Skip to main content
ADS12: Melee

Matthew Rooney

Matt is a designer from London whose interests lie at the intersection between spatial and art practices. His work addresses the implications of our ever evolving political and technological climates, considering the role of design as a tool for speculative change. His first year project at RCA developed a data assisted system of local planning, seeking to position how both behaviour and the built environment might be effected through augmented democratic proceedings. Completed in this same year, his master’s dissertation undertook a historical study of the Berlin street and its convulsions following the fall of the Iron Curtain, observing the battles between capital and hedonism to define new national and urban identities.

Prior to the RCA, Matt completed his undergraduate study at Newcastle University and subsequently worked in architectural practices in both London and Berlin. Simultaneously he has developed a personal practice in illustration, applying his strong passion for image making for numerous commissions including children’s books, articles and websites.


Chthonic Follies looks to the intermingled territories of resource extraction and seeks to understand the increasing role of machine intelligence in the mining and remote accumulation of planetary wealth. Rethinking the extraction site to include the expanded constellation of financial and logistical flows that underpin the subsumption of nature for capital, the research observes the potential for artificial intelligence to extend the market’s gaze over ever greater territory, material and life. As an apparatus of remote control, planetary imaging hence performs the seeking and visualisation of greater opportunities to extract and commodify. Through manipulation of this 'gaze', the project seeks to develop acts of resistance and agency within the often intangible networks inherent to global systems of extraction and exchange.

As the global hub for the trade of subsurface "chthonic" resources, the City of London and its myriad spaces of capital accumulation represent a zenith of planetary control within the capitalist world system. The distribution of the city’s banks, hedge funds, fossil fuel speculators and mining company headquarters portray the uneven geographies of profit and the perpetuation of neo-imperial hierarchies. Through this relationship is uniquely observed the site of the pit, a feature inherent to both territory and city. Beyond the spatiality of one pit, the open-pit of the extraction site lies another, the trading pit, an architectural form intrinsic to the price fixing of natural assemblages. The duality of these sites, of plunder and profit are challenged through the Generative Occupational Network, a speculative platform seeking to explore the potential of digital space to subvert and manipulate the City of London’s institutions and architectures of power.

Challenging extraction capitalism's Faustian pacts with the planet, Chthonic Follies mechanises a machine learning toolset to design from the image of both financial and commodified landscapes. In doing so an anti-architectural language is formed that draws across three scales of commodity, territory and city.

Intermingled Scales; Commodity, Territory & City — A GAN is trained upon image sets of traded chthonic resources [ores, minerals, geological samples], aerial images of mined landscapes [open-pits, slag heaps, refineries, tailings sites] and the facades of financial institutions and corporations associated to mining in the City of London.
Sites of Extraction; City and Territory
Sites of Extraction; City and Territory — The sites do not exist. They are predictions, understandings of territory and city produced to the visual constructs of the market. Opposed are the images of extraction and accumulation within planetary and urban contexts.

Generative adversarial networks [GANs] are algorithmic architectures that use two neural networks, pitting one against the other in order to generate new, synthetic instances of data that can pass for real data. Enabled is the extraction of visuality of a seeded image source allowing the generation of infinite novel iterations of image. When applied to the image of extraction, the surface aesthetics across the scales of market operation may be understood.


Generative Adversarial Network
GAN_Facades {Film}
GAN_Columns — The generative network informs the creation of a catalogue of novel extracted columns, the machine gaze’s interpretation of neo-classicism, symbolic of the aesthetics of wealth accumulation and institutional power.
Folly Facade — A facade that is all facades and simultaneously none. The folly facade alludes to a deconstructed architecture of London's financial institutions. Here proposed as a signifier to ownership upon the extracted territory, a relationship made visible between the landscape and urban-based commodity circulation.

The facade is identified as a key signifier to the hierarchies of the city and a statement of surplus value and possession. Harvesting an image set from London’s financial districts breaks down the image of wealth accumulation to its key aesthetic attributes when understood through the GAN. The reformation of extracted features gives way to the latent space of computer vision - an inference of visual pattern that exists somewhere between the real and the predicted.


Generative adversarial network & scraped webpages
Algorithmic Topography — GAN generated pit relief
Algorithmic Pits — A series of pits are generated through the application of the GAN image of the extracted territory to the floorplan of the London Metal Exchange [LME]. In doing so the unifying landscape and architectural feature of the pit is formed from hundreds of seeded sites, directly linked to the city's markets.
Generative Pits {56 Leadenhall floorplan}
Generative Pits {56 Leadenhall floorplan} — Machine learning places the topography of the extracted landscape within the floorplan of the Exchange producing altered material, height and topographical states.

The Generative Occupational Network represents a strategy of temporary digital occupation within the City of London. Employing the apparatus of extractive capital, the network makes use of machine learning as a means to deconstruct and undo the physical embodiment of the city’s institutions - offering a site of contestation capable of shifting and expanding indefinitely. 

The target site for this occupation, the London Metal Exchange at 56 Leadenhall Street - is the city's last remaining open outcry trading pit and an institution inherently bound to the commodification of planetary assemblages.


Machine Learning 2D & 3D StyleTransfer
Grafted Territories
Grafted Territories — Landscape features generated from the seeded floor plan of the London Metal Exchange and extracted territories.
Grafted Territories — Interpretation of the Exchange's walls through the application of 2D & 3D neural style transfer of landscape to building.

Grafted Occupation

Machine learning places the topography of the extracted landscape within the building producing altered material, height and topographical states. The generated pits carry the legacy of their seeded sites, formed from a thousand mined territories directly linked to London’s markets. 

The grafting continues to the walls of the exchange - the platform’s plan in perpetual flux as the occupation moves through zones of the existing building. A digital landscape is grown, new forms unite the exchange with subsumed territories. The interpretation of walls as landscape conduct a dematerialisation of form.


2D & 3D Neural Style Transfer
Temporary Autonomous Forum — The network begins the initial geometry is seeded from the pit of the Exchange, locating a virtual chamber of participation within its confines. User-created the zone is directed towards organisation and engagement between activists.
The Community Garden — the Generative Network creates a digital garden within the entrance lobby of the exchange, as measure of engagement the maintenance of virtual plant-life conveys levels activity within the platform.
The Board Room — {generative pits & landscape} seeded from the ground floor offices of the exchange a board room is formed as disagreement emerges within the network. The zone is begun as a splinter site within the movement directed towards conflict resolution.
Platform Vandalism — As levels of engagement drop, the platform witnesses acts of individual vandalism. The pit grows to fill the entire ground floor plan whilst elements suffer from neglect. The network moves on.

The Generative Occupational Network; Scenarios of Occupation

The Generative Occupational Network culminates in the formation of a temporary autonomous zone within the London Metal Exchange. Assembling the extracted fragments of built and territorial image 4 scenarios are presented as manifestation of the occupation. Each immersive environment providing a unique temporal zone of inhabitation hosting virtual behaviours directed to the platform’s counter hegemonic intent.

In hosting this occupation - the intention lies to highlight the extractive rituals of London’s markets, manifesting the city’s role within flows inherent to acts of possession and accumulation within the capitalist world system. By doing so the project asks what possibilities exist to manipulate the technological and financial systems that underpin the subsumption of nature for capital? Working to develop alternative forums of participation within complex structures of power present in the city, the nuances of virtual space offer an opportunity to fulfil individual desires to resist and disrupt often intangible structures of power.


Generative pits & landscape