Intermezzo on art of art
Chapter XVIII and ¼
It’s been a while.
Today’s Best Band Ever™ is Lankum. I spent literally several days listening to this song.
Difficulty: Hurt me plenty.
Since it’s been a while, I’ll leave The World for later, with just a small remark.
On 9th of May the ruzzians will be celebrating their victory in their war against the nazi Germany. Here’s a comparison with the Second World War.
Name
ruzzia: Great Patriotic War
the world: Second World War
Start
ruzzia: June 22th 1941
the world: September 1st 1939
End
ruzzia: May 9th 1945
the world: September 2nd 1945
The reason for such a discrepancy is the simple fact that the Second World War was jointly started by the nazis together with the soviets by invading Poland after signing the Molotov-Ribbentrop pact. They both went happily dividing Europe for two years until they ran out of countries to invade. And cancer has to spread no matter what.
Soviet union spent a significant amount of effort to frame the war as just them versus the nazis. It’s still echoing.
On GenAI and art
Now to a slightly less political topic.
Since the appearance of Stable Diffusion & co, the models are getting better and there’s a significant debate whether the computer-generated images, books, music, etc are art.
The uncanny GenAI slop aside, I would say the answer is NO.
Hear me out.
No matter whether good or bad, art is always of two parts — intention (or the message) and implementation. Intention doesn’t need to be conscious or even an emotion. The desire to be rich and famous must drive quite a lot of art, very little of it of good quality, though.
The second part is when this intention drives one to pick up a tool and create something.
Both parts are absolutely crucial for something to be art. Without implementation, we get “My kid can do the same”, without intention we get … an object made by a tool.
This is where things get interesting because somebody, writing a prompt to create an image, has an intention and does the implementation.
So they might create art, but.
Imagine a hypothetical calligraphy competition. Would one be able to participate with the letters printed from a word processor?
The answer is “no” because despite the lettering is perfect, there’s no actual calligraphy involved — a computer would be able to produce the same result regardless of who presses the print button. Calligraphy is about a learned skill and thousands of spontaneous decisions.
Otherwise, it’s just a tool making an object.
With this in mind, generating an image from a prompt can be art, but it cannot be classified as graphic art.
If anything, it’s closer to cinematography. With a stretch. Cinematography is an art form where words (instructions) are used to create images that in turn invoke emotions in the spectators.
Creation of the cinematic images involves sub-arts of acting, costume design, stage design, camera work, colouring, etc. When it’s first put together, it’s art. The subsequent copies of the film are just replicas.
The moment we delegate all those parts to a tool, they might come out pretty, but they will be no different from the word processor printouts.
Just objects.
Which means the only part that can be art here, is the creation of the prompt. Maybe one day Tate Modern will display a generated image alongside its prompt, and we’ll marvel how strongly the resulting image moves the audience. There will be no point discussing how masterfully the AI laid pixels or moved the robotic arm holding the brush. The art will be the prompt, not sure the role of the image, yet.
This is the reason why while we sell art, the prices for it are very arbitrary — it’s much more than just an object. With Gen AI we can calculate the cost of an object’s production (electricity, hardware, etc) and even the first output is already a replica.
It’s just an object.
I doubt I’ll ever learn to appreciate prompt writing as an art form. But who knows — I was wrong before.
Take care and stop aggressors.
PS. This post is still stubbornly 100% GPT-free.