What does it mean to see a painting completely?
Not to simply look at the image it presents but to engage with the strata beneath it. That which includes the revisions and hesitations that led to its final form.
Viewfinder is what we mean when we talk about medium-native works. As a series, it’s one of the most intriguing ones I’ve seen on Bitcoin using the Ordinals protocol to date. With that comes a number of different aspects one can consider.
I could talk about the sheer amount of data we put onchain for it (15+ MB). Or the complex parent child tree. Or how the Belvedere used the largest X-ray machine in the EU to perform these scans.
Or as the venue for its release, we could simply slap some “conceptual” tags on it and consider our work done.
Most of these are certainly nice-to-haves, and if we were other groups on Bitcoin, we probably would lean into these aspects as the core selling point. But at the end of the day, none of these things really matter when it comes to Viewfinder.
Here, I talk about why Viewfinder actually matters:
For the uninitiated, Viewfinder is a collaboration between artist Marcel Schwittlick (@schwittlick_), Ancora (@Ancora_art), and the Belvedere Museum in Vienna (@belvederemuseum). The work seeks to answer that very question by investigating what lies beneath Gustav Klimt’s iconic surfaces.
Using a collection of multispectral scans including X-ray, UV, and infrared from the Belvedere’s Klimt archive, Viewfinder constructs a new kind of artwork. It doesn’t reproduce the paintings or reinterpret them in style. We’ve seen plenty of code-as-paint explorations over the years and we know we can dig deeper.
Instead, it takes pairs of image layers with two different perspectives of the same motif and asks what happens when we treat the space between those layers as a site of visual meaning. The result is a series of generative, animated works, each a response to structural difference rather than aesthetic reference.
This process is rooted in extrapolation. Not interpolation between two known points but projection outward into a new conceptual field. Marcel describes it as “an act of speculation”, one that embraces uncertainty and complexity rather than reducing Klimt’s work to a digital shorthand. It isn’t a simulation of Klimt and it isn’t interested in stylizing his compositions. Viewfinder treats the invisible residues of Klimt’s practice as data to be processed from the sketch beneath the surface, the corrections layered under gold leaf, and the structural decisions that never appear in the final painting. What emerges is less a tribute and more a continuation, a computational interpretation of the image’s internal structure.
For more insight into this idea of extrapolation, find Marcel’s article on Viewfinder here.
This approach stands in stark contrast to much of what’s often described as “innovative” in contemporary art-tech discourse. In a recent post, SHL0MS critiqued a gallery work described as “an oil painting created by a robot arm.” On closer inspection, it turned out to be a digital image printed on top of a randomly textured surface. It was convincing from a distance, but materially dishonest up close. The paint didn’t construct the image; it floated above it, applied by a plotter without any meaningful engagement with form. It’s a photocopy of painterly depth, without the image ever truly interacting with the substrate.
Viewfinder operates in the opposite direction. Where the robotic pseudo-painting uses technology to simulate a painterly effect, Viewfinder uses technology to enter into dialogue with the painting’s underlying logic. It works with evidence of structural data captured from archival imaging, reinterpreted through a generative system. It generates new work based on the latent relationships between what Klimt composed and obscured. This distinction of the difference between simulating a result and extrapolating from process is at the heart of the project.
The visual language of Viewfinder is unapologetically digital. Its palette veers toward saturated flats rendered in compressed, aliased textures that feel more algorithmic than organic. There’s a sense of artifacting baked into the system with colors clashing without blending, detail collapsing into pixelation, and form emerging through computational contrast. This isn’t digital pretending to be analog, nor is it a nostalgic filter. It’s digitality asserting itself on its own terms in a way that is brutal and synthetic. In a visual ecosystem saturated with soft generative gradients and AI hallucinations, Viewfinder resists atmospheric aesthetics and foregrounds its own mechanics, encoding violence into the image via the ruptures of its process.
That brutalism isn’t incidental. It’s in dialogue with the work’s source material that extends beyond simply being stylized Klimt reproductions. They’re speculative constructions built from scan residues. The rawness of the output mirrors the conceptual tension in the project. Just as Marcel’s code interrogates the structural integrity of Klimt’s canvases, the visual system refuses to smooth over the gaps.
You’re looking through a Klimt painting, rendered in digital noise and procedural motion. In this way, Viewfinder echoes the approach of early net art practitioners, who used glitch, compression, and breakdown as primary materials. It’s a visual aesthetic that treats resolution not as clarity, but as a threshold where systems either reveal or conceal their internal logic. Viewfinder chooses to reveal.
It also places Viewfinder in a lineage of systems-based art practices in place of stylistic homage. Artists like On Kawara and Hanne Darboven all treated rules, repetition, and duration as valid materials for making. Generative art pioneers such as Frieder Nake and Manfred Mohr wrote algorithms not to imitate form but to discover new ones. Like these precedents, Viewfinder displaces the image from the center of the frame and turns our attention instead to the process that governs its emergence. The work begins with a rule rather than a picture.
That logic extends all the way down to the way Viewfinder is published. These works inscribed as digital artifacts on Bitcoin is a reflection of the structure of the work itself. These are pieces built on timestamped inputs, algorithmic systems, and trace data. Publishing them on a protocol known for permanence, sequence, and procedural transparency simply reinforces the conceptual underpinnings of the project.
As my colleague Steven Reiss noted during a public presentation at the Belvedere, “If art consists of rules, data, and time, it is logical to choose a medium that brings rules, data, and time as inherent substance.” In this context, Bitcoin is an infrastructure that mirrors the work’s own foundational logic.
Ultimately, Viewfinder is not about seeing more clearly. It’s about understanding that clarity is a layered thing that is constructed and negotiated. Rather than showing us Klimt, it shows us how a Klimt comes into being and how sketches are buried and surfaces layered until something coherent takes shape. It reveals that an artwork is a temporal sequence of adjustments and revisions rather than a fixed object.
Viewfinder offers a kind of attention. It invites viewers to consider the image not as a final effect, but as a condition that arises from underlying rules. Instead of presenting us with a portrait or landscape to see, this lens focuses on how seeing is made possible in the first place.