as deformed fragments get informed

The contemporary progression of urbanisation is driven by systems of order and predictability, tied to practices of uncertainty minimisation. As architecture is inseparable from finance, risk mitigation such as due diligence parallels Machine Learning algorithms that minimise information entropy. This project hypothesises that ML-based decision making will further inform urban architecture toward predictable, or "safer" forms. Inspired by artistic practices embracing disorder, this proposal explores how architectural production might incorporate uncertainty to counteract this tendency.

Through minimisation of entropy and exclusion of ambiguity, information gives the city a form that is increasingly categorised and predictable, while the value of deformation is overlooked in the process of urbanisation. The project critiques this adversarial effect of information and proposes a reassembly pipeline to subvert the parallel condition between financial cycles and machine learning processes. The pipeline assigns various traits to the unused equipments and its parts — weight, locations, images, point clouds and material characteristics data. By relentlessly informing these "deformed" and providing a public interface to query based on the traits, it realises the urban assemblage through the redistribution of the once-discarded materials. This new industry of obsessive capture accumulates data about discarded fragments, allowing new characters to emerge. The captured data is distributed redundantly across London as a repository network, archiving the progression of urbanisation through the lens of the deformed.