Perry's Pen

William S. Burroughs, the counter-culture literary icon and author of Naked Lunch, often seems a figure of the past. Yet, his radical artistic methods—specifically the cut-up technique and his shotgun paintings—reveal a surprising, almost prophetic, engagement with concepts that now define the era of Artificial Intelligence (AI) and Large Language Models (LLMs).

Burroughs’s work, conducted decades before the first publicly available consumer AI, can be viewed as an attempt to find an algorithm for creativity and to remove the human element from artistic production, mirroring core debates surrounding today’s generative AI.

1. I. The Cut-Up Method: A Primitive LLM
2. II. David Bowie’s Verbasizer: The First LLM on a Laptop
3. III. Generative Chaos: Removing the Human Hand
   1. A. The Shotgun Paintings
   2. B. MythBusters and the “Painting with Explosives” Test
4. Conclusion

I. The Cut-Up Method: A Primitive LLM

Burroughs, along with collaborator Brion Gysin, pioneered the cut-up method: physically cutting up existing texts (newspapers, novels, speeches) and rearranging the pieces into a new, fractured, and often deeply unsettling narrative.

This process is strikingly similar to how modern LLMs like ChatGPT and Gemini function.

• The LLM Analogy: An LLM does not think or create from a void; it processes a massive corpus of existing literature, code, and information. When prompted, it rearranges and probabilistically selects the most likely next word, generating a coherent-seeming, but fundamentally derivative, output.
• The Burroughs Protocol: Burroughs’s cut-ups rearranged existing literature into something new by randomizing the sequence of pre-existing linguistic units (sentences and phrases). This act of recombination and randomization fundamentally parallels the stochastic text generation employed by today’s models. It is an act of recombinant literature, using the source text as a dataset to generate “new” meaning through algorithmic chance, much like a primitive, analogue language model.

II. David Bowie’s Verbasizer: The First LLM on a Laptop

The conceptual leap from the analog cut-up method to modern LLMs was bridged by David Bowie, a longtime admirer of Burroughs, who brought the process into the digital age in the 1990s.

• The Transition to Software: Working with programmer Ty Roberts, Bowie co-created a Macintosh application called the Verbasizer. This custom software automated Burroughs’s process: the artist would input his own writings and other texts, and the Verbasizer would mechanically cut the sentences into phrases and then randomly reorder them.
• A True Proto-LLM: This process moved beyond the physical randomness of paper scraps to a digital, algorithmic randomization, creating a “real kaleidoscope of meanings” for lyrics on albums like Outside (1995). The Verbasizer was a dedicated text-generation program designed explicitly to break writer’s block and introduce unpredictable linguistic novelty—a direct ancestor of how writers use modern AI to generate text prompts.

III. Generative Chaos: Removing the Human Hand

The concept of removing the human element from artistic composition, initiated by Burroughs’s shotgun art, found both a high-art example and a spectacular pop-culture echo.

A. The Shotgun Paintings

Towards the end of his life, Burroughs created abstract expressionist paintings by firing a shotgun at cans of paint placed in front of a plywood canvas. The resulting splatter and chaotic patterns became the final artwork.

• The Dehumanized Artist: By using a weapon and the random force of an explosion, Burroughs outsourced the compositional decision-making to a non-human, mechanical force. The artist’s role shifted from creator to curator—he set the parameters (the paint, the canvas, the gun) and allowed the algorithm of physics and chance to execute the final piece. The shotgun acted as his generative algorithm.

B. MythBusters and the “Painting with Explosives” Test

The MythBusters episode “Painting with Explosives; Bifurcated Boat” (2013) provides a fascinating, high-octane echo of Burroughs’s method.

• The Algorithm of the Boom: Jamie Hyneman and Adam Savage tested whether one could successfully paint a room using explosives and paint. The goal resulted in pure chaotic, generative abstract art.
• The High-Energy Algorithm: Like Burroughs’s shotgun blast, the explosives completely removed intentional brushwork. The “art” was the product of chemistry, physics, and explosive force—a high-energy, real-world generative algorithm. The hosts became mere engineers setting the parameters for a non-human process to create an unpredictable, visually striking outcome.

Conclusion

Burroughs’s experiments were not merely artistic gimmicks; they were profound attempts to explore the limits of language and authorship. By embracing the algorithm of chance, whether through the rearrangement of words or the physics of a bullet, he laid the conceptual groundwork for a future where creativity is increasingly mediated and shaped by non-human intelligence, connecting the literary underground of the 20th century directly to the AI labs of the 21st.