Neural Network Generated Zines (2019), by Everest Pipkin
In this project Everest made two zines (printed and digitized) using an "intensive upscaling algorithm" that reinscribed details in highly compressed thumbnails. Essentially, this project employs a powerful neural network (if I understood it correctly) to add more detail to an image whose format was intentionally designed to be a reduced (but visually effective) version of the original image. I'm interested in how this work touches on Hito Steyerl's discussion of the politics of 'poor images'. Steyerl argues in Defense of the Poor Image that there is a hierarchy of images based on their resolution, and that resolution has been "fetishized as if its lack amounted to the castration of the author."
We could simplify, then, and say that the upscaling algorithm is part of the redemptive 'agenda' of the empire of resolution. Yet, Everest's project complicates that assumption by showing us the limits of such algorithms. While we know that the 'poor images' in the zine have gone through a rigorous process of enrichment, they aren't particularly remarkable--they still fail as (en)riched images. Yet, what we are given in the zine is a kind of unheroic visual document of these two systems trying to compromise their difference--and we are asked to pay attention to that. In this project we already see tension between disorder (the reality of compressed, low res images) and order (the desire for high res, rich images). But with regards to how order and disorder function in this work effective complexity, I see this work hovering around simple-disordered.