The Paradoxical Potential of the Post-Lenticular Baroque

Architecture representation has always been closely linked to perspectival and lenticular conditions, whether it be in drawing, image, or model. Yet, Vision-Language Models (VLMs), and emerging world models, are operating outside of these traditional spatial, aesthetic, and representational associations. Although initially trained on visual datasets to simulate physical environments, these multimodal models operate in fundamentally non-visual, translating images into operational coordinates, data points, and machine-readable instructions for autonomous actors.

However, when instructed to render representational or spatial imagery, the statistical nature of these models become apparent, often producing imagery that lacks a true understanding of consistent optical parallax, rigorous lens geometry, or the physical constraints of lens-based media. While these anomalies may be considered inconsequential to autonomous machines and systems, subjecting these images to the lenticular measurement practices of metrology, such as photogrammetry, exposes the distorted spatial territories of these models and their post-lenticular understandings of the physical environment.

This talk will unpack the paradoxical potential of these artificial spatial conditions, as a territory of operational and post-lenticular intrigue, through a design-led research project, The Post-Lenticular Baroque.