The Market in the Model: Latent Diffusion as Neural Economy
This paper critiques generative image models, specifically latent diffusion models, by examining their underlying mechanisms and the problems they were designed to solve for computer vision engineers. It argues that these models function as a "neural economy," abstracting social communication into quantifiable vectors for sale. The analysis traces the training and generation pipelines, revealing how each operation reinforces platform and attention economy logics, and warns that critiques focused solely on copyright may inadvertently uphold the model's fetishistic nature. Instead, the paper advocates for centering social exchange in critiques. AI
IMPACT Examines the economic and social implications of generative models, urging a shift in critique focus from copyright to social exchange.