“All Compressed and Rendered with a Pathetic Delicacy That Astounds the Eye”: Midjourney Renders Ambergris as Constantinople

Abstract

Abstract Text‐to‐image software have become widely accessible over the last few years, and the resolution of the images generated by these machine‐learning software has increased to such a degree that the images attract more and more public attention. Midjourney is one of the few available AI tools that provides images to textual prompts. With early beta access granted, I have put the abilities and biases of the tool to the test by prompting it to blend a fictional city (i.e., Ambergris, created by Jeff VanderMeer in 1993 first for a novella) with a historical one (i.e., Constantinople). The ontological distance of these two cities would not pose a problem, for the textual and visual data for both are vastly available on the internet, which data the software utilizes to generate images. The decision to blend these two is based on Ambergris's quality of being a city influenced by the history of Constantinople, yet diverging from it in other aspects. The software ended up generating images, where it employed various periods and urban features of both cities, all the while conforming to depiction styles that can be associated with both Ambergris and Constantinople.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 12 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.