I asked Gemini to design a futuristic smart building in a tropical city like Kuala Lumpur, including AI features and sustainability. Then, I asked it to redesign the building using a low budget and simple construction methods.
Output: Gemini generated a futuristic design with advanced features such as adaptive facades and vertical greenery. The image was visually strong but complex. After refining the prompt, it produced a simpler and more realistic design with fewer complex elements, making it more suitable for real construction
Reflection: This shows how prompts directly affect AI results. The first design focused on creativity, while the second focused on practicality.
The second version is more realistic, but still lacks detailed technical information such as materials and construction systems.
I asked about passive design strategies for tropical buildings and compared natural and mechanical ventilation in terms of sustainability.
Output: Perplexity provided clear explanations supported by sources. It explained key strategies such as orientation, shading, and natural ventilation. It also showed that natural ventilation is more sustainable than mechanical systems due to lower energy use.
Reflection: Perplexity is strong in research and provides reliable information with references. However, the results are general and not directly linked to a specific design. Compared to Gemini, it lacks visual output, but it is more reliable for factual information. The information can be applied to improve building design, especially in making it more climate-responsive.
I asked ChatGPT to create a concept proposal for a smart and sustainable public building in a tropical climate. Then, I asked it to simplify the concept to make it more practical and affordable
Output: ChatGPT generated a structured concept for a smart tropical community hub. The first version was detailed but idealistic. The second version simplified the design, focusing on low-cost solutions and passive strategies.
Reflection: This demonstrates how changing prompts can improve racticality. The simplified version is more realistic and focuses on passive design instead of complex technology.