This podcast is completely AI-generated with voices that don’t exist, created by real people talking about real stuff.

In the second episode of AI Wednesdays, Alpha and India explore the commonly asked question: "Why doesn’t Midjourney understand specific dimensions or the number of storeys in a building?" We take a deep dive into the underlying technology behind Midjourney—diffusion models—and explain how they work as image synthesizers rather than intelligent tools with a true understanding of architectural prompts.

We clarify the differences between diffusion models and transformer-based AI models, such as GPT, emphasizing the strengths and limitations of each. While Midjourney excels at creative and conceptual visualization, it falls short when it comes to delivering exact architectural precision, such as specific measurements or accurate counts of floors in a building. We discuss how architects can best leverage this tool for inspiration and integrate it into a broader design process by using CAD software for technical precision.

Tune in to discover how to make the most of Midjourney's capabilities, set realistic expectations, and avoid common frustrations when using AI tools in architectural workflows.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit rbdsailab.substack.com