The algorithm as brush: When code becomes a work of art
Imagine a morning in 1965, in a laboratory in Stuttgart bathed in a sickly light. Frieder Nake, mathematician and artist, watches a machine click and grind as a stylus traces geometric lines on a sheet of paper. These drawings, born from a series of instructions written in ALGOL, are not the product of chance: they obey precise mathematical rules, yet their final result partly eludes their creator. For the first time, a work of art emerges from a dialogue between man and machine. Fifty years later, a collector will pay 3.3 million dollars for a similar algorithm, Tyler Hobbs’ Fidenza, whose abstract curves come to life on high-tech screens. Between these two moments, generative art has traversed the ages, shifting from scientific curiosity to major cultural phenomenon.
By Artedusa
••9 min readWhat fascinates about this art born from code is its fundamental ambiguity. Is it the artist who creates, or the machine? Is the result a work of art or merely the product of algorithms? And above all, how do you collect something that, by definition, can be reproduced infinitely? To answer these questions, we must delve into the backstage of this artistic revolution, where equations replace paint tubes and where each work is both unique and potentially infinite.
01When machines dreamed of Kandinsky
The first steps of generative art were not taken in galleries, but in research laboratories, between metal cabinets filled with punch cards and cathode-ray screens with their greenish glow. In the 1960s, as the world discovered the Beatles and the Cold War, a few pioneers dared to imagine a new form of art, where creativity would no longer come solely from the artist’s hand, but from the mechanical execution of predefined rules.
Frieder Nake, the German mathematician who watched his plotter draw in 1965, openly drew inspiration from Paul Klee. His Hommages à Paul Klee were not copies, but algorithmic variations on the Swiss master’s motifs. By coding instructions like "draw a line of random length in a given direction," Nake created compositions that, while evoking Klee’s aesthetic, possessed a new geometric rigor. Alongside him, Georg Nees explored similar territory with Schotter (1968), a work where perfectly aligned squares gradually became disorganized, like stones tumbling in a landslide. Chance, controlled by mathematical formulas, became an artistic tool.
Yet these experiments remained confidential. At the time, the general public knew nothing of these machines capable of creation. Art critics, meanwhile, regarded these drawings with suspicion. "It’s cold, mechanical, soulless," some said. Others, like the philosopher Max Bense, saw in them a revolution: art was no longer a question of manual skill, but of conceptual intelligence. By writing rules, the artist became a choreographer, and the machine, his dancer.
02Vera Molnár and the mourning of the brush
If Frieder Nake was a mathematician playing at being an artist, Vera Molnár was an artist who had learned to speak to machines. Born in Hungary in 1924, she had first studied traditional painting before turning, in the 1960s, to computers. Her journey perfectly illustrates this transition: after decades spent wielding a brush, she found herself in front of a screen, typing lines of code in FORTRAN to generate infinite variations of squares and lines.
Her work (Des)Ordres (1974) is emblematic of this approach. On a 10x10 grid of squares, Molnár applied systematic deformation rules: some squares tilted, others shifted slightly, creating an impression of movement and imbalance. What fascinated in her work was this tension between control and letting go. Molnár defined the rules, but it was the machine that decided, to some extent, the final result.
One day, her plotter broke down. Rather than give up, she took a pencil and manually reproduced the drawings the machine should have traced. This gesture, both poetic and ironic, summed up the entire ambiguity of generative art: even when technology failed, the idea remained. The work was no longer in the result, but in the process.
03The era of algorithms that come to life
With the arrival of personal computers in the 1980s, then the internet in the 1990s, generative art gradually left the laboratories to invade screens around the world. Artists no longer needed cumbersome machines: a simple Macintosh sufficed. It was in this context that Processing was born, a programming language designed specifically for visual artists, created in 2001 by Casey Reas and Ben Fry.
Reas, now a professor at UCLA, is one of the architects of this democratization. His Process Compendium (2004–present) are living works, where organic forms evolve in real time, like organisms under a microscope. What strikes in his work is this impression of life: the lines seem to breathe, the colors blend like fluids. Reas draws inspiration from biology and chaos theory, but also from Sol LeWitt, whose Wall Drawings were instructions to be executed rather than finished works.
At the same time, artists like JODI (Joan Heemskerk and Dirk Paesmans) pushed generative art in more subversive directions. Their work wwwwwwwww.jodi.org (1995) was nothing more than an apparently blank web page, but whose source code revealed an abstract landscape made of ASCII characters. For JODI, art did not reside in the final image, but in the code itself, in this invisible layer that structures our digital experience. Their approach, both poetic and critical, foreshadowed current debates on algorithmic transparency.
04Tyler Hobbs and the mystery of Fidenza
If generative art has experienced a spectacular resurgence in recent years, it is largely thanks to the rise of NFTs and platforms like Art Blocks. Launched in 2020, the latter allows collectors to purchase algorithms that generate unique works at the time of purchase. Among the most famous artists on this platform, Tyler Hobbs stands out for his both rigorous and poetic approach.
His work Fidenza (2021) has become an icon of generative art. Each piece in this series is the result of a complex algorithm that combines precise geometric rules with an element of chance. The curves intertwine, the colors respond to each other, and each composition seems both ordered and organic. What fascinates about Fidenza is this impression of perpetual movement: even when still, the forms seem to dance.
Hobbs, a former software engineer, describes his process as a collaboration with the machine. "I define the rules, but it’s the algorithm that decides the final result," he explains. This approach recalls that of Vera Molnár, but with a major difference: while Molnár worked with slow, unpredictable machines, Hobbs uses ultra-fast computers and modern languages like p5.js. Yet the mystery remains the same: how can a set of lines of code give birth to something so evocative?
05Collecting the immaterial
Buying a work of generative art is a bit like acquiring a musical score rather than a recording. You own the instructions, but the work itself can take different forms with each execution. This particularity raises fascinating questions about the notions of ownership and authenticity.
Take the example of Dmitri Cherniak’s Ringers, another emblematic series from Art Blocks. Each piece in this collection represents a string wrapped around a set of nails, according to algorithmic rules. The result is both minimalist and hypnotic. But what exactly does the collector own? A digital file? A hash on the blockchain? Or simply the right to say "It’s mine"?
The answer is complex. In the world of NFTs, ownership is recorded on a blockchain, which guarantees authenticity and traceability. But the work itself can be displayed, copied, or even reinterpreted. Some artists, like Refik Anadol, go further by creating immersive installations where the work evolves in real time, fed by data streams. In this case, owning the work means owning an experience rather than an object.
For traditional collectors, accustomed to canvases and sculptures, this immateriality can be unsettling. Yet it also opens up unprecedented possibilities. A generative work can be displayed on a screen, projected onto a wall, or even integrated into a virtual environment. It can also evolve over time, like a painting that changes color with the weather.
06Code as heritage
While generative art fascinates, it also poses major challenges in terms of preservation. How do you preserve works created with obsolete technologies? What happens when the programming language used to generate them is no longer maintained? These questions are crucial for museums and institutions wishing to integrate these works into their collections.
The ZKM (Zentrum für Kunst und Medien) in Karlsruhe, Germany, is one of the few museums to have taken up this challenge. In its reserves, one can find works by Frieder Nake and Vera Molnár, as well as vintage computers and emulators that allow software from the 1960s to run. "Preserving generative art means preserving both the code and the machine that executes it," explains a curator at the museum. "It’s like having to preserve both the score and the piano."
For more recent works, like those on Art Blocks, the blockchain offers a partial solution. The code is recorded permanently, and the work can be regenerated at any time. But what happens if the platform disappears? Or if the servers hosting the images go down? These questions remain unanswered, and they highlight the fragility of this immaterial art.
07The future: between utopia and dystopia
Generative art is at a turning point. On one hand, advances in artificial intelligence open dizzying perspectives. Artists like Mario Klingemann use neural networks to create portraits that seem straight out of a dream. On the other, debates over authenticity and intellectual property rage. Can we really speak of art when the work is generated by AI? Who is the author: the artist who wrote the code, or the machine that executed it?
These questions are not new. In the 1960s, Frieder Nake already wondered whether his algorithms could be considered art. Today, with the rise of NFTs and generative AI, these debates take on a new dimension. Some see an unprecedented democratization of artistic creation. Others, a threat to traditional artists.
One thing is certain: generative art will not disappear. It will continue to evolve, driven by technological advances and the imaginations of artists. Perhaps one day, we will all own algorithms capable of creating unique works, just as we own cameras today. Until then, one thing is sure: the dialogue between man and machine has only just begun.