when Pixar lit up theaters and moved increasing ticket sales.
Don’t like it? Don’t use it, your “chastity” will remain intact. But the efforts to ban the work or litigate against the software is going to undermine artists more than it will result in increased sales for existing work. Worse yet, the technology will be sealed up behind corporate moats where only multinationals have access to it and they will step on underdog creatives by lobbying for “protections”.
Hands Over the Marble
Let’s address some of what is required to even get started creating images with diffusion software.
You need a computer with a decent processor, I run an MacBook Pro with an M1 plus 16GB of RAM (not an overpowered machine) and it still takes a while to spit out the images when the parameters are maxed or the prompt is complex. The software and the model are also like 8GB of data as well. If your PC is good, check out Diffusion Bee for a compiled piece of software with GUI otherwise there are versions of Stable Diffusion you can run if you’re comfortable with command line interface (CLI). I’m still learning but I like being able to adjust the specifications like dials and switches.
In terms of CGI versus AI, that gets a little hairier. I have a nuanced perspective on AI, which I’ll address after defining CGI: computer generated graphics, not analog (physical world manifested or existence, classically studied) derived. I really see no difference between using Photoshop, Illustrator, or a piece of software like Diffusion Bee. Call it collage or call it iteration, it’s good.
Chisel to the Stone
Primarily, when I use the term “AI” it means “augmented intelligence”, which to me dials into the aspect of tool use—a number of creatures and perhaps organisms do this as well, in various ways of utilizing abstraction to solve problems. For example, a chameleon can’t use a hammer but it can shift its colors for environmental security, which I find to be quite abstract! So in this sense, tools like CGI or image diffusers become another creation tool that we are using to augment our intelligence and manufacture the realities we’d like to engage with visually.
Now in terms of AI: there is no intelligence which is artificial - it’s only a matter of scale. Can this intelligence accomplish what that one can? I can’t do photosynthesis like plant life but they can’t write about it here on this platform…yet!
People are using ChatGPT to write scripts, it’s definitely interesting in terms of rapidly developing a first draft but I think there’s too much value in working through the process to just give that over to the software. I think we’ll lose out on generating neuronal connections regarding the early steps of creating and familiarity with the story.
We are becoming prompt engineers, which is why I choose to run the software on my machine instead of via the web. The best artists I’m seeing guard their prompt techniques like recipes.
To wrap up my points: let’s not think this is in its final form today. It’s just practice. Let the folks make their pictures and if it’s no good then the market will decide to buy or not. What comes next will be more interesting and compelling, I believe—but we have to learn the scales and chords first before writing dynamic songs.
Hope some of this helps…