#351 - ANDREW KUDLESS, Founder of Matsys on A.I. in Architecture
SUMMARY
This week David and Marina of FAME Architecture & Design are joined by designer and educator Andrew Kudless, Founder of Matsys, to discuss the fundamentals of generative Artificial Intelligence (A.I.); how he teaches Midjourney; it’s potential role in the architecture profession; the evolving role of an architect; how students can make the most of the tools; and more. Enjoy!
ABOUT ANDREW
Andrew Kudless is a designer based in Houston, Texas where he is the Bill Kendall Memorial Endowed Professor at the University of Houston’s Hines College of Architecture Design as well as the Director of the Advanced Media Technology Lab. In 2004, he founded Matsys, a design studio exploring the emergent relationships between architecture, engineering, biology, and computation. The work of Matsys has been exhibited internationally and is in the permanent collections of the San Francisco Museum of Modern Art, the Centre Pompidou in Paris, and the FRAC Centre in Orleans, France. His work on Confluence Park has won a number of awards including a 2019 AIA National Honor Award. In 2019, he became the first American designer to contribute to Louis Vuitton’s Objets Nomades collection. He holds a Master of Arts in Emergent Technologies and Design from the Architectural Association and a Master of Architecture from Tulane University.
TIMESTAMPS
(00:00) Teaching.
“The students are paying for an education… Anyone can learn anything… You could learn almost anything on Youtube but it’s maybe not the best way to learn and it’s not a very focused way to learn.” (04:14)
(05:06) Introduction to Midjourney and AI.
“It’s going to take a lot of resistance… because A.I. is doing some heavy lifting, we then just fill the space that is created with more work for ourselves. I hope [instead[ we take the space and go to the movies or go outside and go on a walk or whatever it is.” (40:17)
“It’s a huge problem that we’re going to have to deal with… [in Midjourney] there’s a model called an ‘aesthetic predictor’ It’s essentially a model that has been trained on a bunch of images that humans have judged from 0-10 [based on what is visually pleasing]. Midjourney doesn’t give you just any image when you enter a prompt, it’s restricting the types of images that come back to you that are probably between a 7.5-10. So it’s not letting you look at any ugly images. So it’s incredibly biased toward ‘beautiful’ often in a Western sense of what beauty is as well.” (44:27)
(47:55) Cost to use A.I. tools.
(54:13) Generating Options.
“This technology is so important that we shouldn’t leave it in the hands of Microsoft, Google, Apple, and Facebook because they're putting their hand on the scale in a variety of ways like the programmers are determining the weights on these models of what you can and can't do. And I don't know if you should trust all these companies to make those decisions.” (01:01:20)
(01:13:47) Teaching Midjourney.
“The reason I think that the architects will still be valuable is that just because you can make a beautiful image doesn't mean that that's architecture.” (01:19:27)
