#521 - Does Digital Technology and A.I. Improve Architecture?

Technology and architecture discussion

0:00This is The Second Studio, hosted by the Architecture and Design Office fame. My name is David Lee, and with me is Marina Bordaroni. The two of us are architects, partners at the office, and hosts of the show. And today it's the two of us, and we're going to talk about technology and architecture, maybe mostly AI and architecture. Maybe, we'll see. And what happened was recently I participated in a digital conference, the DATM conference by Gable Media, G-A-B-L Media, and they're a podcast network, which we're not a part of, but they're good people, and they have a lot of good shows, and they invited me to be a speaker of sorts, a panelist, speaker, whatever, in this digital conference.

0:42And our topic in our session was roughly what I just described. It was myself representing the architecture of people for whatever reason. I don't know why I would be chosen for that. And then there was someone that was more on the contractor side of things, and so we had a session. I believe all that stuff was online, so if you Google it, you can find it. But I will say, I was deathly ill. I became really ill, like, the day before the event. Deathly ill. And that morning, I took all kinds of drugs to be somewhat functional, and fortunately, I think, like, the adrenaline kind of kicked in.

1:21You mean, like, Advil, Tylenol, syrup? Mm-hmm, yeah. Tough drugs, yeah. Yeah, yeah. And hardcore, snorting, you know, the Advil. So, and I think I made it through, but so if you watch the video, and I look very sweaty, or my screen turns off, and I mute myself, it's because I'm blowing my nose. I was in poor condition, so that's also to say, I don't know if it made any sense at all. So we'll call this a take two of that. No, but of course, these topics have come to the surface because of AI recently, which is ever-evolving.

1:52And so we're just going to have a chat about it, and mostly, we're going to put Marina in the hot seat, as usual. Sponsors. Today's episode is brought to you by Acelab. You know what Revit did for building geometry? When a wall moved, every plan, section, elevation, and schedule updated with it. Acelab's Material Hub is doing the same thing, but for materials. So in your current workflow, when a material changes, you're probably manually chasing that update through your spec, finished schedule, keynotes, and sustainability report every single time.

2:23Acelab's Material Hub replaces all of that with one record. Make a selection, and every single document updates automatically, so there's no chasing, confusion, or spreadsheet graveyard. With over 250,000 products, over 10,000 brands, Acelab has real verified sustainability data, not just marketing claims. Check it out by tapping the Acelab link in our show notes. This episode is brought to you by Chaos. They make award-winning design and visualization tools, and they're really pushing the boundaries with AI. Now, AI is shaking up the architecture and design world, and Chaos is taking a responsible approach.

2:58Their tools are built to help architects, designers, and visualization specialists to work faster and explore more ideas without replacing creativity. Basically, Chaos wants AI to make your imagination stronger, not take over. So you get to design with more freedom, create stunning visuals, and stay in control of your work. If you want to see what that looks like, tap the Chaos link in our show notes.

Has digital technology improved architects

3:20Okay, so technology and architecture can be talked about in a myriad of different ways. It could be about building science. It could be about the production of architecture, like drafting. It could be about clients using that technology now, especially with AI. It could be about renderings and presentation. It could be a lot of different stuff. It could be home technology, too. Yeah, home technology. I think the question I had posed was how and whether or not technology, let's say digital technology more specifically,

3:50if it's improved the quality of architects over the years. Well, I would say probably not. And what makes you say that? Because there is great architecture that was done before without technology. And I could not say that by driving around town, I've seen better architecture that's been recently built over the last 10 years thanks to the implementation of technology. So, I mean, I think in a way it makes the architect better because it's like another medium, another tool in the toolkit that we can use to do better architecture.

4:36But I don't think it necessarily created better architecture. And I would probably say that in some cases it probably made architects more lazy or maybe gave more opportunities to bad architects to do more architecture. You know, like, I'm just, again, you know, thinking like, you know, buildings that are, you know, like dumb things like, you know, modeled using standard components from a BIM software that have not actually been designed.

5:08So, there is more architecture that has been put together using technology, but not necessarily better designed thanks to technology. On the other side, though, there is really beautiful building that could not exist without technology. But I would say, you know, across the board, I wouldn't say like, oh, since technology has been around, like it's better all of the architects to do better work. Like, well, when you say that there's some examples of more contemporary architecture, that's beautiful and could not have been created with technology.

5:43What examples are you thinking of? When were they built and who designed them? Oh, wow. I don't know. I'm just thinking. But I mean, like, what do you mean by that? Like, do you mean, well, give us a broad example maybe, but also when you say technology for that, in that instance, do you mean like building technology? Like we couldn't do, you know, certain facades? Like modeling or documenting or, I don't know. I'm thinking, I don't know, maybe Zaha did building would not be able to be built if there is not such technology.

6:14I mean, you could, but you would have to be really good at sculpting physical models to come up with the form and then figure out how to actually build that. Yeah. You know, like, and documenting that to get it built. Like, that would be an immense task to do without, you know, the supporting softwares. So, that's an example.

Building science versus design execution

6:32Yeah.

Building science versus design execution

6:33I think there's a lot of, to unpack with what you said. On that last bit, though, there's sort of a distinction between, there's a distinction between building science and then, let's say, the design of buildings and the execution of the buildings. We need to say, I'm going to define building science as like the performance, the utilitarian functional performance of the building as an enclosure being, right? But, and so maybe the most clear relationship between digital technology and the tools that architecture use and the buildings that are built is something like Hadid or like even shop architects or a lot of architects.

7:14But, you know, where you have, I'm thinking specifically of a facade, the facade is made up of a thousand panels. Each panel is different. Each panel is a different size, shape, and curvature to it. Oftentimes, a complex curve in the case of like Hadid or someone else like the new MAD building, the Lucas Narrative Art Museum, which we haven't seen yet. You know, those are really good examples because that's fusing together directly digital tools, which allow them to produce these types of shapes in the computer and know mathematically exactly the shape of each of those panels, how it gets constructed.

7:54But let's say the shape of each panel and then use a, we would say a digital fabrication method. So it's a machine that's also utilizing the same inputs and math, let's say, to create those panels. Could you do it without that? Yes. Would it be near impossible? Yes. Could you be as precise? I don't think so. And that is, I also want to say that those types of structures, you can look at those structures and say they're amazing because they appear to be complex, which they are. You can look at old historic buildings, cathedrals, for example, and say, well, those are ultra complex.

8:28The difference, though, is that the complexity of a lot of those older cathedrals and whatnot is, I don't want to oversimplify, but a lot of times it more has to do with the decoration on the facade. And it's not like they had a 1,000 page set of CDs. There was a design-build effort, and there were artisans and craftsmen working on site to create the cargoes, to create all the ornamentation without the level of precision that's worked in the DFAB, you know, Hadid example.

9:03Where the relationship between the digital model and the building is like a direct one-to-one. Like, it's almost identical, right? That's not the same thing with the older structures. So, I do, I think... I'm sure the guy who, you know, built Versailles and designed Versailles would love to hear. Well, it's different, though. Like, I think... Oh, it is. So, that is one clear example of the digital technologies for architecture and then bleeding over to construction and squishing down those compressing, right? The gaps and the layers from design to the actual built thing.

9:37Like, before there was much more gap because you were drawing by fucking charcoal or lines in the sand, whatever they were doing. And again, you figure it out as you go. That you don't do that with Hadid, ZHA, facade or a mad facade or like whatever other facade. That's not... You couldn't and that's not what it's about. So, I think that's a really good example of that. But to your first point, that was also... That was the angle with my question or my observation produced my question, which was like,

10:11I actually don't know if architects are able to produce better designed buildings. And I don't mean from a building science standpoint. I mean from a design standpoint as a result of the tools we have in the broad sensor.

Iconic architects and historical innovation

10:30Even if we think of the best architects today versus the best architects 30, 40, 50, 100 years ago. So, could we say that that architecture from 100 years ago... So, you also made the observation of like there's kind of general architecture, which is one scene to critique. And then I'm saying there's another kind of petri dish of like, let's say, just call it the best. The iconic architects, the best architects of these eras. Are the best architects from 80 whatever years ago and their work, is that worse design?

11:04Less innovative, less interesting, less beautiful than the best architects now? Because the architects now have access to more or better whatever technology. And my comment is like, I'm not sure that's the case. No, and in a way, I mean there's a lot of beautiful buildings and whatnot out there. But in a way, I don't know, maybe I would think that when there was not technology, architects would think harder, try harder to create something unique and innovative.

11:38Versus now where we kind of all have the same again, tools in the toolbox.

11:45But they had the same tools as well, which is basically pencil and paper. It's true, but it's like, I don't know, like I'm thinking the Pompidou Center in Paris. Like that's a really innovative, different building, right? They didn't have maybe as much technology as we have today when they did that building. Sure. But it was really different. And it's like, how often do you see buildings like that nowadays? And maybe it goes back to the bigger question of kind of like the design language across the world kind of becoming universal in a sense that nothing really feel like it's super innovative or very different.

12:28Like everything could be everywhere. And maybe that's also kind of like losing that. And maybe that relates to technology. I'm not sure.

AI pattern recognition and creativity

12:35Well, so it does in two ways. One is just the exposure to images, right? Which we've talked about before, like Pinterest, Instagram, this age of like just seeing everything from every corner of the world basically all the time, right? If your feed is not, you know, if your algorithm hasn't narrowed you down to only Kardashian stuff. And then it's also true with the more recent developments of AI technology because AI is basically pattern recognition. So it's a version of something else that's out there or a version of two other things that are out there that are being squished, right?

13:11And that's essentially what it is. And so as probably a lot of people, listeners know, the common position with AI when we talk about creativity and true innovation and what people would say true intelligence is like, well, but if it's just iterating based on previous other stuff, it's not really being innovative. Which I have actually, real quick, with that, I understand that conceptually.

13:46However, you can't tell me when you give inputs to AI to produce a building, a facade, or a video for, you know, streamers, like whatever thing it's producing, that it's not really impressive. And it's not like a good solid B plus sometimes, right? Right. And it's like the, so even if the behind the scenes mechanism of AI is limited in the way that I described, because it is true, it just, it's doing pattern stuff.

14:18It's not actually producing something new. But if you look at the average architect, the average designer, or a junior level designer, or a crappy designer, is AI going to produce something just as good as them? Yes. So really, when we have this conversation about AI, you know, being limited, because it's not like inventing something new, I'm like, okay, well, you're comparing it against the 1% of any creative profession. And, and I also have this other thought, which is, any human being, though, what they produce is a result of them putting together a lot of different inputs and synthesizing that information through their own brain, right?

15:02Right. And I don't know how you draw the line. So in that flow, in that workflow, that process, I don't know at what point you draw the line or make the threshold and say, okay, well, human being, what you've produced is, is a synthesis of this, these other inputs. Therefore, it's not truly innovation, right? Because that's AI is doing the same thing. And at what point, oh, but you've done something that's now like different enough to where we consider to be truly innovative and truly creative, because somehow you cross a threshold.

15:34So I think there's actually a lot of parallels and similarities between the way AI works, and the way a human works, and that everything we produce is a result of other things we see, we might have the capacity currently to draw more inspiration, consciously and subconsciously from a wider set of information, but it's still synthesizing. Yeah, no, I agree.

16:29And that's why AI would, and would keep, you know, would not become amazing unless that person is able to recognize that what AI has proposed is not that great. So that's why I don't think it necessarily better architects in the sense of designing, because you do have to have that understanding of criticizing the work you're doing and how to make it better. And if that's not inherent to you, AI is not necessarily going to do that for you, or you're not going to be able to discern it.

Experience levels and design authorship

17:01I agree. I think also, there's, so when we talk about AI versus the creative person, like this is a very broad, you know, thing to work with in a conversation. Like, it's like, okay, but what level is that, are they a novice, a hobbyist, a person has no, in the case of architect, are they a client, and they're messing around with AI? Is it a fourth year student? Is it a person with two years of experience? Is it a person with 30 years of experience? That makes a really big difference in terms of the, obviously, the competency of that person, what they produce with the AI, and then also, I think, quite fairly, the dynamics between AI and that individual, and the conversation of whether an AI is being of great value.

17:45It's a tricky question, because I was thinking, okay, what if you are, let's say, an architect or a young architect, you're not particularly experienced in designing things, but maybe you have a good sense of aesthetic, or you have a good eye. So you don't have, like, hands-on experience, you know, putting things together, but you kind of know what looks good and what doesn't. And so could AI kind of, like, make you a good designing architect, because you have the understanding on how to, you know, tell it, how to create something, like the qualities of what architecture should have without having to draw the lines, you know?

18:28And maybe, and I don't know, it's kind of interesting, because it's like, well, can you design just based on, you know, words and ideas, and not images and lines and drawings, and give that information, give that information to basically a machine, to spit it out for you, because you, I don't know, you don't have that ability. And how do you consider that authorship, like?

19:00Well, so another parallel with this is how offices actually function in real life when they get to a certain size. So when you have a design director, for example, the design director is the director of design for that, a project or a series of projects. They're the ones responsible for the design, compared to a project manager who's keeping just track of everything. It's not just, it's a lot of work. But that design director in a larger office or mid-size office, they're not actually doing a lot of the drawing.

19:30They're sometimes doing none, and they're literally just giving, literally just giving feedback. Other times, maybe they do like 5% of the actual drawing, and that 5% would be them doing a quick sketch on trace paper by hand, because the design director is still right now of a generation, typically, even for us, you draw by hand or you draw on an iPad or a mirror or whatever. And that's like the ascent of drawing, but it's really just a markup. It's not the actual production of the thing that's being critiqued, right?

20:04And so in that regard, could we say that the design director has any form of authorship over their work? Because it's the same thing, if you think about an individual working with AI, design director working with their team of draftspersons. It's very, very parallel. It is the same thing. Yeah, I think, though, like, I don't know, like, socially, there is something kind of, like, not okay that a machine, you know, would do work for you. It kind of feels like a cheat, you know, versus you have, like, other human in the team doing that work, and you're kind of, like, editing their work.

20:42Yeah. I think it's more of a moral... Ethic. Ethic. Well, that gets into another topic, but I get what you're saying. I understand that and the position that comes along with the ethical argument. But, again, like, if you look at it more diagrammatically or kind of distanced, right, it's pretty close to being the same. I think my point in this last segment is more, like, I really do question this idea that the dismissal of AI because it's not producing true, like, ingenuity.

21:18I'm like, right. But most of the stuff that people produce, creative people, is not that way anyway. And then, again, we're talking about, like, the best of the best. Like, that's super rare. And even then, they're still a product of inputs. So, I do think, like, right now and maybe for the somewhat foreseeable future, to me, so I was talking about that threshold, like, between, like, when does it become, like, true genius or true intelligence and true creativity?

The concept hurdle for artificial intelligence

21:45In architecture, what I see as being that threshold and the final hurdle for AI to overcome before I'm, like, you know, completely blown away is the concept of a project. Because the concept of a project is a really weird thing. Like, it's this, it's this, it's weird in the sense that it is complex, difficult to understand, not from a presentation standpoint, like, it should be clear once you have it figured out.

22:17But it's, like, weird to find that concept. And I say it's difficult to understand because even within the profession, we all talk about it, we all learned about it, 95% of architects don't actually understand what a concept is in architecture. They think other stuff are concepts when they're not, they're just metaphors or just other, you know, formal principles which are not concepts unto themselves. So, it's weird in that regard, but a true concept is also strange because it's a result of, like, really ultra-diverse, unrelated variables and inputs being put together through this one idea.

22:51And that's what a concept is. So, it's, like, in a diagram, it is, like, the concept that you find is a result of you exploring not, like, this of, like, what the building is, but, like, a whole fucking giant three-dimensional, fourth-dimensional universe of ideas, fourth-dimension including time. And then figuring out what the right thesis statement concept is for this project that gets narrowed down to a statement. And then what's fascinating is that thing is robust enough to then produce threads from it, which the whole entire project relates to that core concept idea.

23:28That is a really weird thing. And I don't think AI right now, and for whatever number of years or months, has the capacity to do that, right? It's able to produce things that right now, for, like, iterations of other things, that are good iterations, but they're iterations of other things, but not something where you're taking oranges, bananas, and bricks, and the sky, and being like, okay, the concept is based on these things coming together. It's not, and I doubt it will be, because it all comes from the selection and editing process of the sensibility of a human being.

24:05And AI functions on data collection and analytics, which is a very different way. It is not a sensible way of working, you know, which, I mean, maybe at some point it could get there, but I think it's, again, what would make the difference between humans and robots, it's that sensible understanding of the world. Yeah, the thing, too, is, like, the brain, I'm reminded of our conversation with David Brody from years ago, and he was on the podcast.

24:35David Brody was a, actually, he was never my professor, but he's a family friend who was a long-time professor at the school I went to. Anyway, people listening, go back and listen to it. I think it's worthwhile. He's a fascinating person, was. And in that conversation, he was talking about kind of, you know, he's passed away, so he's from, like, a very different generation. And so in that conversation, he was talking about computers versus, like, you know, not computers. This was, like, well before AI. And he made a remark that, like, he had felt that the human brain is remarkable because it has the ability to kind of, like, flutter and, like, a butterfly almost.

25:15Just kind of, like, flutter around and, like, find weird pockets of thought that seem to be random, like, seem to have nothing to do with, like, the local in-person present meeting. But, like, something comes in and then flutters away and something else comes in. And then when you have a room of three or four people that are doing that, then there's, like, magic. Then you have these different butterflies floating around and all of a sudden, like, oh, oh, oh, oh. And then we have this idea, right? And that capacity to flutter, almost subconsciously, really, is something, you know, like, I think we're both saying, I don't know if AI has that or is going to have that.

25:53There's something almost ethereal about that state of mind where you're just completely open. Maybe it would, like, I don't know if Claude, you know, if you hang out with Claude for five years, it would become a Claude version of David and maybe it would get the, develop the ability to flutter. But I, but yeah, I don't know if that's. Claude, flutter for me, period. But I think, like, it would have to have some sort of, it would have to develop its own personality to be able to flutter.

26:27My thing is that, obviously, AI and the inputs are limited to what's online, right? Oh, yeah, of course. Yeah, and a lot's online. But also, we shouldn't forget, a lot is not online. Like, what does it feel like to move through certain spaces in the case of architecture? What did you feel like because of the sun, the this and that? Like, there's just so many in-person things that we've. Yeah, and that's why technology could also take different forms. Like, right now, it's like keyboard, screen, right?

26:58But there is, like, the vision goggles you could put on, like, things you can put on your temple, things that monitor your body and, like, how you react to. So, if you imagine, like, all of those different types of technology get combined into creating, like, a super AI, right? Like, yeah, we might be able to get to flutter. Yeah, sure, I think that, but that is, like, you would have to be, you are logging every single second of a person's life, of many people's lives from day zero to the end.

27:32And that recording devices would have to capture everything, not just, like, video, not just audio, but, like, what it, what the wind feels like. But I guess, yeah. I just feel like the nuances of that in-person experience, we are far away from capturing truly all of that. And I think when I talk about fluttering, the fluttering is a result, and the ideas from the fluttering are a result of all these, like, random weird experiences we've had throughout life.

28:03And I think that is probably a more realistic limit of AI in terms of making a point than anything else. Yeah, and I think the bigger question is, even if we were able to create an AI like that, what would be the purpose of it?

28:21The larger purpose for getting... I mean, we have a bunch of humans, like... The ideal outcome is that it solves the more serious survival needs of society. So we don't worry about hunger. We don't worry about shelter, protection, for protection reasons. We don't worry about money. That's fine, but do we need them to flutter to do that? Like, if we're basically recreating a human version, but in an AI form, in, like, all of its best qualities that currently AI does not exist in, what would be the point of that?

29:03Like, what would we basically clone who we are into a machine to what extent?

Solving practical problems versus art

29:09Well, I think this points toward the delineation and the handoff between functional, the function of things, like functional stuff, and the art and creative side of things. I know the two go hand in hand. I know they relate, blah, blah, blah, but we can also make that separation. I think that's the difference. So, like, what's interesting is that AI has solved, and by the time this comes out, it'll be different, but has solved, like, half a dozen mathematical problems that humans have not been able to solve to date.

29:44So unsolvable, well, not unsolvable, problems that could not, that were not solved over whatever hundreds of years of math has been around by all the top mathematicians throughout all of human history could not solve these six problems solved by AI. Now, what's, I'm not a math expert, so let me put that out there, but what's interesting, the position, the response from the mathematicians was not what I expected. What they said in response was that, well, AI is not understanding the true purpose of math, and for me, not a mathematician, I'm like, the purpose of math is to solve the fucking problem that you've laid out for yourself, or that's been, you know, you figured out this was a problem.

30:26They said, no, the purpose of math is basically, I'm going to poorly paraphrase, human development, the search essentially for beauty, the quest for knowledge, all of these more, like, ethereal whatever things that you would typically hear from someone who's a painter, right? It's more about the human experience of solving things, like, that's the point of math. Now, I think you can have that feeling for anything that you do, because anything can become a craft and an art in terms of how well you do it.

31:00But, I think that's stupid to say as a mathematician, because math has a purpose, which is for things to be solved, to make life better in a functional way, mostly, right? It's not a painting. It's not an expression of an idea from Van Gogh and his way of seeing night, or whatever else he's drawing, right? Paper painting. Like, it's math. And it would be, to me, it's the same thing as, like, saying, okay, AI has provided a cure for cancer.

31:32And then, um... The point is not the cure. Biologists would be like, well, that's not the point. The point is the quest for the cure. Like, no, the point is to solve the fucking problem so we can use that to build up on that or to cure people. Don't tell me that the point of the math problem is the quest of solving the math problem. That is ridiculous. The only time where... The only situation where that type of position, I think, makes sense is if we're talking about art or the artistic side of things. Like, and buildings have these two sides of the science and the function and the art.

32:04That's it. And so, that was very unexpected for me to hear. And I'm like, I think you're just upset that you're basically... You guys are going to be pointless very soon. Um, and so, but, but, so, to answer your question, no, I don't think the machine needs to know how to flutter or have all those inputs to flutter to solve the more practical things, which is probably the first thing we should worry about, right? Like, taking care... Like, it's the same idea of universal income. A disassociate from AI.

32:35The idea of universal income is that everyone gets a fixed, you know, paycheck per month, per whatever, and that allows everyone to just at least live, right? So, they don't... So, no one has this stress of like, fuck, I might not have a place to live. I can't afford a place. I don't know how to feed my kids. That goes away, right? Now, there's issues with universal income, fine, whatever. But that's the idea of it. And I think AI should be solved, be used first to solve that stuff. Don't worry about creating beautiful pictures. We're about solving that stuff.

33:06The issue is that, can we use AI for that? Why don't we all be like 100% into AI for that reason? The problem is the systems we have in place, i.e. government and the way, you know, things are structured at that level. So, and now a little break for our show sponsors. Dear architects, design leaders, and founders, redefining the industry. If you want to shape not only buildings, but the business behind them, this is your moment. Future London Academy's executive MBA takes you inside the boardrooms of Pentagram, Uber, Zaha Hadid Architects, and Google, where you'll master finance, strategy, and operations from the people who already excel at them.

33:47Study along peers, transforming the industry worldwide, and do it all in just nine intensive weeks over one year, between London and California. The program is award-winning, transformative, and crafted for design and architecture leaders with 15 or more years of experience. If you're ready to grow your impact, apply now by tapping the Future London Academy link in our show notes. Say goodbye to design-unfriendly spreadsheets. Programma, built by designers for designers, provides an integrated suite of tools for every phase of your project. From mood boards to advanced specification tools to project management.

34:19Seamlessly integrated, easily shareable, and always current, Programma reduces redundancy and minimizes costly errors. Share your work in the format that works for you. Be it PDF spec sheets or online branded client dashboards with always up-to-date specification details. Join interior design movers and shakers and start for free by clicking the Programma link in our show notes. Experience the difference with Programma, software built for interior design workflows.

Impact of AI on client relationships

34:44Bringing it back to architecture, I think what's fascinating to me with AI, because the big question, of course, is will it replace architects, is I think, in a way, this technology is going to force the question of, is the value of architecture and architects, is it more to do with the function of the building from a use standpoint, and the function of a building from a building performance building science standpoint,

35:15or is it more to do with the artistic statement that's being made by the creator, and being felt by the users and the visitors of the city? I think that that question is going to be forced upon us more so than it ever has been. I'm thinking, you know, is AI going to help people understand what good design is versus not,

35:47or is it just going to make more people think that bad designs are good designs? Because I do think within the profession, there is obviously advantages and disadvantages, and architects could kind of, again, use AI to be better, right? And I'm not too worried about it. I don't think that architects are going to be replaced by AI anytime very soon, although it could be completely wrong. But I think my biggest worry with AI in our field is what does it do to, like, a client?

36:22And how does that impact our relationship with the clients, and obviously, like, the final product that's being designed? When you say the client, are you thinking specifically of working with them, that they're going to be constantly AI-ing, you know, the architect's work? Or are you thinking more that the client, their education that they bring to the table, their knowledge that they bring to the table, is lesser or greater as a result of AI? Yeah. Understanding the design.

36:52Yeah. It's a good question. It was the same thing, you know, when Pinterest came around, right? It's like, oh, I know those people who come in with their images, and that's what they want, and, like, they're not able to discern, like, is this good? Is this bad? Is that fit my project? Now, it's even, in a way, worse than Pinterest, because Pinterest, you just kind of collect images and throw them on the table. AI, you grab, you transform, you edit, and the client has the ability to do that, and they can think that it is good because it looks, you know, decent.

37:23There's lights and real trees. Because it's sunset, yeah. And, you know, it's moody, and it's sexy, and whatever. And therefore, they kind of, like, get trapped in the image game of AI. The glossiness of the rendering. The glossiness of the rendering. You know, the vision that AI is painting before the vision is even established. So, I have a couple thoughts with that. I do think, as we go forward, if you were to take the basic prompt, let's say my client, and I have a rough sketch,

37:55and I say, AI, fix this sketch of a floor plan, make it more livable, buildable, and then here's a rough idea of what I want facade, make me a beautiful facade. I think, as we go forward, it is going to produce better and better design. Like, not just from a functional standpoint, but, like, let's say it's facade design. It'll produce a better design than it did, than AI did five years ago, two years ago, ten years ago. Right? And so, the output from AI, the base output, I think is going to be raised over time. Yeah. I do not think that that, I think that's completely separate from whether or not the general public becomes more savvy and knowledgeable

38:36and understands what is good and bad design, what is good and bad taste, what is good and bad aesthetics, what is good and bad proportions, all that stuff. I don't think it'll have any impact at all. And to your point, even if, in the future, the AI is, at the first, you know, output, is producing a B-plus level architecture, that client can quickly make it go back down to being a D-minus and think it's phenomenal. So, I firmly do not believe it's going to help educate clients in that regard.

39:07It will from a functional standpoint, because it's fucking data. It's fucking, it's dimensions. Oh, two foot's too narrow for a door. Oh, I didn't know that. Make it three foot. Okay, easy. Learned. That's not difficult. But I'm going to call it the design and the composition of things. That is much more nuanced. And I don't think you can, as a human being, I don't think you can understand that unless you're born with it. Some people are. Or you've been trained in it. Those are the only two ways. You can't just learn it passively by talking to an AI who's going to give you stuff you're requesting.

39:39I do not believe that. But if you want to use AI as a school, then that's different than you're being educated. But without actual education, and when does actual education come into play with AI in this way? I don't know. Without education, it doesn't exist. And because we've seen clients who, I mean, look, any architect designer knows this. You produce something that is good, you think is good, but it is good. You present to the client, and they want to do some wacko variation that makes no fucking sense artistically. You're like, this is horrid.

40:10What are you talking about? And there's nothing we can do really to train them in those kind of extreme circumstances. And so, no, I don't think it's going to improve them. I can understand, though, where it could lead people into a false sense, an even greater false sense that they have a better understanding of design than they do, because they're looking at pretty pictures. Yeah. You know? And because the AI won't tell them, like, oh, Bob, I think, you know, the prompt you just put in is a bad idea for what you're trying to do, right?

40:43Like, the AI will have to push back. And maybe it will. And maybe it will. But so far, it just, like, takes in what you're giving it and, like, feeding you stuff.

Training the human eye and taste

40:54I just think, because you're talking about the knowledge of the human being, right? And so, when we talk about this hypothetical client, the general public, that's the same, roughly, as when I was a first year in architecture school. I probably had a better handle on aesthetics than the average person at that point anyway. That's why I went to architecture. But, you know, day zero, day one, right? You basically don't have anything. When I think about that version of David, and I think about who I am, you know, now, but even at the end of five years, that is a massive amount of change and growth and knowledge and understanding of what beauty and composition is.

41:37And putting aside all the other stuff about architecture, then we think about that year six, in my case, a version of David at the end of undergrad versus, you know, after six years of practicing versus 15 years of practicing. Those are each very different versions of David. And my understanding of these things has evolved tremendously, right? And I'm like, no, no, no, no. A person who starts off with an even lesser understanding of just what looks good than me when I was a first year is going to learn that through AI, through a few chats for a year or so.

42:11No, no, no, it's not going to happen. I think it goes back to, you know, the practice. I mean, can AI make me a better tennis player? I don't know. Like, it could tell me how to play, but is it going to make me a better tennis player? No, it's going to come from the years of practicing, right? The training, the muscles. The dedication. The dedication, you know, the micro-improvement in the techniques and the shoes I'm wearing and, like, the racket I'm using and, you know, all of that stuff. So it's...

42:42Yeah. And in the case of architecture and the complexity of buildings, you know, AI might produce a half-decent option for this facade or this one wall in a room. That has nothing to do with whether or not that standalone item makes sense for everything else in that building. And everything else means all of the rooms or the facades, the structure, the site, the people, all those things. At some point, it's going to be pretty good at doing that as a whole package.

43:12But, you know, there are, you know, what do you say, layers, levels to the game. There are levels to all of this.

43:23But I think that, going back to it, the reason why I think it's going to force the question of, is the value of us human architects more the functional stuff or the artistic things? Is because... You can tell that we're both still sick. It goes back to your point and question of, well, do we need AI to evolve to a certain level to solve the functional stuff in society? It's the same thing with architecture. AI will first be able to solve easily the functional aspects of architecture and the building science aspects of architecture.

43:58And it should. Why not have AI, you know, ensure that all the facades being designed and hopefully built are better than what they are now with current tools? Why not have AI figure out a fucking new material that solves three things at once for us? Why not have AI improve any floor plan so it actually works? And a lot of floor plans I see by architects are garbage. Which, yeah, that's a no-brainer, right? Because that is, again, to me, a more mathematical, you know, aspect of what we do.

44:30There's a right and a wrong.

44:33Fine. I have no issues with that. I think it makes sense. When we get to the expression of things, that's the question. And it's maybe the argument is more the one you'd hear from a pure artist who doesn't do, doesn't produce work that has this dual personality of being both artistic and functional. So let's say a sculptor, right? Like, or a painter. Someone was, I heard someone say recently that, you know, AI producing art.

45:03Like, let's say you go to a museum and there's this gorgeous painting and it was done by AI. The Mona Lisa. Would you care? You would only care because it's like that's an interesting and impressive use of technology. But would you actually care about the painting other than the novelty of, like, other than the power of the technology? You would not care. Like, you just, you wouldn't care. There's something about the art, it being made by a person, that is, is the thing.

45:37Like, Starry Night. It's mesmerizing because of what it is. But I think now we're realizing a lot of it is mesmerizing because we know it's made by a person. And that's coming from an individual who, I'm not a painter and I don't know a lot about art. But when I look at it or anything else, I'm like, it's not just a thing. It's that someone made this at a point in time with the tools they had and the society that they had. And that is, that's worth, that's what caused me to take a pause and to reflect.

46:07So there is something there, I think, that will always be there because we are humans at the end of the day. We're not robots. And I think the question is for architecture, does that apply to architecture at the broad scale, most of the buildings? And does it apply to, and or does it apply to the 1% of the architecture that are the, like, more iconic, let's call it iconic architecture or artistic architecture, whatever you want to say.

46:39And if that goes away, then there's really no use for the architects. But I think that is part of what we do because it is part of what we do. We produce buildings that have to emote poetry, you know, for the city. And so, so yeah. And that was David's mother-in-law about AI. That's where I feel like what's going to happen is that all of the generally good architects are going to get washed out.

47:13Because the tools will become so accessible for the average person that they won't even need to go through an architect as an intermediary kind of bridge to use these softwares. I wonder how the GCs are going to respond to that, though. Like, okay, here is a set of drawings for my home edition remodel that was done by AI. And now go build it. You know, at some point, those drawings, from a technical perspective, will be just as good as the average architect today, if not better. Okay, the GC has questions about those drawings. How is that workflow going to go?

47:43They're going to talk to the AI. Do they want to be... You can barely get GCs to respond to email in some cases. You really think they're going to sit down and type prompts to get answers to their question? This gets into, like, what is the apparatus of the AI? You know, maybe at that point, we're all wearing these pins where we have... I don't know, is the question. I've seen videos of the... Is it the Tesla robots? Some sort of robots walking around in San Francisco. Have you seen those? No. I think it was in the financial district.

48:14You're like, you know, it's like, you know, the towers in downtown SF and people walking around in cars driving by. And there is this, like, awkward walking robot that just walks on the sidewalk. Like, he just belongs there. And someone was like, this is what life is in San Francisco. And you see Waymos and robots. And soon you're going to have, like, you know, flying saucers and things. Flight is a... Teleportation, as you often say. That would be the one. But I saw something about teleportation. Did you see it?

48:45Whatever you saw. No, no, no, no, no, no. Isn't it for Fool's Day? Because this always happens. It was... Well, it wasn't true teleportation, but I saw something. I think it was using, like, dark matter energy or whatever stuff to transfer an image from one place to the next using some, like... I'm not a scientist, so I can't really explain how they did that. But, like, some, like, weird lighting thing, whatever stuff. And I was like, we're getting closer. Are you talking about email? Was it an email? Someone sent an email to you?

49:16It was an image that they were able to recreate somewhere else by using, again, some sort of, like... That's cool. ...dark matter thing. And it took, like, a shit ton of energy, but they managed to do it. I think it might have been in China that they did that. And I was like, of course. They're going to be the ones who come up with teleportation. Yeah, or they're lying. That happens. Says the Chinese one. That happens often. That happens often. Are you lying right now? But so... The other interesting thing I heard from a philosopher regarding AI is that everyone likes to focus on what it can and cannot do.

49:55And then there's a whole ethical thing of, like, stealing other people's work to feed it. And that's all been talked about, but this particular individual was saying something which was kind of, like, getting to the root of all of it, which is AI, to answer all these questions, the core question is what do we want out of life as humans? I think... Chocolate. You know what I mean? Like, that's the thing. And I think that also can directly point to what we just talked about, more, like, functional, mathematical, pragmatic stuff versus artistic things.

50:27And, like, what do we want out of life? I think it's easy to say everyone wants to have the baseline stuff covered, not to worry about getting sick, not to worry about not having enough money to feed yourself. Like, that's obvious. So there is, like, what does the general public want from AI? And then there is, like, what do the AI company want out of AI? That, what they want is just, they want, they want everything. And that's a, I don't want to talk about this, but that's a crazy thing with, like, all the tech CEOs. Like, they want fucking everything.

50:59Yeah. It's like, there's not a single thing they don't want. They want it all. And... And I think that's kind of maybe the most disturbing part of AI. It's not so much, like, the product or the solution that's being offered. It's like, what is the outer motive of coming up with this technology? It's money. It's money. Yeah. And therefore, again, like, are we all little minions that are being manipulated to a greater deal for, you know, one big entity?

51:29The entity meeting one company? Yeah. That's, that's, yeah, that's another, that's where we're headed. I mean, there's only a handful of companies that basically own fucking everything at this point. And it's going to just kind of decrease and decrease, I think.

51:45Yeah. I think that's, that's where we're headed. I think, though, it is a good question. Like, we talk about architecture as being a public service and the value of architecture and buildings and the impact it has on the cities and how people live in cities, how they experience the city and the well-being of, you know, daylight and, you know, beauty and all this stuff. And I don't know. It's sort of like, let me put it this way. You actually don't need a building to be an iconic Gaudi, Gary, Wright, Hadith, whatever building for it to be beautiful and serve the public to the maximum.

52:29You know, and so I could see a world where, again, most of the city buildings are designed by maybe AI. And actually, the hiring an architect and paying them a lot of money is reserved for, like, those iconic structures, the monument structures, the one-off houses. And at the end of it, that's, it's a bleak kind of take for architects in a way. But if our objective as architects is to improve cities and something does that better, more consistently and more cheaply, well.

53:08Actually, you know, like we always say, like, you know, architecture and architects should better cities and make them, you know, more beautiful, more functional, more everything. I actually wonder, like, do most architects think like that or is it mostly us, you know? Like, do most architects actually just think about that one building they're working on and they just wanted that to be good and the rest I don't really care about?

53:35I, my inputs are skewed because I went to urban design school and I taught urban design and was part of an urban program. So I'm, a lot of my inputs maybe are more for people that are like that anyway. But my feeling, even with that asterisk, is that the majority of architects would completely agree with that mission statement. Whether or not they get to put into practice is a different thing more to do with the practicalities and the difficulties of executing and getting clients and budgets and fees and all that stuff.

54:05But I think that's true.

Designing without computers in the past

54:07I want to go back to the, real quick to the very beginning and say, and ask you, if you imagine that you were the same individual, but you existed X decades ago in a time where there was not computers being used, no computers, zero, no CAD, no nothing. Do you think your capacity to create good design, because there's a whole thing that doesn't get built, but that's more to do with practicalities, like I said, of the world and commissions.

54:40I would have been a better architect back then than now. Do you think you would have been better at designing spaces then versus now or that you'd be better now because of computer, Enscape, VR, 3D printing, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah. Well, I was never great at drawing.

55:02So I would tend to think that I'm probably a better architect now than I would have been, you know, like a hundred years ago. I think, you know, floor plan and rational and critiquing the work, I would have been the same. But like designing things in a three-dimensional way. But you would have built up that skill set because you would have had no choice. Well, that is true too. The drawing skill set was stunted because we live in a world where you use computers. Yeah, but you also have like the natural abilities that you have as a, like before training.

55:33Sure. You know, like some people are, they're bored and they know how to draw. Like they're just gifted. Like the eye and hand coordination and expression just flows like magic. Okay, let me kind of help you remove that piece of it. So let's say the marina that we're comparing, I think you're thinking about a self-performing marina that's like you're producing a building by yourself. I'm going to put in another layer of information and say in these two marinas, the scene is that you're the design director. So you have people working under you who do like the bulk of the drawing, like I said.

56:04So you have to draw because there's no computers, but your drawing is not like with the parallel rule. It's, you know, you're doing markups, let's say, and you're thinking conceptually. So it's, therefore, it's less about the skill set of the hand. No, I mean, no, because I think that conceptual thinking is, would have been the same. I think, I think, again, that comes from like inherently who you are and then like what you, like how and what you've learned in school. And I don't think that would have been any different between like now technology or like 100 years ago no technology.

56:36I totally agree. Like it's not like technology is teaching me to think differently. I know. It's helping me with the outcome, like the communication, you know, the thing that's being communicated to other people, but it's not helping me come up with it or like, or like think about it in a different way at all. At all. Right. No. Like not at all. No. I completely agree. So there's two things. The presentation, the presentation bit that you said, that is all, that's true, but that's also only relative to our current context.

57:12Right. So if you go back a hundred, a hundred years, it's not like people are like, well, I don't understand stuff. They had no comparison. They wouldn't be like, give me a photorealistic rendering that, that, that, that vocabulary didn't exist. That idea didn't exist. It's like building a castle and then just go and build it. So even back a hundred, a castle, but a hundred years ago, the presentation aspects of what you said wouldn't apply because the, the, the, the way we produce stuff to present stuff existed for that time. And that's actually an interesting point. Like at what point in history were architects asked to like, do like presentation meetings and put a boards and whatever?

57:46Like it's, it's, it's purely to do with the other broader technological advancements and the way society works. Right. So we only, it says weird vision. Huh? No. What? Like, like the invention of printing television and all of that stuff, right? A hundred percent. So we own, so it's a, it's a, it's a, it's a weird circular thing, right? We only have the perception that these tools allow us to communicate better, but no, we have the perception that we need to communicate better because we can communicate better because of the tool that exists.

58:18Right. Like, it's a feedback, it's a feedback thing. That's what I'm saying. Yeah. We're being forced to use those tools. Yeah. But we didn't need to originally. But it's just, it's just relevant to our current context. Yeah. Right. So. Same thing with AI. It's a, uh, possibly. I, I, I, and I, but I want to reiterate. Or, or maybe we've, you know, maybe we feel like we have to use these tools in order to stay relevant within the period of time these tools have been showing up. Exactly. Like, okay, well, we invented books. So, well, now guess what? We can't draw in the sun anymore.

58:48We've got to put it on paper. So we're part of the, of the club of putting things on paper, you know? And there's always this idea that it's, this version of presentation is better than before, but let's kind of unpack that. It's maybe more, in the case of architecture, it's more real, it's closer to the real thing. And if that's your metric of better than true, but, and this is a really big but, I don't, no one really cares about the quality of the presentation.

59:19At the end of the day, at the end of the next 10, 15, 50 years, does that fucking matter? It does not. The only thing that matters is the building. So do we say that Frank Lloyd Wright's structure is less impressive because his drawing wasn't as realistic back then? Who gives a fuck? No. So it's, it's better, but like really at the end of it, no. But so I think it's kind of an interesting observation because it almost feels like from ages ago, the architects and architecture has been trying to follow along the innovations happening in, you know, other places, right?

1:00:00To stay relevant or more appealing or in the hope that society will understand what we do or like whatever quest, right? Yeah. And it's kind of like, well, if that's the case, then we'll kind of like put it up on ourselves. We didn't have to. We didn't have to. And, and, and, and even to this day, do we still need to do that? I mean. It's only a need because of all the peripheral, peripheral and related professions, right? So if you see that other people who are creating, you know, 2D artworks or videos are doing this tools and it's like architect, why can't you produce an image that looks like that?

1:00:34You know, of course, um, it does, the, the, the more transparency there, how do I, how do I phrase it? The images we show to clients, the closer that is to reality and the more of that in a way is better in ensuring the thing gets built that way. That is true, but you know, I, I also 100% agree that I do not believe that my capacity as a designer, uh, an architect on the design side of things is better now than it would have been, uh, because I'm using digital tools.

1:01:08If I had existed in a time, racism aside, where I had produced everything, produced everything by hand, my capacity to produce good design and my capacity to critique and give feedback to people would be identical. If anything, I think on the broad view of stuff, I could see an argument being made where the quality of the architect now, now and their design skillset is lesser than it was X years ago because the amount of stuff the architect has to learn and do to keep up with society is, is, is 10X what it was before.

1:01:47The amount of time that they have left to spend on design and thinking. Exactly. Yeah. Not for sure. It's all toward BIM, VR, uh, AI producing construction document sets that are like literally 10 times what it was before. I mean emails, correspondence, communication. Documenting everything. Yeah. All that is time away. If we're talking about one person, time away from that person honing their craft of just designing and you need that time. So I don't think, uh, to answer the question, I don't think that actually technology has played any role at all in improving the quality of design and less even design thinking from myself, uh, for myself or architects in general.

1:02:27Um, you know, at, at all, at all. And, and it also is to say that a lot of that knowledge and expertise, where does it come from school? You know, it comes from in person meetings and conversation and stuff. Um, how much of the essence of this knowledge we're describing that was taught to you in school that you gained over the years in offices was not documented online as a thing that AI could learn from.

1:02:57Um, almost all of it. Yeah. That's true that the, the knowledge that we're talking about, it's not documented online. It's not documented online because it is, it is, is so valuable and difficult to grasp. It is a result of hundreds of conversations and pinups, thousands of them in different formats, different people, different projects over time. Then you develop that knowledge. That is not online. That's not, that is not accessible to AI at this point in time. Yeah. All right. Take that.

Episode wrap up and contacts

1:03:27Yeah. So those are our thoughts. So cloud and chat GPT, if you're listening to this episode, please work on yourself. Um, yeah. So those are the thoughts and we'll give you an update on our thoughts. Not for another six months, I think. All right. Outro. Thank you everyone for listening to David's monologue for an hour. If you like what we do, please leave us a review, subscribe on all of our channel. Uh, you can find all of those episodes on a second studio pod.com, the website. We're also on YouTube, social media, Instagram is kind of the place.

1:03:59If you have any questions, any followup topics or guest suggestions, you can always reach out through the hotline two one three, two, two, two, six, nine, five, zero. And, uh, there is a few more weeks before the end of the year. So stay tuned. A few more episodes before we get onto 2027. Sounds good. Talk soon. Bye. Bye.