Matt DesLauriers

Artist

Matt DesLauriers · official site ↗

Based

Canadian artist based in the United Kingdom

Practice

Generative systems, creative coding, plotter drawing, print, interactive installation, open-source software, and research

Raster

matt-deslauriers ↗

In the set

Meridian · 1,000
Subscapes · 650

AB[500]

2 projects · 1,650 editions

Links

Full bibliography ↗

Matt DesLauriers writes software to discover images he could not draw in advance. A Canadian artist based in the United Kingdom, he works across code, prints, installations, websites, and physical drawing systems. He is also an active open-source contributor, building tools through which other artists can develop and export generative work.¹

He had coded since adolescence, but around 2014 began joining that knowledge to an artistic practice. An early JavaScript sketch transformed a photograph into an autonomous impressionistic painting. The browser became a studio in which a brief program could be run, adjusted, and run again. DesLauriers later built workshops and public libraries around that same iterative method.²

The landscape is not the subject placed inside the algorithm. It is the evidence left by the algorithm's behaviour.

Subscapes: surveying a multiverse

Subscapes, released through Art Blocks Curated on 23 April 2021, comprises 650 isometric terrains. DesLauriers wanted his first Art Blocks project to feel natural and familiar while remaining clearly artificial. Each output occupies its own ecosystem of colour, topology, structure, and rendering style. Together they form a multiverse of possible worlds rather than views of one continuous planet.³

The isometric view recalls strategy games and simulated construction. DesLauriers grew up with games such as SimCity, where a world is presented as a system that can be inspected and changed. In Subscapes, however, the viewer has arrived after the process. The terrain sits like a collected specimen or a survey model from an inaccessible dimension.

DesLauriers has described the project as an artificial survey. A conventional surveyor converts measurements of an existing place into a diagram. His code reverses the relation: equations and parameters produce the terrain, and the image records a place that exists only within that generative space. Nature supplies not a picture to copy but examples of how local rules can create coherent large-scale form.⁴

The project's independence is part of its meaning. Its code has no external dependencies and can run in a command-line terminal. DesLauriers imagined it moving beyond the original browser into mechanical plotters, emulations, and remixes. The blockchain stores not only a receipt for an image but decentralised drawing software capable of being performed again.⁵

Meridian: coordinates without geography

Meridian followed through Playground on 28 September 2021. Its self-contained JavaScript program is only 15 kilobytes, yet each output constructs a stratified landform from hundreds of thousands of small coloured strokes. The token hash becomes a coordinate within a multidimensional parameter space, locating one composition along one of many possible longitudes.⁶

The title refers to longitude, but there is no corresponding Earth. Cryptographic seeds replace physical coordinates. A holder's token does not point to a mountain waiting elsewhere; it supplies the position from which the algorithm resolves one mountain from its field of possibilities.

DesLauriers spent months developing the single program, a marked change from the short sketches that had previously circulated briefly on social media. Long-form release meant all 1,000 results would be collected, compared, and scrutinised. The system therefore needed to sustain both individual composition and population-scale variety.⁷

Four evenly spaced works show the flagship algorithm's range.

Marks that behave like matter

Meridian gains its tactile character from accumulation. Thousands of short strokes produce strata, atmosphere, and depth. DesLauriers compares the system to pen plotting, except that a physical pen would have to change colour with nearly every mark. Many outputs use only two or three ink-like colours, but their interleaving creates tonal richness beyond that limited palette.⁸

This makes print more than merchandise attached to a digital work. At large scale, the marks become individually visible. The 272-page Meridian book reproduces all 1,000 outputs with code examples, technical essays, trait analysis, and large details. It also records how to recover the project from the chain if Art Blocks' own interface disappears.⁹

The work's preservation model is distributed. DesLauriers imagines the code suspended by invisible threads among network nodes. Old web sketches can vanish when hosting lapses or platforms close; the on-chain projects persist because many participants maintain the shared system. Permanence remains contingent, but responsibility is no longer attached to the artist's server alone.

Tools as public work

DesLauriers' contribution to generative art extends beneath his own images. His canvas-sketch framework gives artists a common JavaScript workflow for rapid sketches, browser preview, high-resolution export, animation, print dimensions, and shareable builds. Companion utilities provide seeded randomness, geometry, colour, shaders, and pen-plotter output.¹⁰

That infrastructure now extends into formal research. In the 2026 ACM pictorial Synthetic Gestures: An Evolutionary Sketching Machine, DesLauriers presents a gradient-free system that turns text prompts into plotter drawings. Drawing gestures are represented as implicit neural forms and optimised through an evolution strategy, with semantic guidance from CLIP. Because the browser-based system runs locally and gives the artist real-time feedback, it supports noise, discontinuities, discrete operations, and multicolour marks that are difficult to accommodate in end-to-end differentiable renderers. Machine learning appears here as an artist-built drawing instrument, not a finished-image endpoint.¹³

A collector's statement by SonOfLasG gives that research a view from inside one work. Writing about the Bicycle drawing from Latent Dispatch: Studies, he reads its continuous line as a pre-entanglement artefact: the system's distilled idea of a bicycle before audience drawings, noise, disconnected strokes, and multicolour operations complicated its certainty. His own history with cycling sharpens the distinction between a computationally legible archetype and the particular bicycles through which a life is actually lived. The study records the assumptions of one model at one moment, not a universal form.¹⁴

Open-source infrastructure rarely appears in a museum display, yet it shapes what a field can make. A tool encodes assumptions about scale, export, reproducibility, and material translation. By making those decisions reusable, DesLauriers turns private workflow into a public resource. The framework supports other artists while continually feeding back into his own practice.

This reciprocal model also appears in his teaching and historical work. Workshops introduce creative coding to new practitioners; the ongoing documentary project Generative reconstructs lineages that the rapid rise of blockchain art risked flattening. Producing tools and producing history are related forms of maintenance.

Worlds, games, and natural signals

Wayfinder, made with the National Film Board of Canada, turns procedural landscape into an ephemeral journey. Players cross forest, grassland, and tundra, collecting words that combine into poetry while vegetation and animals respond. Visual assets, verse, and terrain are assembled in real time, so each traversal becomes a temporary work that will not recur exactly.¹¹

Later projects continue to move between natural model and artificial trace. The Sferic Project records low-frequency atmospheric radio far from electrical interference. Sierra began with a generative mountain whose chance arrangement suggested alpenglow, then translated the digital palette into fifteen layers of silkscreen ink and four CMYK-linked tokens.¹²

DesLauriers is careful not to claim that code surpasses nature. His interest is in simulating a system rather than simulating an image of nature. The distinction keeps the work from becoming landscape illustration. What matters is how parameters interact, where order breaks, and when an artificial process produces a form that feels inevitable.

Why he belongs in the AB[500]

Matt DesLauriers belongs in the AB[500] because Subscapes and Meridian demonstrate two complementary models of the generative landscape. One presents discrete isometric worlds as surveyed objects. The other maps hashes into a continuous high-dimensional terrain and renders each coordinate through accumulated marks.

Both projects connect code to material practice. Their systems can become terminal drawings, pen plots, archival prints, large displays, or a complete book. They are digital-first without treating the screen as the only legitimate surface.

The canon also needs the infrastructure around the works. DesLauriers did not only use a creative-coding ecosystem; he helped build it. His art and tools share a principle: a compact system should open a large field of possible action. The mountains are the visible evidence. The deeper work is the space that makes them possible.

Bibliography

The references cited here are listed below. For the complete 24-source ledger, see the Matt DesLauriers bibliography ↗.