July 8, 2025
photography AI memory media-theory philosophy

Palm Trees Are Noise

Pleasant Murmur or Distracting Noise

by Markus Oberndorfer

From “Staging the Ordinary”1 to “hallucinating the everyday”: an attempt, traced through the example of the palm tree, and grounded in media-philosophical questions surrounding photography and generative AI techniques.

The palm fronds caress our cheeks like soft brushes, their soothing rustle, carried in from afar, enchants us like the Sirens once bewitched Odysseus on his Odyssey2. They seem to question what is. Such poetic reflections ­— thoughts on mere being3, on documenting, staging, or generating the everyday, and the resulting considerations in this text — mark the beginning of an artistic search for roots and traces within the ambient noise of today’s media landscapes.

An Old Dream

Since 2015, my project REVISITED4 has taken me on a journey — from photography5, through video, film, 360° VR video, on to generated images and 3D-modeled interpretations of the Sunset Strip. Across all these numerous media representations — including that in the computer game Grand Theft Auto V6 — the place remains clearly recognizable, its distinctive qualities even tangible to a certain degree. Why is that? What defines its essence — whether real or virtual, documented or assembled from fragments? What network of features and characteristics is needed to shape, represent, and imbue a spatially limited section — a place — and fill it with meaning in such a way that we feel a sense of rootedness? That its essence, fundamental mood and atmosphere7 emerge tangibly — as a “situational synthesis of everything that appears in a given area?”8

Material and immaterial qualities associated with architecture, non-places, and elements of vegetation are fundamental in defining a place. They are essential for locating oneself in space — for integration, identification, and orientation. Together with events that define it as a lived space for individuals and the collective ­— events about which stories are told — these qualities shape the identity of a place. “This does not only include the presence of these exact things in space, but also the rhythm they inscribe into urban life, or the ways in which people inhabit the (micrological) city.”9 This rhythm is constantly changing “through the necessity, arising from an existential need, to constantly re-appropriate an ever-changing environment.”10

Expectations toward a place also vary depending on perspective. For residents, the focus is usually on quality of life and fulfilling everyday needs. Tourists, on the other hand, visit certain places in search of something they hope to find — often based on a pop-culturally influenced, biased image. In the case of Los Angeles, this might be the glamour associated with Hollywood Boulevard, the “Sex, Drugs & Rock’n’Roll” attitude linked to Sunset Boulevard, or the lifestyle evoked by singing Moog leads11 and Roger Troutman’s talkbox12. 2Pac’s California Love13 which thrives on a controversy: a dream of freedom and self-realization intertwined with an often challenging reality of everyday life.

Just as the Sunset Strip — with its billboards, neon signs, clubs, landmarks, and the legends surrounding it — is deeply embedded in our cultural memory and imagination, Los Angeles and Southern California more broadly embody the promise of sun, beach, lightness, and palm trees.

Yet apart from the California fan palm, none of the species — including the so-called “Skydusters”, which became a symbol of Los Angeles in the postwar era — are actually native to the region.14 They were planted only from the early 20th century15, and are part of a carefully crafted image of the city that today serves as central identity anchor. Their omnipresence owes as much to the local film and television industry as to a culturally rooted, romanticized idea that has elevated palm trees to global symbols of luxury and exoticism. We associate them both with “sex, glamour, and celebrity”16 and with the “old dream” mentioned in the earlier quoted scene from Dune17: the sea breeze on the Riviera, the eternal summer in the tropics, the exotic Orient, the fertile oasis — that sense of lightness and freedom that is both seductive and illusory.

TO LIVE AND DIE IN L.A. 18

The palm tree became an icon embodying both nature and artificiality — suspended in a tension between exotic ornamental plant and marker of urban identity. Yet, as the regions it often represents become increasingly affected by hotter, drier, and more volatile climate — the palm proves ever more incompatible, as it provides little shade, purifies the air only to a limited degree, and demands far too much water and costly maintenance.

For this reason, the city of Los Angeles decided some time ago to stop replanting palms that die of old age or fall victim to pests and fungal diseases brought on by shifting climatic conditions. Instead, the city replaces them with native trees more resistant to drought and wildfires, such as the California live oak.(fig_p.28) Exceptions are made only in locations where palms hold historical or symbolic value. For example stretches of Sunset Boulevard, East 43rd Street, or Highland Avenue.19

During recent fires, palms once again drew attention, as their dead, highly flammable fronds likely contributed to the rapid spread of flames. Esther Margulies, a specialist in climate- and wildfire-resilient urban planning, remarked: “We need to change the perception of LA. When it comes to living things, we really should have a deep respect for the trees that are native and indigenous to this area and really will thrive here.”20

Much like the date palms in Dune that went up in flames during the Harkonnen attack, the so-called Skydusters are expected to begin disappearing from the cityscape for a variety of reasons — well before the end of this century. An event the L.A. Times referred to as “the death of a star.”21 A star which, arranged in serial formation along the roadsides — documented in Ed Ruscha’s Street Archive, in the 360° videos of REVISITED, and in the ever-changing digital archive of Google Street View — has become “not only part of a green paradise, but also a symbol of the endless sprawl of streets and concrete.”22 It is precisely this environment — the one Los Angeles so powerfully proclaims — that makes it difficult for Hollywood to convincingly portray other locations. “You can shoot Anywhere, U.S.A., here, as long as you avoid palm trees,” said a location manager. “That’s our biggest bane, the palm tree, because if you see it, it’s either Southern California or Florida.” As a result, the greens crews in film productions work to turn the city’s floral abundance into a more nondescript vegetation. “They strip the fruit from exotic trees; they add fake bark and leaves to the trunks of palms. Cinematographers choose angles that exclude palms, editors crop them out, and if necessary, special-effects engineers digitally remove them.”23

As the example and work of the greens crews make clear, the (de)contextualization of the palm into a flexible, universally deployable symbol — one that can be placed anywhere ­­— is not merely a digital phenomenon. As a deliberately used symbol, it has long existed untethered from its original habitat. In the digital realm — and especially through AI — it ultimately becomes pure atmospheric pixel density: the code of a globalized visual language stripped of all materiality.

OF MERE BEING24

The title of Wallace Stevens‘ 1971 poem — quoted in two-line excerpts throughout this section and in my series of books, Palm Trees Are Noise — appears to already draw attention to the (semi-)artificial environments we inhabit today, often without even realizing it.

As an artistic approach to virtual visual worlds — and, subsequently, generative AI — my first step was to remodel (fig_p.29) each palm tree depicted in Ed Ruscha’s A Few Palm Trees25 in 3D, thus translating them into digital space. These reconstructions formed the foundational dataset for all subsequent media-based representations of palms, illustrated across an interlinked and continuous series of 40-page books. Step by step, the 3D wireframe depictions (A Few Generic Palm Trees) dissolve into noise (A Few Diffused Palm Trees), before re-forming as AI-generated, realistic looking black-and-white objects (A Few Generated Palm Trees), and finally appearing in colorized, deliberately staged sceneries (A Few Staged Palm Trees).

Ruscha’s book presents 14 black-and-white images of palms, each isolated from its surroundings and accompanied by notes on location and capture method. Their model-like stylization already poses a fundamental question: “real” or “fake”? Building upon these visual references, my representations provide access to a world increasingly shaped and (co)designed by data. In the same manner as the superordinate project REVISITED, Palm Trees Are Noise reflects on the evolutionary trajectories of technology and media — and on the cultural and societal shifts that accompany them.

‘The palm stands on the edge of space.
The wind moves slowly in the branches’ 24

For A Few Generic Palm Trees26, the first part of the series, the 3D models were visualized as wireframes — a translation into the two-dimensional space of the book. The skeletal forms, the base of all 3D objects prior to texturing or 3D printing, resemble templates from a coloring book — open to interpretation. Once filled with textures, surface detail, and color — staged by virtual lighting, and, if desired, animated — they shine anew.

Whereas a coloring book invites us to artistically interpret and intuitively fill the blanks, an image-generating AI performs this act more abstractly: it supplements, alters or reconstructs visual material based on statistical probabilities and creates new visual “realities” from mere noise. For an AI to (be able to) generate something via a model, it must first be trained. During the so-called forward diffusion process, data is deconstructed; in this encoding phase, a neural network captures key features of the original data. In the case of images this means that visual information, such as colors, shapes, textures, and other elements, that may seem trivial to humans, but are highly relevant to machine learning systems, are preserved within the so-called high-dimensional latent vector space. It, however, can not be imagined like the cross-sectional diagrams of the German botanist von Martius, who attempted to classify palms in his 1853 Book of Palms27, nor as the sketches of Caspar David Friedrich, which he used like a modular system to compose his paintings. Here, colors and shapes no longer exist as concrete image or pixel information, but rather as highly abstract mathematical coordinates — as points in an invisible space. To illustrate how this latent space operates, I‘d like to compare it to the processes of remembering and dreaming — even though this analogy ignores essential differences between the functionality of the human brain and AI. “The brain stores experiences as memories. Yet, these must not remain static; they need to be dynamic, adapting to new circumstances over time as the world around us changes.”28 Suppose, that the brain navigates a latent-like space of stored impressions when remembering or dreaming — selecting, recombining, and interpreting them through an internal logic — then dreams in particular provide a close parallel. When dreaming the brain invents people and places or blends real memories with other influential impressions; for example an emotionally charged film scene. It creates narratives that may defy logical structure, yet feel plausible to us until conscious reflection. One could say it moves through “states of affaires (facts, i.e. stored impressions), applies programs (how these are structured and combined, and sometimes encounters problems (discontinuities within the program).”29

Generative AI, operates without ever having experienced (seen, felt or perceived) anything. It analyzes data, processes numerical patterns and recombines stored information — learned abstract representations of data (images, melodies,…) — by bridging gaps between contextual points and creating seamless transitions. Often realistic, though at times far-fetched, these moments are commonly described as an AI hallucinating.

In greater depth: Technically, the latent vector space is traversed by mathematical models (e.g. neural networks). Every point (every position) within it — and all intermediate points created through contextual relations — can be decoded into a perceptible visual representation and meaningful human-readable forms (images, sentences, music). This technique also enables generative AI models to filter and replicate familiar attributes of an artist‘s style because although the exact characteristics that constitute an artist‘s signature are very difficult to articulate, statistical methods allows AI to identify latent attributes that serve as convincing indicators of stylistic similarities. The representation of images and text as coordinates in space makes it possible to create images that are “like” other images (or categories of images) in a new and meaningful sense. This process is reminiscent of image interpolation during which transitions between values or points are calculated and estimated to fill in missing details or create new transitions. AI performs a similar operation. Not with concrete pixels however, but with vectors, enabling it to create fluid transitions and more coherent results. Therefore, if you interpolate an orange and a grapefruit in latent space, you get a tangelo.30

‘The palm at the end of the mind.
Beyond the last thought, rises.’ 24

When AI generates something, for example an image, it visualizes the data and features stored in this latent space through digital denoising and probability distribution, and transforms abstract information into visual forms. To artistically explore the processes of noising (as in training) and denoising (as in generation) I — in A Few Diffused Palm Trees31 — applied Gaussian noise to my 3D wireframes, gradually eroding detail until only spectral traces of their original “selves” remained. This mirrors the fundamental logic behind many generative AI systems — especially those of diffusion models, which are widely used in image, text and music generation, 3D modeling and video synthesis. At the same time, this offers “a glimpse into the nature of the virtual image,”32 because the denoising process ­which reconstructs a plausible image from seemingly chaotic noise, relies on specific architectural logic: the U-Net33. In a cyclical sequence, this encoder-decoder-based network architecture abstracts the noisy data into abstract feature representations, while the decoder creates coherent, plausible visual structures from these compressed informations.

‘The palm stands on the edge of space.
The bird’s fire-fangled feathers dangle down’ 24

In A Few Generated Palm Trees34, the third part of the series, I used AI to approximate Ruscha’s palms based on my 3D models. First, I employed CLIP Interrogator35, a prompt-engineering tool developed by OpenAI and Salesforce that can potentially generate text prompts matching a given image, to describe each palm. One result reads: “A black and white photo of a palm tree with a white background, museum catalog photograph, date palm trees, a palm tree, palm body, archival pigment print, the palms come from the ground, by Robert Rauschenberg, palm pattern visible, style of Hiroshi Sugimoto, palm tree, by Edward Ruscha, 120mm black and white photograph.”36 These prompts were then used to generate structures on the 2D-wireframes originally created for A Few Generic Palm Trees and fed into Stable Diffusion ControlNet37.

‘The bird sings…
Without human feeling, a foreign song’ 24

For A Few Staged Palm Trees38, the fourth part of Palm Trees Are Noise, I implanted the 3D depth maps of my palms(fig_p.31) into Google Street View screenshots — each taken at the 1971 locations indicated by Ed Ruscha. Using ControlNet, forms were extracted from these templates that then served as the basis for new images, once again generated in Stable Diffusion, guided by custom prompts. The resulting compositions thus are products of layered negotiation: the street views as they presented themselves to me; my selected perspectives and “photographic” framing in the 360° environment; the 3D models; my textual instructions — and the location data, which inevitably sent me on a mostly futile search, as many of the original trees no longer exist39 today.

To once fully trace the transformation process from a trees root, to a photograph, and a generated look-alike — I added one palm of personal importance to the series: Palm XV. Based on a photograph taken by myself (Palm Trees I-XIV are based on Ed Ruschas), it went through the same pipeline: it was isolated, re-modeled in 3D, visualized as a generated object, and staged within an arranged environment. Palm XV thus not only serves as another example of this transformation, but also as a reflection of how the photographic appropriation of the pre-existing is always anchored in a tangible relationship between subjective selection and physical presence.

Original, Translation, Interpretation

Instead of reproducing images 1:1 trough medial translation as Sherrie Levine did in After Walker Evans — an iconic work of Appropriation Art — my goal in A Few Generated Palm Trees was to produce look-a-likes that, though adorned in similar feathers, only resemble their reference at first glance but clearly diverge upon closer inspection. By simply photographing Walker Evans’ images — published in his catalog First and Last — and claiming them as her own without further alteration, Levine “triggered controversy on the subjects authorship and original versus reproduction”40 in the early 1980s. Though, like Palm Trees Are Noise, her work — as well as that of Richard Prince41 and other members of the Pictures Generation42 — was never about copying, but rather about navigating the tension between original, translation, and interpretation in a changing media landscape. It is precisely this field of tension that, in my opinion, needs to be further explored and negotiated due to the rise of generative techniques and their appropriation of vast quantities of partly copyright-protected data.

Also through artistic works like these, that draw on the tradition of the Pictures Generation and central questions of Appropriation Art. Instead of a direct medial reproduction, as Levine performed with photographs, my work produces images based on already subjectively interpreted 3D representations — driven by established visual codes and references.

The juxtaposition of Ruscha’s palm trees — based on analog photographs — and my derivative representations and medial excursions into the realms of 3D and generative AI, thus raises key questions about appropriation: what, and how much, of a source (parent) is transferred into, and retained within, a respective medium. At the same time, it examines the concept of the generated look-alike as a virtual image in relation to our concept of reality. It highlights the tension between recognizability and algorithmic re-creation, between resemblance and deviation, and aims to encourage a critical reflection of a complex, constantly changing multi-medial world — a world shaped by a network of data, its analysis and automation, as well as code.

In 1989, Paul Virilio writes in The Vision Machine: “Avenues and public venues from now on are eclipsed by the screen, by electronic displays, in a preview of the ‘vision machines’ just around the corner. The latter will be capable of seeing and perceiving in our place.”43 While machines do not (yet) perceive, they analyze, interpret, describe, generate, automate, and present an endless stream of visual noise on screen.

Systemic Bias and the Illusion of Presence

Even if a model has been extensively trained on images of date palms until it recognizes and abstracts their characteristics and properties, it does not know, what a date palm actually is.

Does the “presence” of the generated palm then still correspond to an extension of its trace — like the light-induced mark on a photographic negative, signifying the palm‘s past presence and rootedness in the physical world? What does the pixel echo, hallucinated by AI, truly depict?

To begin answering this question, we must ask: Assuming that a neural network or model has been trained exclusively on human-made photographs, is there still something like a final, lingering trace of the referent — the real-world object — that we have been trying to mark since the dawn of photography? And what does it mean if there isn‘t?

Could an AI-generated look-alike, a product of statistical patterns, ever be considered an index of a plant? Similar to Ed Ruscha‘s isolated palm trees, which — despite their abstraction and though stripped of species-specific information — might still function as abstract references to broader botanical categories? Or is the AI-generated object closer to the works created by Karl Blossfeldt‘s students in his class Modeling After Living Plants — deriving not from nature directly, but from his photographed plant parts?

According to Charles Sanders Peirce‘s theory of signs, an index is a sign that maintains a direct, causal, or physical connection to its referent — like a footprint in the sand which indicates a person, or smoke that signals fire. “The indexical sign is thus connected to the referent (i.e., the depicted object) by a physical link.”44

While analog photography captures the trace of a real-world object, AI-generated images distort this reality through statistical reconstruction. In Baudrillard’s terms, such images evolve through a gradual detachment from the real: from mere representation (first stage), through distortion (second stage) to simulation (third stage), and ultimately existing as pure signs without real foundation (fourth stage) — as simulacra45. The palm hallucinated by AI no longer possesses indexical qualities; it resembles a self-referential image constructed by palm-like patterns — an echo of media representations and labeled keywords, encountered during training. Ultimately, the AI-generated images are neither indices nor conscious abstractions, neither scientific diagrams nor didactic models. The AI‘s look-a-likes are simulations of something it does not know and whose nature it does not comprehend. And yet, they are reflections of how humans have imagined the thing for centuries — shaped by cultural and individual bias, internalized in machine learning. They are algorithmic collages, statistical condensations, and mimetic constructions, that simulate resemblance without knowing what resemblance means. They depict without truly depicting; refer without knowing what to refer to. They are a form of hyperreality that poses as reality. Simulacra in Baudrillard‘s sense — synthetic constructions without physical foundation.

ORCHESTRATED (Staged) Seduction

The palm has long been understood as a cultural symbol — a stylized emblem shaped by theatrical construction — culturally charged, and deliberately staged. AI-generated content picks up on this exoticized, symbolic, and surface-oriented image of the palm. It not only reproduces it, but amplifies it through pattern recognition and recomposition. This marks the beginning of a “self-autonomization” of the sign — a play of meanings that manifests as staged “seduction,” a kind of “fetishism of the surface.” “This aesthetic staging encourages every form of fetishism, in which artificiality, rhetoric, and pose always outweigh reference to actual persons or bodies (and objects). In the light and mobile play of coquettish allusions, everything becomes a signifier of unstable eros.”46

Even though Ed Ruscha‘s palms, in these terms, are stylized, abstracted types of plants — isolated cutouts that present a distilled, typified idea of palm-likeness — they remain rooted in a physical context and connected to their referents through the location details provided in A Few Palm Trees. Not least through their analog-photographic origin and the trace that can be followed from the illustrations in the book, through mock-ups(fig_p.31) and contact sheets(fig_p.28), all the way back to the original negatives — and ultimately, to their precise location at the time they were captured. This referential chain is broken in AI-generated imagery. The synthetic palms do not depict any particular object but instead simulate the idea of a palm through statistical reconstruction. They are condensed constructs born of Con[dif]fusion47 — untethered from the material world — emerging from no specific viewpoint, no subject’s position within the spatial arrangement of things.

/Remember? The trace and its detachment from the Referent explored through the example of historically significant Objects

What becomes of historical context when nothing remains directly tied to the materiality of the analog — to the referent, and thus to a place (and a subject), but instead unfolds arbitrarily within imaginary worlds? Worlds that lack spatiality; that no longer possess a “system of places that determine one another through positions and distances?”48

Can an AI-generated image that depicts a structure resembling a World War II bunker be considered a historically relevant document if the original visual traces of the bunkers used in training have dissolved into noise, have reduced to fragmentary features in latent space (if at all)? Or is such a claim ulti-mately reserved for “photography in the narrower sense,”49 where light directly strikes a photosensitive surface, triggering a chemical reaction in silver halides embedded therein?

Take, for example, my photographs of the Atlantic Wall on Cap Ferret, shot exclusively on Reala negative film over more than 20 years. Due to staging the pre-existing in a consciously composed frame, their truthfulness to reality lies not so much in mimetic precision, but above all in their ability to situate a specific object in a specific scene at a specific moment — and, thereby, to locate us within post-Nazi Europe.

“According to Foucault places are marked by relations of placing. These placings, tying themselves to real places, also reconnect the utopias — the non-places, with the places. These heterotopias50 are “places outside of all places, even though it might be possible to indicate their location in reality.” They reflect power relations as well as phantasms, desire just as the people who are placed in these places and thus somehow construct them. The “heterotopias”, which are marked by deviance, by illusion and compensation, always refer to the order of society — they demand to be historicised. By that Foucault’s ideas provide interfaces to reflections on history and memory.”51 As Inge Marszolek writes in Looking at Bunker (Photographs) ­— Thoughts of a Historian.

AI-generated bunkers are not real places referring to other places. They are, in Baudrillard‘s terms, simulations — visual constructs entirely detached from reality. While actual bunkers are material traces of a specific historical moment, and the photographic staging of such sites constitutes a media-based interpretation, AI-generated bunkers are abstractions of a different order. They exist solely as algorithmic constructs, unanchored in space and time — simulations of bunkerness that have never left a trace. They are simulacra — untethered from referential grounding and historical context. If Foucault‘s heterotopias are real places that refer to other places, then AI-generated bunkers lose this heterotopic quality. They may contain bunker-like structures, but these structures are only statistically plausible approximations. “Bunker-likeness” — analyzed and composed through the lens of “a system in which linear continuity and dialectical polarity no longer exist, in a field unhinged by simulation.”52

In order to explore that AI systems probably only recognize visual patterns without understanding meaning, I began analyzing all 53 photographs in my book Foukauld53 using CLIP Interrogator. My goal with the project A BUS IN THE SAND54, which I started in 2023, was to determine which elements the AI identifies in these images — images that document the appropriation and disappearance of the Atlantic Wall bunkers on Cap Ferret — and how the model describes them. For instance, the photograph Foukauld, that gave my book its title (fig_p.36), was interpreted by the AI as follows: “There is a bus that is sitting in the sand on the beach, graffiti in an abandoned bunker, stunning large format photograph, naval landscape, penned with thin colors on white, full width, San Francisco, 85mm photograph”55 Such results demonstrate that the neural networks I used — CLIP and BLIP56 — primarily detect formal and stylistic parameters from the images. They are incapable of grasping deeper symbolic, cultural or historical meaning. These systems — trained to correlate visual and linguistic data — can recognize patterns, but not context. Accordingly, CLIP outputs “bus in the sand,” “graffiti in an abandoned bunker,” or “naval landscape,” yet remains oblivious to the affective dimensions embedded in the photograph or to the place‘s political and historical charge.

And yet, even if AI-generated images are built on surface-level patterns and lack inherent historical substance, these simulacra can still serve as screens for projection of historical and cultural discourse. Foucault often describes heterotopias as real places with specific functions or social and symbolic significance — cemeteries, prisons, theaters or museums, for example. Since AI-generated doppelgängers of culturally passed on images absorb, condense and remix the symbolism of the “real”, I wonder whether they represent an extension of heterotopia into the virtual realm. And whether, even if only metaphorically (since their documentary qualities have been lost in noise), they can also be understood as “space of thought.” Marshall McLuhan’s notion of the Electronic Age, in which, “all senses are stimulated simultaneously, and linear perception dissolves in favor of networked, mosaic-like thinking,”57 is also relevant in this context.

Media creates new perceptual realities, as the act of seeing and understanding objects, places, and events is increasingly shaped by screen-mediated representations. Consider how much time we spend wandering through the world with a smartphone in hand. The omnipresence of digital imagery is transforming the way we experience spaces, reality, and history — shifting away from a purely place- and event-bound felt-bodily perception (see also: “Se Souvenir”58). According to Hermann Schmitz, this shift would imply that the simultaneous stimulation of all senses, described by McLuhan, feeds back into the situation itself, thereby affecting the embodied perception Schmitz emphasizes — intermodally, as a “half-thing.”59

Even though AI-generated bunkers lack physical localization and have lost their indexical grounding in translation, they remain condensed representations of “bunker-likeness”, and — when credibly staged(fig_p.36) — can function as compressed symbols. Not by directly referring to a specific, locatable object, but instead by operating through the associations and emotional responses they can evoke in the liminal space between simulated bunkers and their real-world counterparts. Paradoxically, though they are not direct signs of a concrete past, they still articulate a digitally constructed pseudo-culture of remembrance ­— one that has detached itself from actual sites, yet continues to resonate with cultural memory through associative resonance.

Photography and its role as a medium of memory and cultural identity are facing new challenges in the light of generative AI and the paradigm shifts it brings. The question of how we evaluate photographs, and simulations generated from noise, in terms of their informational value is, and will remain, one of the great challenges of our time — especially as they become almost indistinguishable — especially on screen.

The uncertainty about whether — and what — we actually see is reminiscent of a passage in Virginia Woolf‘s Orlando: “Our most daily movements are like the passage of a ship on an unknown sea, and the sailors at the mast-head ask, pointing their glasses to the horizon: Is there land or is there none? to which, if we are prophets, we make answer: ‘Yes’; if we are truthful, we say: ‘No.’”60

Labyrinth of Desires

Media — particularly photography and screen-mediated virtual imagery of the “world out there” — also play a central role in the series SILO61, adapted from a screenplay by Graham Yost and based on the science-fiction novel of the same name by Hugh Howey. In this dystopian narrative, humans have been living in an underground silo for 140 years, following an environmental catastrophe that remains unspecified until the end of the second season. The “Pact,” a quasi-constitution, administrates social life through clearly defined hierarchical structures and areas of responsibility. Restricted access to technology and knowledge, as well as control of information, serve as instruments of power. This power not only operates through direct repression, but also — in the sense of Foucault’s concept of governmentality62 — by inducing people to control their own behavior in accordance with the expectations of the system in question.

Photographs and direct visual representations — along with any knowledge or relics from the so-called “Before Times” ­— are classified as “Red Level Artifacts” and reserved for the ruling elite, known as “IT.” By contrast, symbolically abstracted interpretations of life in the silo, drawings for example, circulate freely. The only visual connection to the outside world is mediated through screens, which display a continuous feed transmitted from a single camera, mounted above ground at the exterior of the silo. As the narrative unfolds, doubts about the authenticity of what is shown intensify: Is it truly a real-time feed of the surface, or is it filtered, manipulated — or entirely synthetic? The desire for an inhabitable outside world — a world like the one depicted in colorful illustrations in forbidden books — one that perhaps once was — becomes as illusory and entrapping, as the bleak, dystopian “reality” depicted on screens. SILO rightly presents photography as an instrument of power — one capable of raising hopes and desires, and of rendering the distant tangible in thought.

Photography in the Age of Generative AI - Regarding its role and essence

“Photographs are inextricably linked to, and anchored within, our experienced reality.”63 Yet photography has undergone a profound transformation and been in constant flux — at least since the advent of digital photography, which has drastically simplified retouching and editing, and not least due to the emergence of technologies such as 3D, CGI, and generative AI (which has itself been extensively trained on photographic datasets). Nowadays, almost every photo captured with a smartphone is enhanced using AI-powered features. In the hope of achieving an ostensibly richer and more striking image, we consciously accept that these optimized results often deviate significantly from what has originally been captured.64 Even if we must be aware that this additional “detail” never truly existed, we indulge in the illusion. What, then, does this mean for the “depicted” object or the post-photographic hybrid? At what level can photography — be it as medium, technique, or concept — persist in a world increasingly shaped by digital image generation and optimization?

“For until this day, no representation could assure me of the past of a thing except by intermediaries; but with the photograph, my certainty is immediate: no one in the world can undeceive me. The photograph then becomes a bizarre medium, a new form of hallucination: false on the level of perception, true on the level of time.”65 This classical photography that Roland Barthes refers to, was once regarded as a “reliable” medium for depicting an actual state of affairs. Alongside cinema, it has fundamentally shaped the visual culture of modern society. It plays a crucial role in how we perceive and understand the world around us and influences public opinion — regardless of whether it is used as a form of artistic expression or for the documentation purposes. And although photography creates constructed imagery — be it through framing, the use of props, or the staging and costuming of the pre-existing — it nevertheless (in the narrower sense) remains a confirmation of the past existence of a scenery created though spatial design.

Just as the world within a photographic frame is constructed, so too is our lived reality — shaped by subjective perception and the “power of aesthetic delusion.”66 Similar to how we often describe a cultivated alpine environment as “nature,” even though, like farmland, it is in fact a man-made, cultural landscape and thus only almost nature67, we assume (culturally predisposed) that anything that looks like a photograph is one (that carries some inherent claim to reality).

Surrounded by screens, we constantly surrender to a synthesis of image, text, and language — curated by algorithms and driven by the analysis of personal data. We are enchanted by its rhythm and attuned by atmospheres of sensation, intentionally shaped via technique of impression68 — atmospheres69 that carry embodied, affectively-involving powers.70 Like spectators immersed in a film, we are gripped by joy and sorrow, ease and tension, boredom and excitement… as if doubt itself had been suspended. Willingly, we participate in this felt interplay of “contraction and expansion”71 — a vital rhythm embraced in a voluntary “suspension of disbelief.”72

“We must surrender to the spell of fetishes (in order to feel anything at all), while at the same time we must be capable of directing our own fetishism.”73 — Hartmut Böhme writes in Fetishism and Culture. This demands a high degree of cultural competence — a competence that must be recalibrated in the light of AI — for “demands for identification, through which one surrenders to the flow of presented pleasures, stand in opposition to demands for acts of distancing that limit and frame this very surrender. This contradiction cannot be resolved; it is therefore a paradox.”74 A paradox that confronts us at any time and place in everyday life, as the spatiotemporal boundaries of immersive experiences have become increasingly blurred in an environment where information continuously passes by on screen — reel to reel. When and where to surrender to the spell of fetishized images ­­— and when to resist the seduction of audiovisual stimuli?

“Surrounded by his screens and subject to video control and the discipline of programs, as well as to the rules of interactivity, this new photosensitive being turns into a consenting victim of a progress that amputates his private life, with electro-optical addiction to information more and more alienating him from his sense of self.”75 Digital transformation and evolution of media not only reshape the processes of image production, but profoundly influence our perception of reality. Personalized and algorithmically curated content, flashing past us in an endless feed, makes us increasingly susceptible to manipulation. The repeated exposure to the same perspectives generates a kind of tunnel vision — a self-reinforcing, self-validating cycle that narrows perception, inhibits critical engagement, and, in extreme cases, renders it virtually impossible.

“In forgetting the ‘constructedness’ of images, their apparatus structure, lies the possibility of following the magic of images entirely,”76 and of losing oneself “in the labyrinth of desires and addictions.”77

Learning to critically engage with a world in which photographs and other media representations of the documented moments and the pre-existing intermingle with highly convincing “forgeries” is of profound cultural and political significance. Simulacra blur the essence distilled of information, distorting reality by dissolving the boundary between documentation and construction to the point of indiscernibility. Ultimately, giving rise to a visual world in which authenticity becomes less a matter of verifiable fact than of belief.

This challenge — rooted in the difficulty of verifying the referent in digital space — is also being addressed by major players in press photography, technology development, and news media. One example is the Content Authenticity Initiative, a collaborative effort involving companies such as Fuji, Nikon, Nvidia, Adobe, alongside media organizations like The New York Times and the BBC. Together, they are working on technical solutions to make the origin and integrity of digital images transparent and traceable. Canon and Reuters, for instance, are currently testing blockchain technology in a proof-of-concept pilot that ensures end-to-end traceability throughout the image production chain.78

A responsible and ethical practice in working with generative technologies, alongside philosophical and artistic examination of the evolutionary shifts brought about in the digital age — and the precise use of terminology when employing or discussing media — has become more urgent than ever. Just as we differentiate between oil, acrylic, or brushstroke in painting, it must become second nature to distinguish — and clearly label — analog, digital, hybrid, and fully generated images. Otherwise, we risk losing the ability to classify and contextualize information — and with it, the capacity to make informed decisions.

Is it even possible to preserve an authentic image in a world increasingly saturated with mirages? Or are we, in truth, already dwelling within a clouded, self-enchanted perspective — a subjective illusion no less constructed than the digital and simulated realities we scrutinize? The answer, it seems, lies in our relationship to the referent — that anchoring point which connects the real to the image, reminding us that photography has always been more than mere depiction. A world in which the referent loses its relevance — whether through cognitive overload, apathy, or total immersion in the “flow of presented desires” — becomes a world without facts. A world in which everything must be believed, yet nothing can be believed — or even meaningfully asserted.

“While the trace of the referent may provide a possible foundation for the ontological claim of photography in the narrower sense, it does not account for the ontological effect that photographs in the broader sense (including digital photo-graphs and generative images), (can) exert upon us.”79

Even if generative images (and other virtual worlds) are devoid of any direct referent, they can produce profound real-world effects. They have the power to stir emotions, shape perception, and influence political narratives and cultural discourse. They form memories and, in doing so, condition how we think, feel, and act. In their consequences, they are as real as the perceptible environmental impacts of data centers and increased clickwork. Though, artificial intelligence, unlike human beings, operates solely on probabilities. It does not feel; it knows neither felt-bodily communication nor existential concern. It lacks lived, embodied experience and remains confined to a kind of machinic, computational constellationism80 — devoid of subjective thought, driven not by rationality, but by relational logic.

Paul Cézanne, quoted on the back-cover of the German edition of Virilio’s The Aesthetics of Disappearance: “You‘ve got to hurry up if you still want to see things.”81 A pressing invitation that reverberates in today’s hyper-accelerated visual culture — a culture in which the real and the virtual blur into one another, often thoughtlessly and at even greater speed.

Hallucinating OR Staging the everyday

The act of creating imagery by means of generative AI hinges primarily on the precise formulation of prompts and negative prompts — What do I want? <> What do I not want? — and on triggering the generation process with the simple keystroke combination “Control + Enter”. Beyond linguistic skill, prompting also demands a certain technical sensibility, especially when fine-tuning parameters. Using seed values82, for example, allows one to precisely develop previously formulated prompts or reproduce specific variations of an image. The creative process usually unfolds in loops: adjusting, specifying, executing — again and again — until, confronted with a generated result, one feels the AI has more or less hallucinated something resembling what was imagined. If not a finished image, then at least a foundation for further refinement and approximation.

Much like a tourist, I set out in search of “palm-likeness,” “Ruscha-,” “Los Angeles-,” and “Oberndorfer-likeness”. Thousands of generated images and countless hours of graphic labor — including 3D, but excluding the writing of this text and the layout of this book, totaling to roughly 1700 hours — were required to develop a visually coherent and consistent language for the specific subprojects. Because even with guiding-tools like ControlNet and prompts, the results — especially when viewed as part of a unified series — often seem arbitrary.

The final selection — guided by aesthetic and book-design criteria — was refined through multiple iterative post-processing steps. These involved repeatedly interpolating and enhancing with AI technology, layering, retouching, editing, composing and collaging in Photoshop — again and again — until a print-ready, conceptually coherent series of images and scenes emerged. A body of work in which I had the feeling, that I could rediscover some of what I had imagined prior to my departure.

In contrast, photography often requires nothing more than pressing the shutter at just the right moment in order to create something meaningful within a clearly defined frame. Why is that? For me, the most decisive reason is that an AI — attempting to weave together the information it has been force-fed (in the most Nurembergian fashion) — lacks the most vital capacity for staging atmospheres: the ability to sense them at all. This doesn‘t mean, as previously noted, that generative AI can‘t produce something capable of evoking associations and emotions. But: “The more you know about the spatiality of feelings/sensations, the higher the competence to built spaces of sensation. Creating feelings means staging atmospheres.”83

The act of taking a photograph, in this sense, is — to me — an act of “staging affective atmospheres”84 through the technique of impression — always driven by aesthetic intention directed toward the image itself, and, depending on the concept, anchored in historical and documentary relevance. The negatives, stored in folders, form an indexical archive. At its core, this practice is a technically skilled act of drawing from the pre-existing — from spatially diffused atmospheres of feeling — in a moment of felt-bodily attunement. It is an approach toward a singular, already-past, impressive, emotionally charged, and deeply personal situation85. One does not wander around with the camera permanently at the ready, eye pressed to the viewfinder. Rather, something breaks into one‘s perception — arising from what Schmitz has called the “primitive present”86. And if this something appears worthy of a negative, the camera is raised — and one attempts, in a newly unfolding original moment (the previous has already passed), to feel, reconstruct and capture what one has just been moved and touched by.

This, I believe, is what has long fueled my fascination with analog photography, and why it — despite all the years spent distinguishing it from other media that appropriate its visual language, has remained my principal medium. Its limitations. The conscious act. The immediacy. The rootedness in the encountered. The “relationship between the qualities of the surroundings and the human condition”87, which originates in a specific place, at a specific moment, surrounded by optically and climatically charged stimuli: season, morning, noon, evening.

I regard the enlarged photograph as a “methodical translation (of felt-bodily experienced, spatially extended atmospheres88) into the medium of the visible image”89 — and as yet another approximation. This time, an approximation of my mental image: of how I remember the past situation, that was to be represented and translated, in the present (while working on the print). One which, “in the best case, salvages something of the original moment‘s atmosphere and immersiveness, making it accessible for the viewer through associative resonance”90 (see also “Sense of Balance”91).

Whether something is salvaged or not depends not only on the scenography but also on nuances of color and density — and on the viewer‘s readiness to yield to what is being presented. Just as atmospheres of light possess a felt-bodily perceptible presence that affects us directly on site, so too do half-things, intermodal qualities, and synaesthetic characters — accessible through sensible atmospheres92 — affect us in and through methodical translations. That is precisely why it has always been important for me to be present in the darkroom when photographs are enlarged from a negative. Images serve a purpose: to create the conditions that allow viewers to immerse themselves in their own emotional worlds — guided by the presence of the object (the framed image and the referents embedded in the scene). It is at this moment, that not only the act of photographing, but also the act of viewing a photograph, becomes an event — and “in a certain and indeed decisive sense: performative.”93

Meaningless or Meaningful Noise

Images — including those generated by AI — can, as Eugène Ionesco writes in The Colonel‘s Photograph, become “a journey without moving an inch.” “Irresistible — the crudest means, from which we cannot escape.” And yet, they can just as easily remain “lifeless” if they fail to evoke suggestions of motion, an affective response within us. “The radiant land, in which I had already put down roots, suddenly alien ­— now only a picture in a frame, a lifeless object.”94Meaningful or meaningless noise — which, and this is the crucial point — carries no factual or documentary relevance without its referent. And it is precisely this referential bond that grants classical analog photography — particularly the negative (which Steffen Kammler, in Smoke Signals from the Labyrinth, once defined as “photography in the narrower sense”95) — a unique status. One that collaged, generated, 3D modeled, or digitally constructed visual worlds will never fully attain.

“Photography (once again) outlives itself.”96 For only “the light of the future casts the (photographable) shadows of tomorrow”97 — shadows that, to an AI, are nothing more than silhouettes dissolved into noise, stripped of any past. Pseudo-memories without lived time.

Hallucination of A Reality

In my artistic explorations of the original, reproduction, reconstruction, and generation, the palm tree has emerged as a particularly resonant and versatile motif. It is more than “just a plant.” Like the bunker, the palm itself is a “crudest means” ­— a vessel of emotion and a bearer of layered meanings that extend far beyond the visual. Working with it — with its medial representations, with the palm-likeness as image material — the ambivalence between scientific classification, colonial gaze, romantic idealization, and its role as a pop-cultural icon becomes tangible. A generative AI composes its probabilistic version of a palm tree from the distillation of precisely this cultural bias and visual tradition — a textbook example of the principle: “bias in; bias out”.

The juxtaposition of digital and generative (potential) spaces with the so-called “original” constitutes a call for a kind of media literacy attuned to the aestheticization of reality — as once described by Walter Benjamin. In a world where images are increasingly produced by apparatuses — in the case of AI, by apparatuses in conjunction with neural networks that facilitate imitation — the focus, as Benjamin already outlined in the Age of Mechanical Reproduction, shifts from the experience of the original to the effect of its staging. And in our present moment, that shift continues: toward a hallucinated reality distilled of such stagings: a post-factual “reality” shaped less by material existence than by its medial reproduction and (re)presentation.

In this entanglement of the Real, Symbolic, and Simulated, a threshold space opens up — reminiscent of Michel Foucault’s concept of heterotopia — where the boundaries of image, place, and meaning dissolve into noise and must be renegotiated. The detachment from the thing itself — from the real places and societal structures to which they once referred — emerges from the recursive recomposition of extracted visual codes, accompanied by a dissolution of space and time. To highlight the perceptual distortions in virtual content — content that, at first glance, may mimic the original in appearance and effect yet at second glance diverges fundamentally in informational substance — it seemed appropriate to revisit and further negotiate the palm motifs that Ed Ruscha isolated with white marker in the 1970s. This time, however, by way of synthetic image production — and via processes of interpretation and translation that reveal the medial conditions of the 2020s.

Once again, we find ourselves at a turning point — the age of Artificial Intelligence — where pleasant murmur and distracting noise seem to merge, and where the confrontation with new realities is permanent. “The tasks imposed on the human apparatus of perception in the present age,” as Benjamin noted, “cannot be solved by mere optical experience, i.e., contemplation. They are mastered gradually by tactile reception through habituation. But even the distracted can grow accustomed.”98

Whether — and how — we manage to preserve our gaze for the Real in the age of simulation will depend on our willingness to pause, to attune perception anew, and to learn how to navigate a media landscape that is shifting rapidly and profoundly.

KEEP PALM:

Virtual images, in particular, are vulnerable to manipulations that go far beyond the conscious choice of framing. This is inherent to their digital nature.

But analog photography was never entirely immune to intervention either. A well-known example is Grigori Goldstein’s image of Lenin, in which his rivals Trotsky and Kamenev were retouched out and replaced with stair steps. In this case, manipulation happened directly on print. With negatives or Polaroids, this process is more difficult — yet even then, altered prints could again be reproduced as negatives or polaroids.

The question of the original versus the reproduction has accompanied photography since its beginnings. We have long known it is not an objective medium, and over the decades, we have grown accustomed to this fact — and to engaging with it as such.

In the virtual realm, the stakes are high. Where once single images were deliberately and painstakingly altered, today content farms, bots, and trolls generate mass quantities of content, spreading it across social networks. Their goal: to influence opinions, shape worldviews, and steer political tendencies. They have become a kind of “Deutsche Wochenschau” of our time. But the problem today is not — as it once was — a lack of information, but quite the opposite: an overload of information, a multitude of channels, and echo chambers.

For me, analog photography and its related processes represent a counter-world to the overproduction of digital content and the “efficiency” of generative AI. Perhaps that‘s why it has been experiencing a renaissance for years now — much like the vinyl record — even if, for many, it is less about the media qualities and more about the aesthetic surface, the analog grain and look.

Maybe the special value it holds for me lies in my decades of experience with the analog photographic color darkroom, where I began to understand photography. Or perhaps it stems from my general interest in patient, focused work on content — a practice I still prefer to pursue in darkness rather than under the harsh light of everyday life.

This essay was written for “Palm Trees Are Noise”
Published in German and English.

Terminologies:

Felt-bodily: He describes the term as follows: “When I say ‘felt-bodily’ or ‘embodied’ (leiblich), I am not speaking of the visible or tactile body but of the ‘felt body’ (Leib) as the carrier of such embodied impulses as, for instance, fear, pain, lust, hunger, thirst, disgust, vigor, tiredness and being in the grip of emotions.”99

Situation: Characterized by wholeness (i.e. internal cohesion and differentiation from the outside) as well as by an integrating meaningfulness deriving from states of affairs, programs and problems, and an internal diffuseness of this meaningfulness, which means that the meanings contained in a situation are not altogether singular – in pre- personal experiencing they never are.100

Impressive Situation: Meaningful, polyvalent impressions that are pre-reflexively and holistically perceived, without decom- position into distinct meanings. The majority of our perception and interaction with others and the world is the processing of such impressions, often related to inconspicuous impressive situations.

Current Situation: A situation that occurs right now. Its chronological sequence can be split up into procedural phases.

(Zuständliche) State-like Situation: A situation whose course cannot be followed from moment to moment, but only across extended time spans — for example, language, proce- dures for presenting states of affairs (facts), programs, and problems.

Impressive, State-like Situation: Things appear to us with a typical or even individual character. This character is an impressive, state-like situation.101

(Gefühlsträchtige) Emotionally Laden Situation: When atmospheres emanate from something, the source from which they arise is an impressive situation — a meaningful impression in which a diffuse, internally structured, never fully explicit but suddenly present aura of meanings (i.e., states of affairs, programs, problems) is filled with the atmosphere of a feeling (including mixedfeelings). If an atmosphere radiating from such an impressive, emotionally laden situation affectively involves a person and draws them into its spell, the mediating link that makes such transfer possible is felt-bodily communication.102

Through felt-bodily communication—across the bridge of suggestions of motion and synaesthetic characters these charged atmospheres are transferred into embodied-affective involvement. That‘s where they act upon humans as gripping forces. This spontaneous mechanism can be technically simulated; this forms the basis of aesthetic work in the sense of Gernot Böhme.103

Feelings: According to Schmitz, feelings are not private inner states of the soul, but spatially extended atmospheres. One comes under their spell through bodily feeling. Atmospheres are not cognitively “understood,” but bodily (leiblich) experienced. They are feelings rooted in lived situations and conveyed through synaesthetic qualities of impression, which connect the feeling that constitutes an atmosphere with complementary symbolic meanings. Feeling and meaning form a unity.104

Half-Things (Halbdinge): They modulate the way in which nature appears, tune its vital tone, and give atmospheres their colors and sounds.105

The intertwined tendencies of “contraction” and “expansion” as the basic pair of categories of the felt body: The felt-bodily condition constantly fluctuates between “the maximum of possible contraction (of the mere primitive present) and the maximum of possible expansion (pure, immoderate expansion) in thousand fold alterations without ever entirely reaching one of these poles. The felt-bodily dynamic of contraction and expansion constitutes the human vital drive.106

Atmosphere: See Gernot Böhme‘s book Atmosphären. The Concept of Atmosphere in the Philosophy of Hermann Schmitz, 5. The Thing and its Ecstasies, 6. The Making of Atmospheres, and 7. The Critical Potential of an Aesthetics of Atmospheres.107

Photography in the broader sense: Every digital image is only a secondary image of the result of light exposure. A – no matter how precise – remix of this memory.108

Hermann Schmitz has not been widely translated.
I’ve tried to the best of my ability.

Footnotes:

Footnotes

  1. Markus Oberndorfer, Staging the Ordinary, in: Destined To Return, 2022, p.10. ↩

  2. Odyssey, Twelfth Song, German Gutenberg Edition 16, Lines 184-191. ↩

  3. Wallace Stevens, Of Mere Being, The Palm at the End of the Mind: Selected Poems and a Play, 1971. ↩

  4. Markus Oberndorfer, In Dialog with the Sunset Strip, Reflektor Magazine, 2018, p.12-13 & Markus Oberndorfer, Destined To Return, 2022, p.4. ↩

  5. Ed Ruscha, Every Building On The Sunset Strip, 1966. ↩

  6. Markus Oberndorfer, In Dialog with the Eclipse Blvd, Destined To Return, 2022, p.13. ↩

  7. Jürgen Hasse, Atmosphären der Stadt, Jovis 2012, p.6-8. ↩

  8. Jürgen Hasse, Atmosphären der Stadt, Jovis, p.11. ↩

  9. Cf. Ibid. ↩

  10. Saskia Hebert, Gebaute Welt/Gelebter Raum, p.180. ↩

  11. Effect device that channels sound (e.g., from a keyboard) through a tube to the mouth, shaping it in real-time through mouth movements and capturing it with a microphone. ↩

  12. A high-pitched, modulated synthesizer sound, rooted in P-Funk (Parliament, Funkadelic, Zapp, Ohio Players) and later becoming iconic in West Coast Hip-Hop. ↩

  13. 2Pac, California Love feat. Dr.Dre & Roger Troutman, All Eyez On Me, 1995. ↩

  14. Jared Farmer, Trees in Paradise, 2013, p.377. ↩

  15. Jared Farmer & Nathan Masters: In the 1920s and 1930s, L.A. was busy building its street network. In the age of parking lots, sidewalks, sewers, and power lines, plants with leaves and roots like acacias, eucalyptus, and pepper trees had a bad reputation. In contrast, palms promised a symbiosis with the built world and infrastructure. www.pbssocal.org/shows/lost-la/how-did-l-a-become-a-city-of-palms-and-other-questions-about-californias-trees, 2014. ↩

  16. Cf. Ibid. ↩

  17. Denis Villeneuve, Dune, 2021, 00:43:47-00:44:36. ↩

  18. 2Pac (Makaveli), To Live & Die In L.A., 1996. ↩

  19. Ally J. Levine, L.A.’s palm trees are dying and it’s changing the city’s famous skyline, www.latimes.com/projects/la-me-palm-trees-dying-skyline-los-angeles, 2017. ↩

  20. Esther Margulies, landscape architect and professor at the University of Southern California (USC), comments on the role of palm trees in Los Angeles in the context of conspiracy theories and fire hazards. Victoria Namkung: Conspiracy-laden, fire-prone icons: what will happen to LA’s palm trees? www.theguardian.com/us-news/2025/jan/23/los-angeles-wildfires-palm-trees; ↩

  21. Ally J. Levine, L.A.’s palm trees are dying and it’s changing the city’s famous skyline, 2017, www.latimes.com/projects/la-me-palm-trees-dying-skyline-los-angeles. ↩

  22. Hanna Wüste, Grüne Oasen - Die Rolle der Palmen in der Geschichte von Los Angeles, www.noteverypicture.de/grune-oasen-die-rolle-der-palmen-in-der-geschichte-von-los-angeles; ↩

  23. Jared Farmer, Trees in Paradise, 2013, p.417. ↩

  24. Wallace Stevens, Of Mere Being, The Palm at the End of the Mind: Selected Poems and a Play, 1971. ↩ ↩2 ↩3 ↩4 ↩5

  25. Ed Ruscha, A Few Palm Trees, 1971. 27 Markus Oberndorfer, Generic Palm Trees I-XV, 2024. ↩

  26. Von Martius, Das Buch der Palmen, Taschen, 1853. ↩

  27. Denis Villeneuve, Dune, 2021, 00:43:47-00:44:36. ↩

  28. Flavio Donato, Vilde A. Kveim, Laurenz Salm, Talia Ulmer, Maria Lahr, Steffen Kandler, and Fabia Imhof; www.science.org/doi/10.1126/science.adk0997 ↩

  29. Hermann Schmitz, Situationen und Konstellationen, Wider der Ideologie totaler Vernetzung, p.44-52. ↩

  30. Cf. Benjamin L.W. Sobel, Harvard Journal of Law & Technology, Volume 38, Fall 2024, Elements of Style: Copyright, Similarity, and Generative AI, p.62-64. ↩

  31. Markus Oberndorfer, Diffused Palm Trees I-XV. ↩

  32. Paul Virilio, The Vision Maschine, 1994, p.59. 34 U-Net: www.youtube.com/watch?v=NhdzGfB1q74 ↩

  33. Jared Farmer, Trees in Paradise, 2013, p.417. ↩

  34. Markus Oberndorfer, Generated Palm TreesI-XV. ↩

  35. CLIP: Contrastive Language-Image Pretraining: A neural network that maps visual concepts to language. www.openai.com/index/clip. ↩

  36. CLIP (Prompt), A Bus In The Sand, 2025, p.22. ↩

  37. ControlNet is a group of neural networks that allows for precise artistic and struc- tural control in image generation and enhances standard Stable Diffusion models by including task-specific applications. ↩

  38. Markus Oberndorfer, Staged Palm Trees I-XV, 2024. 40 Google Street View, April 26, 2024. ↩

  39. Google Street View, April 26, 2024, Markus Oberndorfer, Palm Trees Are Noise, p.38. ↩

  40. Of Images: Strategies of Appropriation. Museum of Contemporary Art Basel, Manual No.4, 2015. ↩

  41. Richard Prince photographed images and advertisements, altering the shots by blurring, cropping, enlarging, and grouping them. ↩

  42. Pictures Generation: The term was defined retrospectively by an exhibition at the Metropolitan Museum of Art. This loose group of American artists was active in the 1970s and early 1980s, particularly in New York, and concerned with mass media and image production. ↩

  43. Paul Virilio, The Vision Machine, 1994, p.64. ↩

  44. Steffen Kammler, Rauchzeichen aus dem Labyrinth, Rostocker Phänomenologische Manuskripte, 2009, p.16; Cf. Charles Sanders Peirce, Logik als Untersuchung der Zeichen (1873); Die Kunst des Räsonierens. Kapitel II (1893), Kurze Logik (1895), in: Charles S. Peirce, Semiotische Schriften I. ↩

  45. Jean Baudrillard, Simulacra and Simulation, University of Michigan, 1994. ↩

  46. Cf. Hartmut Böhme, Fetishism and Culture. ↩

  47. Markus Oberndorfer, CONDIFFUSION: A term I developed in the context of this project. It describes a symbiosis of diffusion, confusion, and fusion, reflecting a process or state in which boundaries blur, meanings dissolve or reconfigure, and previously unconnected elements coalesce into something new. ↩

  48. Hermann Schmitz, Situationen und Konstellationen, Wider der Ideologie totaler Vernetzung, 2005, p.186. ↩

  49. Steffen Kammler, Rauchzeichen aus dem Labyrinth, Rostocker Phänomenologische Manuskripte, p.17 & 27. ↩

  50. Michel Foucault: Andere Räume, in: Karlheinz Barck, Peter Gente, Heidi Paris (Hg.) Aisthesis. Wahrnehmung heute oder Perspektiven einer anderen Ästhetik, 1990, p.34-46. ↩

  51. Inge Marszolek, Looking at (bunker)photographs - Thoughts of a historian, in: Markus Oberndorfer, Foukauld - Disappearance, 2012, p.8. ↩

  52. Jean Baudrillard, Simulacra and Simulation, University of Michigan, 1994, p.16. ↩

  53. Image analysis by KI: A Bus In The Sand, 2023. ↩

  54. Markus Oberndorfer, Foukauld, Fotohof, 2012. ↩

  55. CLIP Interrogator, Prompt from 25.8.2023. ↩

  56. BLIP: www.huggingface.co/Salesforce/blip-vqa-base. ↩

  57. Marshall McLuhan, Letter to R.Murray Schafer, 1974, as quoted in: Before Publication, Reto Geiser, Verbi-Voco-Visual, Park Books, p.80. ↩

  58. Markus Oberndorfer, Se Souvenir. Autrement on devient fou, Eigenverlag, 2014, p.70-79. ↩

  59. Hermann Schmitz, System III 5, 1990, p.216-219. ↩

  60. Virginia Woolf, Orlando, Fischer, 2023, p.68. ↩

  61. Graham Yost, SILO, www.imdb.com/de/title/tt14688458 ↩

  62. Michel Foucault, lectures at Collège de France 1978-1979. Michel Senellart (Hg.), 2006, Geschichte der Gouvernementalität II, p.261. ↩

  63. Steffen Kammler, Rauchzeichen aus dem Labyrinth, Rostocker Phänomenologische Manuskripte, p.29. ↩

  64. An illustrative example is AI-assisted photography of the moon. During optimization, it is often enhanced to such an extent that the results differ significantly from what was actually captured. Details are not so much emphasized as they are invented and replaced. ↩

  65. Roland Barthes, Camera Lucida, 1989 (2014), p.115. ↩

  66. Jürgen Hasse, Atmosphären der Stadt, Jovis, p.158. ↩

  67. Markus Oberndorfer, Almost Nature, 2015. ↩

  68. Hermann Schmitz, Was ist Neue Phänomenologie, Ingo Koch Verlag, 2003, p.256. ↩

  69. Jürgen Hasse, Dichte, p.66. ↩

  70. Hermann Schmitz, Was ist Neue Phänomenologie, Ingo Koch Verlag, 2003, p.255-256, p.178. ↩

  71. Hermann Schmitz, Situationen und Konstellationen, Wider der Ideologie totaler Vernetzung, 2005, p.207. ↩

  72. Lenson Piquet, Suspension of Disbelief, 2022. ↩

  73. Cf. Hartmut Böhme, Fetischismus und Kultur, p.480. ↩

  74. Cf. Ibid. p.480. ↩

  75. Paul Virilio, TheGreat Accelerator, p.45. ↩

  76. Cf. Hartmut Böhme, Fetischismus und Kultur, p.478. ↩

  77. Cf.Ibid. p.481. ↩

  78. www.canon.co.uk/press-centre/press-releases/2023/08/reuters-new-proof-of-concept-employs-authentication-system; www.contentauthenticity.org ↩

  79. Steffen Kammler, Rauchzeichen aus dem Labyrinth, Rostocker Phänomenologische Manuskripte, p.28. ↩

  80. While classical constellationism is tied to human experience, “machinic constellationism” could be understood as a system that operates without subjective thinking, purely via proba- bilities, correlations and relational pattern recognition. ↩

  81. Paul Virilio, Ästhetik des Verschwindens, 1986. ↩

  82. A seed value controls the starting state of the AI‘s random generator. It enables the reproducibility of generated content - with the same seed, the computer it can generate the same “pseudo-random sequence”. ↩

  83. Cf. Michael Hauskel- ler, Atmosphären erleben. Philosophische Untersuchungen zur Sinneswahrnehmung, 1995. ↩

  84. Hermann Schmitz, Was ist Neue Phänomenologie, Ingo Koch Verlag, 2003, p.250-255; Anna Blume (Hg.), Zur Phänomenologie der ästhetischen Erfahrung, Hermann Schmitz: Über das Machen von Atmosphären, Verlag Karl Alber, 2005, p.33-35; Jürgen Hasse, Atmosphären der Stadt, p.12. ↩

  85. Hermann Schmitz, Situationen und Konstellationen, Wider der Ideologie totaler Vernetzung, p.52-56; Hermann Schmitz, Hitler in der Geschichte, Bouvier, 1999, p.21-22 ↩

  86. Hermann Schmitz, Situationen und Konstellationen, Wider der Ideologie totaler Vernetzung, p.42-43. ↩

  87. Gernot Böhme, Atmosphäre, Neue Ästhetik, Suhrkamp Verlag, 1995, p.22-23. ↩

  88. Hermann Schmitz, Was ist Neue Phänomenologie, Ingo Koch Verlag, p.244. ↩

  89. Jürgen Hasse, Atmosphären der Stadt, p.39. ↩

  90. Cf. Ibid., p.39. ↩

  91. Markus Oberndorfer, Sense of Balance, Mixed Media Installation, Analog C-Print, 13x test-stripes (framed) + filter-values, Sound, 2015. ↩

  92. Hermann Schmitz, Was ist Neue Phänomenologie, Ingo Koch Verlag, 2003, p.254-255. ↩

  93. Anna Blume (Hg.), Zur Phänomenologie der ästhetischen Erfahrung, Karl Alber, 2005, p.35. ↩

  94. Eugene Ionesco, Die Nashörner, p.55-58, 1994. ↩

  95. Steffen Kammler, Rauchzeichen aus dem Labyrinth, Rostocker Phänomenologische Manuskripte, p.17, p.27. ↩

  96. Markus Oberndorfer, Legends Never Die, 2015. ↩

  97. Madvillain feat. Quasimoto, Shadows of Tomorrow. ↩

  98. Cf. Walter Benjamin, Das Kunstwerk im Zeitalter seiner technischen Repro- duzierbarkeit, p.466-467. ↩

  99. Hermann Schmitz, New Phenomenology. A brief introduction. pp.64–65. ↩

  100. Hermann Schmitz, Situationen und Konstellationen, Wider der Ideologie totaler Vernetzung, p.22, ↩

  101. Cf. Ibid., pp.52-56; and in: Hitler in der Geschichte, 1999, pp.21-22. ↩

  102. Cf. Anna Blume (Hg.), Zur Phänomenologie der ästhetischen Erfahrung, Hermann Schmitz: Über das Machen von Atmosphären, Verlag Karl Alber, 2005, pp.33-35. ↩

  103. Cf. Hermann Schmitz, Was ist Neue Phänomenologie, Ingo Koch, p.250-255. ↩

  104. Jürgen Hasse, Atmosphären der Stadt, p.12. ↩

  105. Cf. Ibid., p.159. ↩

  106. Hermann Schmitz, System der Philosophie. Band 2, Teil 1: Der Leib. Bonn: Bouvier. p.75. ↩

  107. Gernot Böhme, Atmosphäre, pp.28-47. ↩

  108. Cf. Steffen Kammler, Rauchzeichen aus dem Labyrinth, pp.14-15. ↩