Dialectic of simulation
A recent 404 Media piece by Emanuel Maiberg observed that image searches on Google now return lots of generated images rather than archival results — an increasingly unstable distinction as the archive indexed by the search engine comes to consist more and more of generated images. So not only can an image search query can be treated as a prompt to a generative model, which would create spurious images on the spot, but those images can be ranked (which ones were clicked on?), stored, and later regurgitated to accommodate similar queries, so that the most popular searches will be able to draw from the deepest well of previously generated and feedback-sorted images.
In these kinds of search results, the distinction between a query and prompt is being deliberately collapsed. Requests for information are treated as requests for simulation, as if simulations were just better or purer forms of information — ideal realizations rather than contingent and imperfect examples of whatever concept was requested. What is the use of history if not as a dataset to produce a cleaner and clearer instantiation of the static world of forms? What is the point of an archive if it is not feedstock for a model that can refine it into pure conceptuality?
Keep reading with a 7-day free trial
Subscribe to Internal exile to keep reading this post and get 7 days of free access to the full post archives.