In 1983, a Leningrad collective calling itself the Anonymous Typing Pool circulated a samizdat detective novel titled The Case of the Missing Archive. The book followed the formula of a Soviet-era whodunit: a missing dossier, a cast of suspicious bureaucrats, a chain of interrogations, a final revelation. Readers who knew the genre could predict the beats. But those who knew the city could read something else. The missing archive was not fictional. The names of the suspects, slightly altered, belonged to real KGB officers. The interrogation transcripts reproduced, almost verbatim, the language used in closed trials of dissidents. The detective genre was not a disguise; it was a machine. It processed political reality into a form that could travel, be copied, be read aloud in kitchens, and, if necessary, be disavowed as mere entertainment. The plot was the alibi.
This is not a story about how censorship forces artists into metaphor. Metaphor is what you use when you want to say one thing and mean another. The Anonymous Typing Pool was doing something more precise: they were building a narrative engine whose moving parts—suspect, clue, interrogation, resolution—were already legible to every reader of Soviet detective fiction. The engine did not hide the political content; it produced it. The constraints of the genre became the conditions of legibility. The formula was the message.
I return to this novel now because the question it poses has outlived the regime that provoked it. What does it mean to build a plot generator under pressure? Not a metaphor generator, not a symbol generator, but a device that takes raw political material and processes it into narrative form according to rules that are themselves a response to power. The Anonymous Typing Pool understood something that Western creative-writing manuals rarely admit: constraints are not obstacles to creativity. They are the factory floor. The question is who owns the factory.
The Detective as Document
Soviet dissident culture had a peculiar relationship with genre fiction. Science fiction, detective stories, and historical adventure novels were among the few mass-market forms that the state permitted, and therefore among the few that could circulate without immediate suspicion. But permission is not the same as control. Writers in Moscow, Leningrad, Kyiv, and Tbilisi learned to treat genre conventions as a kind of legal code: you could do almost anything inside them, provided you respected the surface architecture. A detective novel needed a crime, an investigation, a solution. What the crime was, who investigated it, and what counted as a solution—these were the variables that turned a formula into a document.
Consider the Riga-based group that produced a series of samizdat thrillers in the late 1970s. Their novels always began with a disappearance. The missing person was never a political figure; it was a librarian, a factory foreman, a retired schoolteacher. The investigation would uncover a network of informants, a file kept by a local Party official, a room in a municipal building that did not appear on any floor plan. By the final chapter, the detective—often a journalist or a disillusioned militiaman—would find the missing person in a psychiatric hospital, committed for “anti-Soviet agitation.” The plot was identical across five novels. The repetition was the point. The formula was a machine for converting individual disappearances into a pattern, and the pattern into evidence. Each novel was a deposition wearing a dust jacket.
This is not the same as allegory. Allegory requires a stable correspondence between the fictional world and the real one: this character stands for that politician, this event stands for that historical moment. The samizdat detective novel worked differently. It used the genre’s demand for causality—every clue must lead to the next, every motive must be explained—to force a kind of narrative logic onto events that the state insisted were random, isolated, or imaginary. The detective form did not hide the political argument; it made the argument, by insisting that disappearances have causes, that causes have names, and that names can be written down.
The Plot as Survival Technology
When I say that these writers built plot generators, I mean it literally. They developed reusable narrative structures that could be fed different inputs—different names, different locations, different dates—and produce legible outputs. This was not a metaphor for resistance; it was a technology for survival. A plot generator is a device that takes constraints and turns them into stories. Under censorship, the constraints are external: you cannot name certain people, you cannot describe certain events, you cannot publish through certain channels. The generator internalizes those constraints and makes them productive. It says: given that I cannot say X, what can I say that will carry X’s weight, X’s shape, X’s consequences?
This is the opposite of the Western romantic notion of the artist breaking free of constraints to express a pure inner vision. The dissident plot generator does not seek freedom. It seeks functionality. It asks: what story can I tell that will survive a search, a denunciation, a trial? What narrative form is sturdy enough to be copied by hand, read aloud, memorized, and reconstructed from fragments? The answer, time and again, was genre. Genre is durable because it is predictable. Predictability is a liability in art criticism; it is an asset in a police state.
Contemporary artists working under authoritarian conditions have inherited this logic, though the technologies have changed. In Belarus, where independent media has been systematically dismantled since 2020, a group of video artists began producing short films structured as instructional videos. The films show how to pack a suitcase, how to clean a wound, how to reinforce a window against a blast wave. The instructional format is a plot generator: it has a fixed sequence (problem, materials, steps, result), a neutral tone, and a plausible deniability. No one can be arrested for teaching first aid. But the films are not about first aid. They are about what the state does to bodies, and what bodies need to survive it. The instructional genre processes political violence into a form that can be uploaded, downloaded, and watched in a browser tab that also contains a recipe for borscht. The plot is the alibi.
The Machine in the Museum
Western art institutions have a habit of treating these narrative strategies as aesthetic choices, as if the artist selected a genre the way a painter selects a palette. This is a category error with political consequences. When a curator frames a Belarusian instructional video as “exploring the poetics of the everyday” or a samizdat detective novel as “interrogating the boundaries of fiction and reality,” they are performing the same act of depoliticization that the state performs when it calls a political prisoner a hooligan. The genre is not a poetics. It is a technology. It was chosen because it works, not because it is interesting.
The Authors Guild, in its AI Best Practices for Authors, warns that AI outputs are “generic mashups of pre-existing works” and that the writer’s “original voice, thinking, and creativity” are what make the writer who they are. The document is concerned with the ethics of using large language models trained on pirated books. But its deeper argument—that a tool that homogenizes voice is a threat to the very thing that makes writing human—has an unexpected resonance with the history of dissident plot generators. The samizdat writers were not using AI, but they were using a different kind of generic mashup: the detective formula, the instructional video, the historical adventure. The difference is that their mashup was not a surrender of voice but a preservation of it. The formula protected the voice from being erased. The generic was the condition of the specific.
This is the paradox that Western criticism consistently fails to grasp. When an artist in Minsk or Tashkent or Bishkek adopts a mass-market genre, they are not making a pop-culture reference. They are activating a survival technology. The genre is a container that has already been tested against state power. It has a known legal status, a known readership, a known set of expectations. The artist can fill it with whatever will fit, and what fits is often more than the state can afford to acknowledge. The plot generator is not a neutral tool; it is a weapon that looks like a toy.
Reverse-Engineering the Homemade Plot Device
What would it mean for contemporary artists and critics to reverse-engineer these historical methods? Not to imitate them—the conditions that produced the samizdat detective novel are not reproducible, and nostalgia for Soviet dissidence is its own kind of kitsch—but to understand the underlying logic and adapt it to present constraints. The logic is this: a plot generator is a set of rules that converts raw material into narrative form. The rules are chosen for their durability, their legibility, and their capacity to carry weight that cannot be carried openly. The generator is not the story; it is the machine that makes the story possible.
This is where the contemporary tools become interesting, not as replacements for human creativity but as conceptual parallels to the homemade devices of the past. A novel plot generator that structures narrative around genre conventions and iterative constraints is, in one sense, a descendent of the samizdat detective formula. Both are machines for processing material through rules. Both treat constraints as productive rather than limiting. Both produce outputs that can be revised, locked, and regenerated—the Anonymous Typing Pool revised their manuscripts through multiple samizdat editions, each one incorporating new information, new names, new disappearances. The difference is not in the mechanism but in the stakes. The samizdat generator was built because the alternative was silence or prison. The contemporary generator is built because the alternative is a blank page. The urgency is different, but the structural insight is the same: narrative is not a spontaneous outpouring. It is a manufactured product, and the manufacturing process matters.
Reedsy’s Plot Generator offers a useful case study. It asks the writer to define genre, tone, structure, protagonist, conflict, stakes, and supporting characters. It then produces a plot broken into acts, with the option to lock certain acts and regenerate others. This is, in effect, a narrative machine that externalizes the decision-making process. The writer becomes an operator of the machine rather than a conduit for inspiration. For a commercial novelist, this is a productivity tool. For a dissident writer in 1983, it would have been a survival strategy. The same architecture—input constraints, process through rules, output narrative—serves different functions depending on who is operating it and under what conditions. The tool is not the politics. The politics is in the relationship between the operator and the constraints.
The Unfinished Archive
In 2021, a researcher in Yerevan discovered a stack of samizdat detective manuscripts in the basement of a former printing house. The manuscripts were incomplete. Some were missing the final chapter; others had pages torn out, or names scratched out and replaced with different names in different handwriting. The researcher, Anahit Grigoryan, spent two years cataloging the fragments and published her findings in a small-run academic journal. Her conclusion was that the manuscripts were not unfinished works but finished processes. The missing chapters were not lost; they were never written, because the information they would have contained was too dangerous to commit to paper. The scratched-out names were not errors; they were updates, as the real people behind the fictional characters were arrested, released, or killed. The manuscript was not a text; it was a living document that changed as the political situation changed. The plot generator was still running.
Grigoryan’s discovery is a reminder that the plot generator is not a product but a practice. It does not end when the book is bound or the film is screened. It continues as long as the conditions that produced it continue. The samizdat detective novel was not a finished artwork; it was a machine for processing an ongoing emergency. The contemporary equivalents—the instructional video, the Instagram story, the encrypted chat—are the same kind of machine. They are not art objects. They are survival technologies that happen to produce art as a byproduct.
This is what Western criticism misses when it treats Eastern European dissident art as a genre. Dissident art is not a genre. It is a condition of production. The plot generator is not a stylistic choice; it is a response to a specific set of constraints that include surveillance, imprisonment, and the systematic destruction of archives. To treat the resulting work as “political art” is to misunderstand the relationship between the work and the politics. The work is not about politics. It is produced by political conditions, and its form is a direct consequence of those conditions. The detective novel is not a metaphor for investigation; it is an investigation, conducted under conditions that make investigation illegal. The instructional video is not a commentary on state violence; it is a response to state violence, distributed through channels that the state has not yet learned to monitor. The plot is not a representation of reality. It is a tool for surviving it.
What the Generator Cannot Do
There is a limit to the plot generator as a critical concept, and it is important to name it. A plot generator can process material, but it cannot create material. It can structure experience, but it cannot have experience. The samizdat writers knew this. Their novels were powerful because they were filled with specific, observed details: the color of a particular corridor in a particular KGB building, the exact phrasing of a particular interrogation, the name of a particular informant. The generator gave these details a shape, but the details themselves came from lived experience. The machine was nothing without the raw material fed into it.
This is the danger of treating contemporary plot generators—whether AI-powered or not—as replacements for the difficult work of observation, memory, and testimony. A tool that structures narrative can be useful, but it cannot replace the act of witnessing. The Anonymous Typing Pool did not use a plot generator to avoid the risks of documentation; they used it to make documentation possible. The genre was a vehicle for the real, not a substitute for it. Any contemporary artist or critic who wants to learn from these historical methods must start with the same commitment: the plot generator is a means, not an end. The end is the truth that the generator makes it possible to tell.
In the summer of 2024, I visited the archive in Yerevan where Grigoryan’s manuscripts are stored. The building is a former Soviet printing house, and the basement still smells of ink and damp paper. The manuscripts are kept in acid-free boxes, each one labeled with a date range and a list of the real people whose names appear in the text. Grigoryan showed me one manuscript in particular: a detective novel from 1985, set in a provincial town, about the disappearance of a local historian. The novel is unfinished. The final chapter is missing. In the margin of the last page, someone has written in pencil: “He was taken on the 14th. We do not know where.” The note is not part of the fiction. It is a message from the operator of the machine, reporting a malfunction. The plot generator had encountered a constraint it could not process. The real had broken the frame.
This is the final thing to say about plot generators, whether homemade or algorithmic. They are powerful, but they are not omnipotent. They can process almost anything, but not everything. There are events that resist narrative, experiences that cannot be shaped into a story without violence to the truth. The dissident writers knew this. They built their machines to handle as much as possible, but they also knew when to stop the machine and write in the margin: We do not know where. The margin is not a failure of the generator. It is the generator’s conscience. It is the place where the machine acknowledges its limits and hands the work back to the human, who must now do what no plot generator can do: bear witness without a plot.