Yasmine Ayman
student
Ain Shams University
Egypt
Urban Design and Landscape
ATHAR is a sensory landscape and wellness destination rooted in the unique natural and cultural identity of Siwa Oasis. The project was inspired by the need to… more
Yi Wang
advisor
Washington University in St. Louis
United States of America
Yi Wang is a licensed architect based in New York and the founder of Yione Studio, an award-winning… more
The project show us the possibilities and potential of leveraging a designer’s capability through cutting-edge technology. It’s an intriguing site and an ambitious scope of work. What impresses me the most is the page 9 in handbook, where the designer translates the site’s unique natural topography, the moving dunes, into an architectural volume language.
My overall suggestion is to establish a performance-driven strategy for using AI image-generation technology rather than a visual-driven one. An architect wears multiple hats: designer, engineer, negotiator, consultant, moderator, and more. Knowing when to switch hats increases the credibility of your works. I’d say the first step is to differentiate which drawings are suited to AI-generation and which aren’t. Quantitative analyses like Sun&shading study should rely on real solar simulation (software like ladybug for rhino, or Insight for revit) or physical models to test outcomes.
Second, prioritize the QA/QC process within your established workflow. For example, the index map in brochure doesn’t show the site is surrounded by water on three sides, which could mislead visitors. Current Chatgpt image generation tools also have real limits on rendering quality. After two rounds of editing, texture start to look cheesy. And this gets even worse for plant rendering specifically. You probably have noticed that already.
19.07.2026