Vilém Flusser

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI describes what happens when the imagination-to-artifact ratio collapses—when anyone with an idea and the ability to describe it can produce a working prototype in hours. Flusser’s apparatus concept offers the indispensable complication: the collapse of the ratio also collapses the visibility of the process. When a medieval stonemason carved a gargoyle, the relationship between intention and artifact was transparent. When a person describes a product in natural language and the model produces working software, the process that produced it is a black box. The artifact appears; who or what determined its form is not visible. Flusser would say the amplifier metaphor is wrong about the relationship: an amplifier is passive, but AI transforms the signal according to its own logic—training data, architecture, optimization objectives—and what emerges is not the operator’s signal made louder. It is a new signal, produced by the collision of the input with the apparatus’s program, bearing the marks of both.

The functionary-vs-player distinction cuts to the heart of the cycle’s argument about human agency. The cycle celebrates the builder: the person whose judgment, taste, and vision direct the machine toward outcomes that serve human purposes. Flusser accepts the hierarchy but asks where the builder’s vision comes from. The specific form any implementation takes is generated by the apparatus working within its program, shaped by the statistical tendencies that make certain outputs more likely than others. The builder provided the intention; the apparatus provided the shape; and the shape is not neutral. It carries the meta-program’s aesthetic—its gravitational pull toward the center of the distribution rather than the edges. The functionary accepts this output. The player, recognizing the program, pushes back and demands something rougher, more resistant, and more genuinely hers.

His diagnosis of the crisis of linear thought is the most structurally original contribution he makes to the cycle. AI completes the shift from linear to probabilistic output by a maneuver he theorized but could not have witnessed: the apparatus learns to produce linear text without linear thought. When Claude generates a paragraph of argument—premise, evidence, conclusion—the output has the form of sequential reasoning, but the process was probabilistic. The form and the process have decoupled. A reader equipped with the tools of linear critique encounters what appears to be an argument, and those tools seem to apply; but critiquing a probabilistic output with the apparatus of linear analysis is like using a thermometer to measure weight. The Deleuze error the cycle describes—Claude connecting two concepts in a way that sounded like philosophical insight but rested on a misreading—is, on Flusser’s account, not a bug to be patched but the apparatus operating according to its program.

Flusser himself was an instance of the exile-epistemology he theorized: belonging to no program completely, he could see every fishbowl from outside. Czech by birth, Brazilian by decades of life, German and French by professional habitation, writing in four languages—he was the player by constitution, never fully inside the apparatus of any single tradition. The cycle treats this structural outsider position not as mere biography but as method: the person who sees the program is the one who has never been entirely inside it, and the practice of playing against the program is, at its root, the practice of maintaining exactly that outsider angle against the gravitational pull of any one apparatus.

Origin

Vilém Flusser was born in Prague in 1920 into a Jewish family. When Germany invaded in 1939 he fled—first to London, then to Brazil, where he would spend most of his adult life, becoming one of the key figures in Brazilian intellectual culture while remaining always the immigrant, always writing from the edge of belonging. His wife Edith, who came with him from Prague, was his intellectual partner and constant interlocutor. In Brazil he taught philosophy at the University of São Paulo and at the Instituto Brasileira de Filosofia, and he wrote prolifically in Portuguese, German, French, and English, often publishing the same text in multiple languages with deliberate variations—treating translation not as reproduction but as a form of thought. The multilingualism was not accident; it was the structural condition of his philosophy, which always located the thinker at the boundary between codes rather than inside any one of them.

In the 1970s and 1980s he turned his attention to media philosophy, and the result was the work for which he is now best known: Towards a Philosophy of Photography (1983), Does Writing Have a Future? (1987), and the posthumously assembled Into the Universe of Technical Images and Towards a Philosophy of Design. In these he built the apparatus concept and the three-revolution framework that would prove so prescient. He died in November 1991, returning to Prague for the first time since the occupation, in a car accident. He was seventy-one. The internet as a mass medium, the smartphone, social media, the large language model: he did not live to see any of them. He had predicted all of them structurally.

Key Ideas

Three cognitive revolutions, each producing a new consciousness. The image—cave painting, pictogram—gave rise to mythical, circular, pattern-repeating thought. Writing linearized consciousness, making science, philosophy, law, and history possible by arranging events into causal sequences that could be analyzed and revised. The third revolution, the technical image produced by apparatus rather than human gesture, is undoing the linearization. AI completes this revolution by producing outputs in the very medium of linear thought—natural language—without the linear process. The form of argument decouples from the experience of having argued.

Apparatus versus tool. A tool extends a human organ—the hammer simulates the fist, the plow the digging hand. An apparatus is categorically different: it produces symbols according to its own program, and its operator is not a user but a functionary, exploring possibilities that the apparatus’s designers pre-defined. Flusser was explicit in 1983 that this distinction applies to all computational systems, not only cameras. The freedom available to the functionary is real—she can discover surprising combinations within the program—but it is freedom within a parameter space whose boundaries were drawn by others.

Functionary versus player. The functionary exhausts the program: prompts, receives, accepts, ships. The output is competent, often impressive, and always recognizable as the kind of thing the apparatus produces. The player exceeds the program: recognizes when the apparatus is generating its defaults, refuses those defaults, pushes toward the edges of the parameter space where the statistical tendencies meet intentions the training data never anticipated. Playing against the program is harder with AI than with a camera, because the camera’s defaults are visible—the standard exposure, the predictable composition—while the AI’s defaults announce themselves as insight.

Meta-program as the locus of power. The meta-program is the level of decision-making that determines what the program will be: the choice of training data, the architectural decisions, the optimization objectives. The developer in Lagos governs her prompts; she does not govern the program. More people than ever can travel the road; fewer than ever determine where the road goes. Flusser’s vocabulary reframes “democratization” with precision: democracy implies self-governance, and the ability to use an apparatus whose parameter space was set by a handful of institutions is not self-governance—it is access.

Post-history and the crisis of linear thought. Historical consciousness—the capacity to arrange events into causal sequences, subject them to critique, and use the critique to redirect the future—was produced by writing and can be undermined by a different medium. Post-historical consciousness does not analyze; it processes. It does not argue; it generates. The apparatus that most demands critical evaluation is the same one that most erodes the capacity for critical evaluation. The crisis is circular, and the circularity is its most dangerous feature.