From Window to Hyperobject

In a unified linear perspective, the image's surface functions as a window onto the world. Erwin Panofsky described it as the result of a geometric cut through the viewing pyramid of a monocular observer fixed at a single point in space. Importantly a linear perspective in renaissance painting was not simply a technical innovation, but a symbolic form implying a specific worldview Panofsky, 2012. This metaphor framed linear perspective as rational, neutral, and universal. Emmanuel Alloa counters that such symbolic coherence may never have existed – arguing instead that dominant regimes of seeing merely conceal their character behind claims of objectivity Alloa, 2015.

The epistemic fiction of the objective gaze was inherited and amplified by photography. Compared to human perception, photography was seen as more mechanical and thus presumed more truthful. Yet photography did not emerge as a neutral record-keeping technology. From the beginning, it was embedded in institutional, scientific, aesthetic and colonial systems. It actively framed, classified, and controlled – organizing bodies, landscapes, and knowledge through the lens of authority. What appeared as visual neutrality was a construct shaped by the apparatuses of power. As Dewdney argues, photography became a socio-technical system that not only reflected dominant modes of seeing, but actively reproduced and naturalized them Dewdney, 2021.

Harun Farocki coined in 2003 a new category of images emerging from military, industrial, and surveillance contexts: operational images. Farocki, 2003 These images are not intended for human interpretation, but for action within automated systems: to target, track, or measure. They are often still lens-based, but their function has shifted from visual representation to operational execution: the photograph no longer serves as a record of human experience, but as a logistical element in a closed feedback loop of operational decision-making, such as drone strikes, automated manufacturing, or biometric surveillance.

With the rise of networked images in recent decades, the representational logic of photography has collapsed. Photographic meaning is no longer confined to the frame, but distributed across metadata, algorithmic modulation, platform logic, and patterns of circulation Steyerl, 2003. Building on this shift from representation to function, Hölzl and Marie argue in their theory of the Softimage, the photograph has ceased to function as a sign pointing to a referent; it now operates as a signal within a networked system Hölzl & Marie, 2015.

The rise of Computer-generated Imagery (CGI) pipelines, computational photography, and Artificial Intelligence (AI) image generation has further severed the photographic image from its traditional indexical base. Even when light and lens remain part of the process, the resulting visuals are increasingly shaped by synthetic logics. The loss of indexicality does not simply destabilize realism – it transforms the very purpose of photography. What is now encountered are not traditional photographs, but photographic images: visual constructs that resemble photography and evoke familiar empathic responses, yet may be entirely generated through computational processes. No longer discrete representations or static objects, images today function as systems – dynamic, procedural, and deeply embedded within computational infrastructures.

As photography becomes procedural – defined by algorithmic operations rather than manual capture – so too do its aesthetics. No longer only shaped by composition, light, or narrative, images are now evaluated through their performance within networked systems: frictionless clarity, entertaining immediacy, seamless resolution, and high-speed compression – even in the degraded form of the poor image. They are optimized for attention and virality, designed to circulate rapidly. Their visual language is algorithmic. What is now popularly called a “photograph” is often a product of prediction, reconstruction, and enhancement. Its smooth appearance is not a return to realism, but a simulation calibrated for maximum legibility.

The operational code of these images is not external – it’s visible in their aesthetic. From compression algorithms to beautification filters, from face detection to exposure weighting, from perspective control to complex computer vision processes stitching multi-lens cameras to a seamless single frame; computational processes shape not only what the image shows, but how it appears. These processes produce a visual style marked by seamlessness, brightness, and clarity – qualities engineered for both machine readability and user engagement. What is perceived as aesthetic choice is detached from physical limitations and an artifact of optimization Hölzl & Marie, 2015 Keller et al., 2021

As Hölzl and Marie argue in Softimage, these images internalize their operational code, both visually and functionally. Their aesthetic is inseparable from the computational systems that generate, process, and distribute them Hölzl & Marie, 2015. Hito Steyerl’s concept of the poor image – low-resolution, degraded, endlessly copied – was in my understanding an early symptom of this shift Steyerl, 2009. It once resisted the high-resolution economy of visual capital by circulating through the informal networks of digital culture. But as platforms grew more integrated and machines became more powerful, even the poor image was drawn into the logic of optimization: upgraded, accelerated, and rendered hyperreal.

This transformation echoes Flusser’s warning: the photographic apparatus operates according to a program that determines what can be seen, produced, or imagined. Humans do not control the camera; they act as functionaries, enacting decisions already structured by the photographic program. Photographic agency has never resided in ownership of a camera, but in the capacity to resist or redefine the limits imposed by this program Flusser, 2000. In the post-photographic condition, the program is embedded in automated, networked infrastructures that govern how images are generated, selected, and circulated. The photographer’s role is increasingly marginal, absorbed into algorithmic systems of optimization and platform governance. ¹ Pressing the shutter becomes a symbolic gesture within a sequence already predetermined by the program. As Flusser provocatively states, “it’s not the world out there that is real, nor is the concept within the camera's program – only the photograph is real” Flusser, 2000, p. 37. In this condition, understanding photography requires more than interpreting images; it demands systemic thinking about the infrastructures that shape visual experience.

Photography's quality is the unique ability to distill reality into seemingly simple forms. This is both its strength and its limitation. The ability to distill makes photography ideal and very efficient for communication and advertisement. Though today's issues, are very often highly intertwined hypercomplex objects. Morton, 2013 And in a same logic, Photography no longer resides in a single frame, place, or moment: it has become a surrounding condition, diffused across satellites, networks, sensors, and platforms.

Varoufakis frames these platforms as sovereign powers in a new digital order: entities that rule over communication channels with the authority of feudal lords, profiting not by selling services but by owning the infrastructures that others must use to be seen and heard Varoufakis, 2024. ↩