Uninhibited Synthesis: Karan4d’s Very Real Invocations

An Exploration by Julian Brangold

M○C△
6 min readSep 22, 2022

“We believe we invent symbols. The truth is they invent us; we are their creatures, shaped by their hard, defining edges.”

-Gene Wolfe

examination” (2022) by Karan4d

Karan4d’s work inhabits a network of dream-fantasy conjuring, a hound-node envoy roaming in between visual disciplines. Advanced attractor becomes organized, pixel-by-pixel, in an obsessive cosmic-mechanic ritual.

A surface forms, one of archetypical sensibility, a surface, more specifically, of painting, for if a painter is one who creates paintings, then Karan4d is, at length, a painter.

Within the discipline of art created with code is an idea that the artist is a form of curator, an administer of applied sensibility that toy-crane plucks out the relevant from a billow of computational production, something which can, potentially, be infinite. In the ever-expanding field of AI prompt-engineering, Karan4d represents the now rogue subset of AI trainers: they who teach the machine to see by hand-selecting the dataset from which it learns, before asking it to spew out its unlimited experiments, thus creating a sandbox for the computer to play within before visualizing. This may be why the resulting creations look so different, and escape the now all-too-familiar frayed fabric AI aesthetic.

Karan4d is an techno-artisan, his hand petting all instances of the process of AI commandeering. He is most definitely not “collaborating” with the machine, but utilizing the labyrinthine tool in a most proficient way, which demonstrates his deep commitment to the cyber-coalition. If we venture further down into oblivion, we might even say he has a cyber-fascination. The process of data-setting (the manipulation of the data-set that the AI trains on to then create imagery) in his practice is most deliberate, as he handpicks, from a foundation of clear aesthetic intention, the sets of images the machine will learn from before it creates. Recuperated piece by custom-written piece, he handcrafts the code which the model interprets. He runs tests, trials, exercises until the image he is looking for emerges from the screen.

“theoxeny” (2022)

His most recent works evoke an artistry forged within tradition, while debasing a most enigmatic foundation to the point of uncanniness. The work theoxeny (above) is accompanied by a poem:

Saffron-kindled threads of wind

blowing from the west

they made my name become familiar,

brushing its breath lighter, fresh and wholesome.

My birth-festival the king’s

gracefully accompanied

by soma drinks

for exhilaration, O lord.

In the center of the frame, a two-headed figure rests with their eyes closed solemnly. Both faces are placid, calm, restful. They respite untroubled by what’s in front of them; they have accepted us, the intricate viewers standing in front of them, and deemed us nonthreatening. An almost libidinous texture of oil painting and brush, you can smell the turpentine. Colors frolic in an eternally familiar harmony: pink column, brown-orange background, some blue fondling the shadows, all warm, all congenial. We even detect the impression of a canvas background. It is a painting in its composition, its light portrayal, in its sensible depiction of expression, in its dialogue with an incipient surrealism from the turn of the 19th-century. It is a painting that speaks to the history of artistic development, but hides a secret, anthological line slitting a path straight from industrial revolution through to AI uprising.

“dyophrenesis” (2022), by Karan4d

Karan’s isomorphic portraiture work invokes a form of frail, lull darkness. His subjects never existed, and knowing this gives the experience of viewing his pieces a patina of uneasiness. The same thing happens often with machine explorations.

We (sometimes unjustly) equivocate digitality with frigid coldness, though that’s a feeling of our own devising. But machines are utensils.In Karan4d’s case, they are the craftperson’s implements with which he shapes the clay of machinic paraphernalia, outputting an earnest territory of emotive conveyance, a corner of sensibility, and, above all, meaning.

In his generative series for fxhash (the generative art marketplace built on Tezos) “Diffusion Walks: Series 1” Karan creates a random walk program that reveals an AI-generated image, the line dancing and quivering as if birthing the color areas that conform the composition, as if a hand was scribbling down the image fed to the algorithm. As a time-based piece, the duration and speed of the process is highly relevant. In a dialogue with randomness, how long the piece takes to complete itself varies from iteration to iteration. Some are completed almost instantaneously. Others have a very slow speed and end up unfinished, while some have a speed of 0, a process that results in a blank canvas. All of these outcomes belong to one single artistic compendium that speaks to the possibilities of an artist’s process. Some drawings simply emerge and would widely be considered works in-and-of themselves, some are never finished, some don’t begin at all. This proceeding elicits a figurative formulation from an abstract one. It feels as though the code is limning the gradual operation of manufacturing the image in real time, evidencing the computerness of the artificial depiction, and, at the time, emulating the very human process of hand-drawing. It is no minor gesture to combine two worlds of computing that normally are used separately, or at least, classified dissimilarly within market structures: the allegorical AI generation and the more abstract, code-based procedural invocations. They are both generative processes but conform to different computer-art traditions, and so tend to correspond to different aspects of the market. Karan boldly combines the two routines to create a work that makes it feel like the machine is talking to itself.

“Pure Liquidation” (2022), by Karan4d

It would be unbecoming were we to talk about an AI-generative artist without taking into account the vertiginous, recent mainstreamification of tools for AI image-generation. Dall-e, Midjourney, Stable Diffusion, they all often inspire thoughts in the ballpark of, “Artists are dead. This is the future of aesthetics.”

Maybe so, but true, clinical instrument artificers like Karan, the ones who are devoted (or obsessed) enough to submerge themselves in the procedural viscera of machinic production, the ones who thirst to understand the inner workings of the contrivance in front of them, they will continue challenging the run-of-the-mill zeitgeist, rejecting the untouched, default monster.

Such an undermining of the given parameters and groundings these tools provide out-of-the-box, and with them the all-too-similar-looking results, prepares the experiential terrain for questioning what the big entities behind these systems want to establish as a normative aesthetic value. Artists like Karan, who have the power of a heightened awareness, and thus the means for a more experimental use of the tool, are able to deform expectations. The output becomes work that enthralls, galloping to arouse a relevant experience, one that, after the dust clears, leaves us in the uncomfortable place of affixing our own, very personal meaning to the machine’s results, and extracting feeling from the callous, quiet grain we’ve displaced.

For more about Karan4d, read our three-part interview with the artist .

--

--

M○C△
M○C△

Written by M○C△

At the crux of a digitally-native creative space, The Museum of Crypto Art (M○C△) preserves history, elevates artistry, and ignites decentralized culture.

No responses yet