We believe city-making should not be a human-centered practice and that there is a lot to learn from topological formations and natural systems mediated by biological intelligence. In this series, we explore Artificial Intelligence as a form of co-authorship in large-scale projects. In Landscaping, mountainous terrain is usually the opposite of highly urbanized centers. Due to its irregular terrain, high intervention costs, compact soil mass its normally one of the most expensive urban projects to develop. How does this intertwine if we think about a mountainous land in one of the least urbanized areas of Argentina? This is how Salta drew our attention as a project site research. We trained an artificial intelligence network [GAN] with a series of aerial images of the region of Salta, Argentina and we re-scripted the code that relates them to understand what logic emerges from these topological formations and how they could affect the construction of urban sites.