Billy Riggs (00:42)
Welcome, welcome, welcome. Welcome back to Rewiring the American Edge. Today, we have a special episode with just me and close friend of mine, a guy who I shared a desk with when I was a PhD student studying design and transportation at UC Berkeley, Stefan Al.

an amazing Dutch designer, now at Hunter College in New York. We're going to talk about from Blueprint to Bytecode, how AI will shape the skylines of the future. But before we do that, I wanted to encourage you to take a look at our episode on Mars.

We had an amazing conversation about the future of where design will go extraterrestrial and how it connects to all the work that we've done on blockchain. It's really interesting. And I think there's a lot of lessons it holds for how we think about this planet and how we can develop more sustainably on earth. And so check it out. And also,

PSA, related to that, if you like the discussions we're having, Vipul and I really want to encourage you to leave us a review, give us stars, do whatever you do to bump our discussion up on whatever podcast streaming service that you use.

This is the way that we expand and that we get our academic message outside of its little bubble. This is one thing that I'm doing to broaden the discussion beyond academia. And for me, it's very important that what we do as professors goes beyond the wall of the professoriate. So if you're listening,

and you're like my mother and you're in the middle part of the United States or if you're listening and you're in Hong Kong or you're in Ireland or you're in the eastern part of Europe, we're grateful for you. Send in your questions. I know we have some questions that we're going to get to as a part of our show. And we're really grateful for those.

But leave us a review and that's just gonna increase our listenership, but it's also gonna give us an opportunity to expand the message and really educate more people on the future of their workforce.

So let us know, send us your feedback, but also leave us a review. We really appreciate it. So without further ado, let's get to it.

Billy Riggs (03:04)
All right, welcome back everybody to Rewiring the American Edge. I am here with professor and architect, Ahl, very good friend of mine. We shared a desk at UC Berkeley. So I have known him for a long, long time. I'm one of your hosts, Billy Riggs. Today we're without Vipul Vyas, my University of San Francisco colleague. But we're diving into a topic with Stefan that's as old as...

human civilization, architectural design, and Stefan is an amazing architect. His work has taken him to Hong Kong where he was a professor at first, but he's designed a lot of buildings. His work has always fascinated me. The architecture of the strip, you should check out his book, The Architecture of Spectacle, his book.

is called, believe, Super Tall, is that right? And The Strip. And he's thought a lot about artificial intelligence and how it's reshaping the way we design and build. Welcome, welcome, welcome, Stefan. Thank you so much for being here. How you doing?

Stefan (04:07)
No, thanks for the

invitation and good to see you.

Billy Riggs (04:10)
Yeah. So, Stefan, I was playful in my introduction, but, I thought we could really talk about, AI and design. You know, we started this to think about the way, that automation and innovation are reshaping planning design and the economy, for the U S and the world.

But before we do that, I thought we could talk about that and the big picture of that to start off with. But I think it'll be great for people to know who they're talking to, to start off with and to know your story. And I said I met you when we were both PhD students at UC Berkeley, but that's just a small sliver of this fascinating Dutch man that I'm talking to, who's actually sitting in New York City. just showed me, before we got on, you showed me

the upper east side and and likely these beautiful tall buildings east of Central Park in New York and for those of you that are watching the stream on YouTube he just showed us that picture so Stefan maybe you can give us a little a little history of of of who you are

Stefan (05:12)
Yeah, of course. So I'm originally from the Netherlands. I grew up in a small town there and I went to study architecture. And coincidentally, my first real job was to work on a design competition for what was then the world's largest TV tower in the South of China in Guangzhou. And to our surprise, we ended up winning this competition. So I found myself designing

what was then the world's largest tower and also designing this very large kind of master plan, you know, roughly 50 hectares and coming to the realization that, that, you know, this was actually pretty cool to think about cities on such a large scale, but also realizing that they're very complex and it's hard. It's a very hard problem to solve when you, when you're dealing with such a large scale.

I found myself drawn to that subject. I lived in Barcelona for some time where I was also an architecture student and also worked professionally. But what was interesting there is I saw how the city had used urban design to really reinvent itself. It kind of re-envisioned its entire waterfront. It hosted the Olympics, created all these beautiful mini plazas.

And I saw how the power of urban planning and urban design really reinvented that city almost. So that drew me to urban planning and did a PhD at Berkeley and became also a researcher and professor focusing on various topics related to more the physical side of urban planning and urban design. So I published on topics related to like climate resilience, sustainable mobility.

the Las Vegas Strip and things like that. So yeah, to this date, I continue in that role. So right now I'm the director of the Urban Planning Program at Hunter College in New York City. And yeah, a lot of our students, they work in this area, they're working for New York City public agencies where they have a chance to implement some of the ideas that we discuss in the classroom. So that's quite satisfying.

So yeah, I continued to practice architecture, although on a smaller scale, and I continued to do research. And you'll find me here teaching in the classrooms of Hunter College.

Billy Riggs (07:23)
All right. Yeah.

So if you're to want to go learn from Stefan and you're interested, he teaches in New York. But really his work when we were students and looking at some of the work he was doing and the history of how buildings got built in the ⁓ Las Vegas strip is really fascinating because these buildings were designed all around spectacle.

creating visual spectacle and just I mean just as an aside for this episode Can you talk a little bit about that because I always thought this was You were one of the first people that that I felt had had brought this to the surface and some of the the stories that you uncovered in your book were really Really interesting. Can you can you share just just one little ⁓ snippet with us?

Stefan (08:07)
Yeah, of course.

Well, when I first went there, I actually went there on a Thanksgiving weekend because when I moved to California, I didn't have any relatives. it was a Thanksgiving holiday break and I decided to go check out Las Vegas. But what surprised me so much about the strip was how technologically advanced it was, know, coming from Europe, of course, where, you know, there's a lot of tradition and buildings are supposed to be contextual.

I was really surprised to be in Las Vegas where they were fully integrated with technology, signage, LED lighting, sound systems. And it was that kind technological aspect that I thought was really new. But then when I delved into the history of the strip, it turned out that this was really part of the very beginning of the strip. If you go to the 1960s, for instance,

The strip had neon lights that were as tall as 200 feet. So that's 20 story buildings. And they were integrating the latest colors of neon, even before they would become commercially available. so it was a laboratory for interesting and technologically sophisticated architecture. So, so yeah, that's.

That drew me to it. And why is that? It's because they're all competing. They're essentially one street. You have the same services being offered, which is kind this resort services and, of course, gambling. But what really differentiated one resort from the next was its branding, its architecture. So that's why there was this intense competition.

by always wanting to do the newest with the latest with the mostest. And it's that race that really made a strip to what it is today. ⁓

Billy Riggs (09:51)
Yeah, and I think

it's important to put in place, you know, I think we always talk about contextual design and what it elevates is the appropriateness sometimes of confrontational design. And I think that really what you see there is confrontational design all over the place.

Stefan (10:10)
Mm-hmm. Yeah. Yeah. You know, one good example is the Luxor, right? So yeah, the Luxor Resort, which is this black pyramid. So up until the Luxor, all the buildings were very, you know, glimmery and very bright. And then suddenly you have this very stealth monolith, this kind of dark pyramid. And by virtue of it being so different from everything else, it kind of attracts attention. So.

So yeah, there is a lot of that there. And of course, a dark pyramid would never be contextual. But in the case of the Las Vegas Strip, it does seem to make sense, right? Because all the buildings are like that.

Billy Riggs (10:48)
Yeah, I think it gives opportunity for lot of other follow on architecture too. It makes me think of lot of the Libeskind buildings that come after that really have these very strong gestures that carry visual interest. And I think that's really important as we think about them, thinking about

the addition to the Contemporary Jewish Museum in San Francisco, where we have these very dramatic contemporary additions that are made to be non-contextual. ⁓

Stefan (11:21)
As I said, there is a Libeskind building on the strip. It is a large mall. So yeah, the star architects have also made an appearance.

Billy Riggs (11:31)
Well, let's pivot to the big picture of now we have software and AI that are confronting the design profession and you know AI has entered every sector and that's what we've been talking about

here and we started with ⁓ transportation and we've been pivoting to what does that mean to the future of work? What does that mean to the future of economies? So architecture and design is no exception. So if you were explaining the state of architecture to someone who hasn't been watching or following it closely right now, I know you've been, I've seen some of your work and playing with

playing with some of the Mid-journey software and things like that. How would you say the state of AI in architecture and design is right now?

Stefan (12:21)
Yeah, so right now for architecture, AIs be used already in several of the workflows. I would say in kind of early visualizations, that's probably the biggest use case right now, creating different design options. So before you would have to model in 3D all the different design options, it may take you one or two hours or

sometimes even two days to model a different design option, depending on how detailed and how fast you are. Then the next step is you need to visualize it. So you need to render it. You may need to do some post-production and Photoshop. So that whole workflow, it's time consuming. So if you can circumvent the entire modeling and rendering and post-production just by using one text prompt, that is truly... ⁓

groundbreaking. So on the visualization side, especially for the early conceptual design, when you're not quite sure yet what it's going to be, I think AI is really quite relevant. Now, the other ways in which I see AI is, for instance, with code, so building code. So if you look at the New York City or the zoning code, so if you look at New York City zoning,

It started off being a modest document of only 11 pages back in 1916. And now it's a document that has more than 4,000 pages and hundreds of definitions and so many loopholes. It's very difficult for any practicing architect to get their heads around that. But with AI, being able to absorb this information and find you, you

relatively good answers. They may not be perfect, but at least give you a ballpark. That really helps that side of the work as well. So yeah, it is being applied. I do think that it's still very heavy on hype though. There is a lot of potential to apply it even more. But until now, we've really seen

most of the applications being focused on visualization, but this is not precise visualization. So this is not kind of 3D model where you can actually have very precise dimensions, precise materials, precise quantities. So we haven't reached that stage yet. And why is that the case? Well, I think one of the main reasons is that the industry is relatively risk averse.

the construction industry is still very much brick and mortar. And it's not like the automobile industry. For instance, it's very capital intensive, heavily regulated, very fragmented into different jurisdictions and different types of companies. So we don't have that kind vertical integration that you may see in other industries.

The other reason is that it may not be interesting enough, at least like the architecture design industry for a lot of the big tech players. Where are we seeing a lot of the AI being applied is where there's probably more money than delivering professional services, like in Hollywood, for instance, or in the financial sector.

If you look at the rendering software, for instance, 3D Studio Max or other types of rendering software, very often it comes from the entertainment industry and then architects start to apply that in their work. So yeah, those are some of the reasons why we're still not where the potential lies. We're definitely lagging behind, but who knows? Maybe in five years, things will be very different.

Billy Riggs (15:44)
So if I was going to summarize, it sounds like AI is kind of like the cool intern. A little more hype than reality, maybe a little overconfident, but kind of sloppy and needs supervision. Maybe has too much sense of humor, but doesn't have the context of the project. But at the end of the day, we did have my

My brother-in-law, on, ⁓ who manages a sheet metal factory, a couple of months ago. And he was talking about how he's, he's using AI to help some of their cost estimation for, some of their raw materials and some of the stuff that they ended up going in fabbing. So I think there's also some things downstream where, there could be some stuff

on the actual construction side, where there's just mechanical processes that happen in the bid process, in the contract side, that AI can just do those processes for people with the RFP side, with the bid side, with the contracting side. And those may be repetitive processes that...

Stefan (16:28)
Yeah.

Yeah.

Billy Riggs (16:52)
You know, architects don't need to spend time doing that. And that's where I'd say, I'm not sure how much we're using AI for that now, but perhaps we should be. And I do agree with you that it takes human oversight and human judgment still matters enormously, but...

We're probably still too early in too much of the early days here to say that this is a reality, but I'm glad to hear that you're saying we can advance this to generate some early ideas. I guess the question I have back to you and this brings up the next segment I wanted to go to is

Stefan (17:30)
Yeah.

Billy Riggs (17:36)
What does that do to the creative process? ⁓ And one of the things I hear frequently, whether or not it's people that are in academia, the movie and film space, music space, in the design space, is that AI will homogenize design process, that it will homogenize the creative process.

Stefan (17:40)
Mm-hmm.

Billy Riggs (17:57)
that will end up with products and buildings that all look the same, like they were spit out by an algorithm, some kind of flavor of really boring vanilla. And we've all kind of seen these suburban developments that looked kind of cookie cutter. ⁓ And I guess...

Stefan (18:05)
Mm-hmm. Yeah.

Yeah.

Yeah.

Billy Riggs (18:19)
the Ronald McDonald McMansion ⁓ that we all have a vision of loops and lollipops and ⁓ little boxes made of tiki-taki. So you've worked around the world, diverse urban landscapes. How do you see AI influencing creativity?

Stefan (18:21)
This is

Well, it's a little bit, there's a strong parallel with the introduction of parametric design. So with kind of the Excel for architectural modeling, was parametric design, various tools like Grasshopper became available maybe like 20 years ago, which allows you to automate modeling. It's not the same as AI, it's still doing the modeling for you, but it's just automating it just like you can put a formula into Excel.

When that first started, at first all the buildings started to look the same because they use the same type of functions. And you can see that too with these AI tools. So what you see is all starts to look the same. Like so many prompts in which people would do, make this Zaha Hadid-style building, or for instance, right? Or make this a true suburban

mansion, and I'm sure you would get into these stereotypes of what these things look like. But what happens as a tool matures, the same for parametric design, is people just become better at using the tool. the tool doesn't do the thinking for that. They just use the tool, just like you're using a calculator.

And it starts to disappear a little bit, the limitations of the typical functions of that tool. And I think we're going to see the same with AI. mean, these models are becoming more more powerful. We'll be able to control them more. I think at first, you're really almost at the whim of whatever goes on behind the scene, these neural

networks and you don't really know how your text prompt or whatever prompt you may give it is going to influence it. But that is changing. Now, how is that going to influence aesthetics? I think that is a truly good question. In this more mature phase of AI, I think one thing we may see, it will probably democratize architectural design to a very large extent.

probably beginning with interior design in which people are just, know, consumers or users are going to re-envision their own homes using AIs, kind of bypassing interior designers to a large degree. So we may see that this really puts design into the hands of, you know, the everyday person.

And of course, it remains the question, when are we going to see kind of the meme style architecture? That probably remains to be seen. But I do think that this will make it easier for people to have a say into the design process rather than be solely dependent on the architect to sketch, to model, to render this for you, which was it takes some craft. takes

you know, lot of time and money to do this. And now everybody can do it, right? Just with a computer. that may impact the aesthetics quite a bit.

Billy Riggs (21:28)
Yeah, think there's, I like the way you're framing it. if I was to make the parallel to the autonomous driving systems that I work on is that, you know, the autonomous systems can optimize the routes. But we still have to frame the rules of driving, you know, and we still have to decide the destinations. And I think that's an interesting parallel to your parametric design.

kind of decision. And if we were to apply that over to art, it's kind of like, we can think of jazz, right? AI can throw out the riffs, but maybe you still have to decide the notes. And so there's a lot of parallels we can draw. So I like the way that you're framing that as an evolution. But maybe we can pivot now to

sustainability because I think there's also been a lot of challenge on emissions as we increase both the processing burden of AI but also the counterpoint on can we use AI to help reduce emissions and improve efficiency. For example, in transportation,

Stefan (22:35)
Ciao

Billy Riggs (22:36)
Can we actually reduce emissions and improve efficiency with AVs? Would there be an architectural equivalent to using AI or using automation in the design process for sustainable design?

Stefan (22:50)
Yeah, for sure, definitely. Just to give you one example, so in skyscraper design, when a building becomes taller than 30, 40 stories, the dominant force is the wind. Now, the wind is very hard to simulate. You need computational fluid dynamics. It's basically a model in which all these little agents, all these little air particles bumping into each other. It's very kind of time intensive to do this in wind tunnels. You can simulate it, but it takes a long time.

⁓ The other way to do it is through AI. If you train an AI to look at lots of kind of CFD images, it can pretty much predict what a particular type of geometry would mean for a tower geometry would mean for kind of the wind. it would help you much faster to iterate your design and to optimize it to kind of reduce the wind load on the building. And that makes...

know, foundation smaller, structure smaller, you save energy, you save materials. So yes, it's very efficient there. We're already seeing it, by the way, not just in the design, but also in the operation of buildings for what we call digital twins, which of course you also know in the urban planning world, but basically what it means, outfit a building with a lot of sensors. Those sensors all go into this virtual model.

And in this virtual model, you simulate all these different types of scenarios. And through AI, you can kind of see which one works the best in terms of keeping everybody comfortable with the least amount of energy for cooling or heating. So those type of solutions in which you apply AI have a lot of promise. Of course, they require more energy to consume or to produce.

Right? So there's that aspect of the energy needed for all these kind of AI type of procedures. But of course, when you apply it to improve sustainability, that's better than using AI to, let's say, generate another cat video or something like that. So yeah, I think it has a lot of potential in many ways.

Just to give you another example, we've become pretty good at reducing the operational carbon for buildings, meaning the energy for heating and cooling and lighting. But the embodied side of construction, which is also massive, the amount of materials that we use in the construction process, is very hard to calculate. You need to know exactly how many units of material are inside a building.

You can calculate this and you can automate it to some extent, but it's still very time consuming to do that. AI could pretty much predict this much faster based on a certain type of design. So I think if we want to tackle these much harder challenges, I think AI could be quite useful. It's almost like using AI for drug discovery, right? So this is using AI to solve a problem that would be

very difficult for humans to do by themselves, even if they have computers or calculators. And I think applying AI to these more complicated problems could be quite beneficial. But unfortunately, we're not doing it enough yet.

Billy Riggs (25:50)
Yeah, think it's also we can say it be done pre and post construction and pre and post occupancy for all these building types. And it doesn't have to be limited to the building. It can be predicting sea level rise. It could be predicting heat waves. And what you said about materials and

Design time also got me thinking about something you referenced earlier about code and permitting. so that triggered this idea that you and I have talked about in past about modular construction and things like that. so, you know, what about affordability and the housing crisis? so, you know, could AI make a really big dent in housing delivery and help address the housing crisis.

Stefan (26:35)
Yeah, theoretically, yes. So we're already seeing some levels of automation. A lot of companies are integrating automation in their construction equipment. We're going to see humanoid robots on the construction sites. We're already seeing drones monitoring progress, robotic dogs. But still, it's very, very limited. And just to give you one.

example. So 3D printing, of course, can really reduce the cost of walls and structures. But it's still only one small piece of the whole construction process, right? Because you have always so many entities that are doing different things. So to get to full automation is still going to take a long time. So we may be able to automate different pieces.

of the construction process and we may be able to generate savings there, but to do the entire project autonomous will be very hard. Now, like you said, I think modular probably has the biggest potential, meaning you could construct large pieces of the building. Typically this is done for housing or hotels in a factory setting. So in a factory setting that's much more controlled,

you could apply automation much easier than on the construction site. So I think that's where I agree with you that we probably have the biggest potential to apply AI and kind of reduce the cost, at least on the construction side of housing. Of course, the construction piece is only one side of housing costs, kind of land value is another.

But I think we can save that discussion for another episode.

Billy Riggs (28:11)
All right. All right. So now that we talked a little bit about the opportunities, what do you think the risks are? every technology has some unintended consequences. And what keeps you up late at night? And what worries you about AI and design?

Stefan (28:28)
Well, one thing is we could become quite sloppy and then you're characterizing the intern. think that's AI as an intern is probably the right way to think about it right now. You still need human oversight. So what that means is that it's just a starting point. You can never use it as the final output. And in construction, that's very much true. So for instance,

AI is being applied to spec writing, right? Which is all the text that's on construction documents. Those are technically legal documents. So to outsource this to AI and to not have proper human oversight, make sure everything is correct. I think that that is a problem. Now in the learning environment, I think it can also lead to people being a little bit lazy also and try not fully understanding

the whole process anymore because this is being done by the AI. And I think that can really short circuit learning. And at the end of the day, for a lot of problems, at least as they relate to architecture, you need someone who has done that task 100 times, 1,000 times to really master it to be able to provide that oversight. So if you don't do the work,

If you send everything away, then how can you ever truly, really have a master ship? So I think the weakest link here may be the human link and us not getting lazy and sloppy and trying not to supplement these tools for hard work.

Billy Riggs (29:56)
Did you hear that students? Do the work, do the work, do the work.

Stefan (30:00)
Yeah, and what we see in education is that maybe you want to have the first year of architecture school completely AI free, almost like a digital detox in which they just teach them the fundamentals without them trying to outsource this to the AI. And I think that may be a good strategy. We're seeing in the academic environment how some universities are going back to using blue books, right?

because it doesn't make sense to have students do reading responses when they can just feed the readings to the AI and it takes them one minute to do their homework. So I think we may want some environments, at least for a period of time in which basic skills are being acquired that are AI free.

Billy Riggs (30:49)
According to Billy and Stefan, your most important weapon in the AI future is a pencil. Always remember, it's a pencil. Okay, so I guess before we wrap, I just wanted to hit you with a quick lightning round with a couple questions. This has been fantastic. first off.

Stefan (30:54)
Hahaha

Billy Riggs (31:07)
For fun, is there a building that's been designed with AI and do you have a favorite?

Stefan (31:12)
I think most new buildings you will see today going up AI would have been used in some part of the process, whether it's generating ideas or whether it's changing the wording in the email from architect to client or whether it's in the post-processing as that happens in Photoshop or other types of software. I would say most new buildings going up today will have some level of AI.

But to truly see a building that's entirely generated by AI, whether on the design side, also construction side, I think we're still a little bit away from that.

Billy Riggs (31:43)
Okay. So do you think there is a planning or design trend that AI should kill?

Stefan (31:50)
I would like to see AI reduce construction waste. So a typical single-family home has about half of the materials get wasted, which is pretty insane when you think about it. So almost 50 % of the waste is construction waste. the typical thing you always see at construction sites is there's big containers. And I think we can be more efficient.

Yeah, so AI is good as a tool for beautiful images, but it would be good to also apply it to other parts of the building process.

Billy Riggs (32:26)
Okay, so what's one human skill that you think AI will never replace?

Stefan (32:32)
I think at the end of the day, we are still people and we like to connect with other people. And once we know we're talking to an AI, I think there's some sort of mystique that disappears. So to be authentic and to be yourself and AI free, I think that will become a virtue at some point.

So yeah, I'm far away from wanting to pick up a book that has been entirely written by an AI, for instance.

Billy Riggs (32:58)
or listen to a song that's been entirely written by AI. Apparently that's the thing these days. But I love that, Stefan. Thank you and thanks for humoring me with a couple rapid fire questions there. Any final words from you?

Stefan (33:01)
you

Well, of course, we're all using AI. There's no way around it. you know, for beginning architects and urban planners, you you need to learn it. And, you know, the faster you can integrate it, the better. But you should never let that... Never not do the hard work yourself as well. So the only way to really apply properly is to do the hard work and to understand what's going on. So it should never short-circuit your learning.

Billy Riggs (33:40)
And that's good advice to close us out. This field is rapidly changing, and you should never stop learning. Remember, technology should serve people, not the other way around. Stefan, it has been a pleasure. Thank you for joining us on Rewiring the American Edge. We will see you next time.

Stefan (34:01)
Thanks again, great to see you.