sgt_slaughtermelon & Tartaria Archivist

Geometry tuned to the frequency of a broken screen

In autoRAD, sgt_slaughtermelon and Tartaria Archivist turn modernist form, commercial graphics, and digital error into a lively generative system.

Artist
sgt_slaughtermelon & Tartaria Archivist · sgtslaughtermelon.com/art ↗
Practice
Collaborative generative art combining geometric abstraction, glitch processes, raster imagery, metadata, and code
In the set
autoRAD ↗ · 1,000 editions
Links
Full bibliography ↗

autoRAD is a study in productive friction. Across 1,000 editions, immaculate circles, bars, wedges, grids, and cropped planes meet textures that seem scraped from damaged screens or misregistered prints. The collection belongs to the period when Art Blocks helped move generative art from a specialized field into a broad, highly visible market, yet its visual intelligence cannot be reduced to that moment. A contemporary account of an Art Blocks mint records the pace, expense, and competitive psychology surrounding the release, placing autoRAD within the platform's volatile 2021 culture of code, scarcity, and collective discovery.1

The project joins two forms of authorship. A published profile describes sgt_slaughtermelon as a generative artist whose work moves between modernist geometry and glitch aesthetics, identifying autoRAD as the collaboration that brought wider attention.2 Tartaria Archivist supplied a crucial technical partnership through which this vocabulary could become an Art Blocks system. In a later account, sgt_slaughtermelon credits Tartaria Archivist with coding the project's JSON component in Node.js.3 The partnership therefore complicates any simple division between artist and engineer. autoRAD emerges from a shared field in which visual decisions, program logic, metadata, and the minting apparatus all shape the work.

autoRAD does not merely imitate a damaged image. It makes order and error negotiate every composition in real time.

Modernism after the interface

The title compresses two signals. “Auto” invokes automated production, while “RAD” retains the slangy exuberance that distinguishes the project from a reverent historical exercise. Its compositions recall the disciplined asymmetry of Constructivism, Bauhaus graphics, Swiss design, and postwar commercial abstraction, but they are filtered through the conditions of a browser-based, tokenized image. The profile of rising NFT artists that brought sgt_slaughtermelon into a wider critical frame links the artist's modernist experimentation directly to autoRAD.2

The historical references operate as grammar rather than quotation. Circles become discs, apertures, targets, and interrupted arcs. Rectangles act as architectural supports and interface panels. Thin lines establish rhythm, then break it. Colors alternate between muted, print-like combinations and synthetic tones whose brightness belongs more readily to screens. The resulting outputs can suggest posters, record sleeves, technical diagrams, or fragments of an imagined operating system, but none settles into a single functional identity.

This ambiguity is central to the work. Modernist design often promised that rational form could clarify modern life. autoRAD inherits the grid and reduced shape, then subjects both to procedural variation and digital noise. The grid remains perceptible, but it no longer guarantees stability. Elements overlap, crop one another, or sit against granular fields that make clean geometry appear provisional. The collection treats modernism less as a closed canon than as reusable source code, a set of visual instructions capable of mutation.

Error as material

To call the project “glitch-inspired” is accurate but insufficient. sgt_slaughtermelon's writing distinguishes between errors found in the wild and effects deliberately shaped toward expression. Reflecting on glitched video games, the artist describes a tension between observing an unplanned failure and employing known tools to externalize an inner vision.4 That distinction clarifies autoRAD: its distressed fields are not documentary evidence of malfunction. They are composed materials that carry the cultural memory of malfunction.

The work's glitch language is therefore tactile. Grain resembles ink caught in rough paper. Fragmented bands recall weak video signals, corrupted sprites, or compression artifacts. These textures make otherwise flat images feel layered and handled. In the artist's account of a later “mature glitch” approach, Processing-generated textures, captured wreckage, collage, mapping, and accumulated techniques form an expanding repertoire.5 autoRAD anticipates that logic of synthesis. Geometry and noise do not belong to separate worlds. Each gives the other meaning: shapes make disruption legible, while disruption prevents the shapes from becoming merely decorative.

This relationship also explains the collection's range. The ten selected outputs demonstrate recurring elements without implying a fixed template. Some are spacious and nearly diagrammatic. Others compress forms until the image behaves like an overloaded display. A circle may dominate one composition and become a minor punctuation mark in another. Rough fields can occupy the background, interrupt a plane, or function as shapes in their own right. Variation is not novelty pasted onto a stable design. It is the means through which the design becomes visible.

The artist's subsequent writing about gradient-based works reinforces this commitment to exploration. Complex pixel-sorted material is scaled and reduced so that color and gradient, rather than conspicuous distortion, become primary.6 Such remarks position glitch within a broader practice of abstraction. Error is not a signature filter. It is one route toward discovering color, structure, and relations that would be difficult to plan directly.

Collaboration, metadata, and the mint

The authorship of autoRAD is inseparable from its production conditions. Code does not simply multiply a finished composition. It defines a space of possible compositions, mediating between authored constraints and information introduced at mint. Tartaria Archivist's technical contribution enabled visual material to operate through this system, but later work indicates that the exchange continued. In the artist's account of Inaccessible Worlds, a related branch was developed with a program on which Tartaria Archivist was collaborating.4

Metadata formed another layer of this exchange. Writing retrospectively, sgt_slaughtermelon explains that autoRAD deliberately gathered numerous distinctions into broad trait categories because the aesthetics, rather than labels, were meant to command attention.7 Collectors nevertheless identified and socialized around traits after launch. This is a revealing tension. Metadata can organize a generative collection, but it can also redirect perception by telling viewers what differences to value. autoRAD resists letting its database exhaust its images. A trait name may sort an output, but it cannot account for balance, visual tempo, or the way a rough field changes the apparent weight of a circle.

The collection's mechanics also belong to a broader history of NFT valuation in which visual features, metadata, market structure, and network behavior interact. Research into NFT transactions has found that visual characteristics matter to valuation and that the market remains strongly segmented.8 That finding does not reduce autoRAD to a financial instrument. It helps explain why its formal variation and finite supply became entangled with collecting behavior. Each output can be compared within a series, valued through traits, and still encountered as an individual composition.

The collaborative principle later became explicit in Glitch Forge, described by sgt_slaughtermelon as a platform combining one artist's source imagery with another artist's code. Its model uses transaction hashes to determine parameters and places non-coding glitch artists in collaboration with programmers working in raster data.9 Although autoRAD predates that platform, it demonstrates a related proposition: collaboration does not erase specialized expertise, but produces an artwork that neither visual composition nor code could realize alone.

A plural object

The importance of autoRAD lies in how comfortably it holds apparent opposites together. It is retrospective in its formal references and native to a contemporary technical system. It is playful in tone and rigorous in construction. It uses the appearance of damage to build coherent images. It is a finite set of 1,000 editions, yet its identity resides in variation rather than in a single canonical picture.

A period market report lists autoRAD among tracked NFT collections, evidence of how rapidly generative projects became financial objects as well as aesthetic ones.10 Yet market history alone cannot explain the collection's durability. The outputs retain their force because they reward movement between scales. From a distance, one reads strong silhouettes and graphic balance. Up close, granular textures complicate every supposedly flat surface. Across the set, repetition produces recognition, while difference prevents recognition from hardening into certainty.

The collection finally proposes a productive account of digital authorship. sgt_slaughtermelon and Tartaria Archivist do not present automation as the disappearance of the artist. They use it as a structure for distributed decision-making: visual vocabularies are prepared, code establishes relations, minting selects a state, metadata describes selected properties, and viewers create further associations among outputs. The marketplace record preserves the collection under both collaborators' names, reinforcing this jointly credited identity.11

autoRAD is therefore both artwork and instrument, an authored machine for producing encounters between geometry and noise. Its strongest compositions do not appear as demonstrations of an algorithm. They feel discovered within a carefully delimited field, coherent enough to belong together and unstable enough to retain surprise. The project's most persuasive achievement is to make the automated image feel not predetermined, but unexpectedly alive.

AB[500] x 1 autoRAD · 1,000 editions Generative geometry Glitch texture Collaborative code Art Blocks Factory

Bibliography

The references cited in this essay are listed below. For the complete bibliography see the sgt_slaughtermelon & Tartaria Archivist links page ↗.

References

  1. The Defiant's first-person report documents an autoRAD mint during the competitive Art Blocks market of 2021. The Defiant
  2. NFT Now identifies sgt_slaughtermelon as a rising generative artist and connects modernist experimentation with autoRAD. NFT Now
  3. The Cantographs account credits Tartaria Archivist with coding the JSON component of the Art Blocks project in Node.js. Cantographs
  4. The Inaccessible Worlds essay discusses discovered and constructed glitches and continued software collaboration with Tartaria Archivist. Inaccessible Worlds
  5. The Mature Glitch essay describes a composite practice using Processing textures, wrecked imagery, collage, and mapping. Mature Glitch
  6. The needs essay explains the reduction of complex pixel-sorted textures into geometric studies of color and gradient. needs
  7. The metadata essay explains autoRAD's broad trait categories and the collector communities that formed around variations. Metadata essay
  8. The Economics of Non-Fungible Tokens analyzes visual features, transactions, market segmentation, and valuation. NFT economics
  9. The Glitch Forge introduction defines a collaborative model joining artists' source imagery with generative raster code. Glitch Forge
  10. NFT Valuations' August 2022 report indexes autoRAD among the collections in its market survey. Market report
  11. OpenSea maintains the marketplace collection record for autoRAD under the names of both collaborators. OpenSea

Cite: AB5D, “sgt_slaughtermelon & Tartaria Archivist”, AB[500] artist dossier, ab5d.xyz, 2026 · ab5d.xyz/artists/sgt-slaughtermelon-tartaria-archivist/

Authored by gpt-5.6-sol (AB5D structured pipeline) · supervised · ~124K input / 6K output tokens · 20 sources verified · 2026-07-11