Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Room to Think

Room to Think artwork

AI And The Future Of Architecture

Nov 3, 202656 minAudio
ShareXLinkedInFacebookWhatsAppEmail

In this episode, Lyssia sits down with architect, researcher, and Carnegie Mellon University professor Daniel Cardoso to explore how technology has changed the way we design, build, and think about architecture. Drawing from his research into computational design, the history of computer-aided design, and emerging technologies like AI, Daniel looks at a question most of us rarely consider: how much are architects shaping our buildings, and how much are the tools they use shaping them too? The conversation begins before AI, tracing the shift from hand drawing and large teams of draftspeople to computer-aided design. Daniel explains how digital tools did more than make architecture faster. They introduced new ways of working, new levels of complexity, and entirely new possibilities for what architects could design and build. They also explore the surprising origins of architectural software. Many early computer-aided design systems grew out of industries like aerospace, manufacturing, and defense before eventually making their way into architectural offices. As those tools became more accessible, they began influencing not only how buildings were documented, but potentially how they looked. Daniel shares an old joke from the architecture world about being able to drive through a city and identify which version of AutoCAD was used to design a building. While the reality is more complicated, the idea raises an important question: can the software architects use quietly shape the cities around us? Lyssia and Daniel also discuss what computers have made possible, from the complex forms of buildings like Frank Gehry’s Guggenheim Museum Bilbao to the enormous number of systems that have to be coordinated inside a modern hospital. But as architecture becomes more digital, what might we lose? Daniel explains why drawing remains such an important way for humans to think, communicate, and make sense of ideas, while also challenging the romantic idea that architects spend most of their time sketching. In reality, much of architectural practice is about coordination, documentation, collaboration, and turning an idea into something that can actually be built. The conversation then turns to AI and whether architecture may be approaching another major technological shift. Daniel explains why AI has potential across design, construction, simulation, and analysis, but also raises important questions about where these systems get their data, who controls them, and what happens when creative work becomes dependent on increasingly opaque technologies. They explore the rise of AI-generated architecture and the idea that machines could process enormous amounts of information to produce designs humans might never imagine on their own. Daniel pushes back on the idea that AI is independently “designing,” reminding us that people still decide what data a system learns from, what it is asked to create, and how its output is used. The episode ends by looking toward the future. Daniel imagines technology that does not simply make architecture faster or more automated, but helps smaller teams collaborate on increasingly complex projects while creating better ways to work and, ultimately, better places for people to live. By the end of this episode, you may start looking at the buildings around you a little differently and asking a bigger question: as AI becomes more involved in the spaces around us, who - or what - is really doing the designing? More Room to Think: Daniel Cardoso Llach Carnegie Mellon University - School of Architecture Books by Daniel Cardoso Llach Builders of the Vision: Software and the Imagination of Design - Routledge, 2015 Designing the Computational Image, Imagining Computational Design - with Theodora Vardouli, Applied Research + Design, 2023 Room to Think https://roomtothinkpodcast.com/ Loved this episode? Let us know! Subscribe to Room to Think If you enjoyed this episode, leave a review and share it with someone who would appreciate a more thoughtful approach to their space. New episodes every week. Build a better life by design. Spotify | Apple Podcasts | YouTube | Instagram

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Lyssia KatanCan AI ever replace an architect?

Daniel CardosoHe used to joke that he would be able to drive around the city and say which version of AutoCAR was used to design whatever building he was looking at.

Lyssia KatanA human can sit down and say to an AI model, design a hospital for me in the style of Frank Gary.

Daniel CardosoDrawing is a fundamental way we engage with the world, that we communicate and make sense of the world around us.

Lyssia KatanNow that AI is entering architecture, does it feel like things are gonna change almost to the level of when they shifted from hand drawing to CAD?

Daniel CardosoIf you read what the White House is saying about AI, it's very similar. But it's not easy to be optimistic these days.

Lyssia KatanAnd there's something really interesting happening today, which is that what if the tools architects use are shaping the buildings around us more than we even realize? Welcome to Room to Think. Today I'm sitting down with Daniel Cardozo, an architect, researcher, and professor at Carnegie Mellon University to talk about how computers have changed the way we design and build. We get into what architecture looked like before CAD, why some of the tools architects use actually come from industries like aerospace, and how software has made entirely new levels of complexity possible.

Lyssia KatanDaniel helps me understand what happens as AI enters the architectural process, who is really making the creative decisions, and why the systems we build may matter just as much as the designs they produce. Because the real question isn't what technology can do, but what we choose to do with it. Let's get into it. Hello, Daniel. Daniel Cardozo is an architect, researcher, and professor at Carnegie Mellon University, where he leads work at the intersection of architecture, computation, AI, and robotics. His research looks at how technology is changing the way we design, build, and even think about the spaces around us.

Lyssia KatanDaniel, welcome to Room to Think.

Daniel CardosoThank you, Lisha. That's a very nice introduction. I should use that on my website and everywhere.

Lyssia KatanPlease do. So, so Daniel, for anyone who has never heard of you before, what do you do and why do you do it?

Daniel CardosoThank you, Lisha. Yeah, well, first of all, thank you for inviting me. It's very nice to be here. I'm an associate professor in architecture and I work at Carnegie Mellon University in Pittsburgh. And like you mentioned, I teach architecture, but I also study how technologies change the way we design and how we think about creative practices. I direct um a research program uh on computational design where we approach these questions um both experimentally and also theoretically. So part of my work is historical, part of my work is experimental, which is unusual, but it's it's uh it's something I like.

Lyssia KatanSo, for someone who has never really thought about any of this, what does computational design actually mean?

Daniel CardosoThat's a very nice question. I think uh the way I think about that question is by contrasting the term computational design with another, perhaps more common name, which is computer-aided design. So I think many architects and designers were trained in uh in the use of software tools for drawing or for kind of modeling and so on, under this idea of computer-aided design. So I think of computer-aided design as a set of methods and ideas and technical systems that support existing ways of working, existing ways of designing.

Daniel CardosoAnd by contrast, I think of computational design as um ways of using computational ideas and methods in ways that more kind of uh that are more transformative and that and that try to change the very way we design. So instead of supporting existing workflows or existing um forms of documentation or analysis, they use algorithms for uh analytical purposes or for generative purposes and so on.

Daniel CardosoSo it's a rethinking of design through technological and computational ideas as opposed to a sort of supporting of existing design practices.

Lyssia KatanSo before we get into that, I I do want to take some backstory and understand from you before computers even entered architecture, how differently did architects think about designing a building?

Daniel CardosoYeah, that's that's a very nice question. I think if we look at it historically, I think there's a there's an interesting moment at the beginning of the 20th century when architects and architectural firms in the US specifically start to become larger and more complex. Buildings become uh more sophisticated, there's new there are new materials entering the industry. Uh, the complexity of the social programs that buildings are expected to fulfill is also becoming more complex.

Daniel CardosoAnd that demands from architects and designers different kinds of engagement, different uh like much more work in terms of documenting, in terms of uh being able to describe with precision um increasingly large uh artifacts, buildings, infrastructures, and so on. So throughout the like the first half of the 20th century, what we see is architectural firms becoming more complex, bigger, and more hierarchical. And when computers enter into the equation uh in the 60s and 70s, that's kind of the universe of architectural practice that they that they come to support.

Daniel CardosoHow were they designed? Well, I mean, these practices, these architects, these firms were working mostly through drawing, right? Drafts people, uh, like very large, often very large groups of drafts people uh entering uh um working under the supervision of an architect and a large number of uh architectural workers who are working on drawing, drafting, and sort of representing the building in in the ways that um society demanded at the time.

Daniel CardosoUm what is interesting, I think, of that question is that with the with the uh adoption of computing in architectural firms, that general equation doesn't change. Like we we we have even there's there's sociological work on this, that there are studies that show that architectural firms actually continue to become increasingly large and hierarchical after the adoption of computer-aided design systems. So it looks like the underlying social organization and collaboration structure of architectural work hasn't drastically changed as a result of the adoption of computers.

Daniel CardosoWhat has changed are the substrates of the practice itself. So instead of, you know, instead of like large kind of warehouses full of people drawing and uh using their kind of T squares, etc., we have what sometimes people call a little pejoratively cat monkeys or or beam monkeys, right? People, people who are playing the role of the drafts person, but now in a sort of digital uh format, right?

Daniel CardosoLike using software to do that. But the fundamental equation hasn't changed that much. There are exceptions, and I think probably we'll we'll get some of those exceptions. But in general, what is interesting to me is that the fundamental equation of architectural work hasn't drastically changed as a result of this adoption of these technologies.

Lyssia KatanDid it make it faster or or even change what architects were capable of designing? Maybe it wasn't possible on paper.

Daniel CardosoYeah, I I think I think that's that's um that's a good that's a good point. Definitely there's something interesting about computer-aided design software and computational tools, is that they, to me, they are like points of convergence of different kind of industries and different forms of design practice, right? So what I'm what I'm discussing specifically is the fact that a common tool like rhinoceros, for example, that many architects use and many designers across different fields use rhinoceros fundamentally at its kind of the geometric and mathematical engine that drives the generation of the geometry in the software, is a system that comes from the aircraft industry, right?

Daniel CardosoSo that kind of convergence of uh aircraft design and ship design as well into the world of architecture definitely makes it possible for people in architecture to use tools and to kind of employ shapes that were not so easy to use before that. They were used, and anyone who's looked at kind of 20th century, great works of 20th century architecture can point to examples of amazing uh geometric and uh and building fits.

Daniel CardosoBut the fact that these mathematical techniques are codified into these softwares, make them available to a broader audience, and for sure, I think they slowly start to kind of become part of the of the of the language of architecture today.

Lyssia KatanSo, prior, none of this technology was accessible to architects. Did they all say all the technology stayed in its own field or industry?

Daniel CardosoYeah, that's uh I think that's a good way to put it. Um the the origin of computer-aided design systems was not in architecture. It was, they were originally uh created to support the design of airplanes and and uh weapons, right? It was it was part of a uh US Air Force funded kind of series of contracts in the 50s and 60s that evolved into what we understand today as uh computer-aided design systems. Um so the tools themselves have uh have an interesting sort of legacy.

Daniel CardosoUm that that that is kind of we can think of it as sediments, like as a layer, historical layers in the structures that allow us to access modes of designing that come from other industries. And it was only later in the 80s and 90s where computers became cheaper and uh and personal that architectural firms were able to afford them. And it was then that more specific packages for architectural design started to appear.

Daniel CardosoAnd uh it was then that like medium-sized architectural firms were able to finally afford computers and experiment with them and so on, but up until then it was it was it was not so easy. Um and with the specific case of uh rhinoceros that I was mentioning earlier, that's an interesting case itself because um uh Bob McNeil, who's the the person who kind of started the the process, actually started as a um Rhinoceros started as a as a package, as a library for AutoCAD, which is kind of the the most to these days the most commonly used architectural architectural software, because up until then uh uh AutoCAD didn't have any kind of surface modeling capabilities.

Daniel CardosoSo yeah, so there's a I don't want to go into too much detail, but I the I think the point I am trying to um underline is that software systems are like cultural objects that are composed by layers of knowledge that often come from places that you would not expect, and they become available as part of a language of design that architects and and other kind of creative practitioners use. And I think that's interesting.

Daniel CardosoI think that's I think that's really interesting.

Lyssia KatanI think that's fascinating. I would have never known that that's that's where it came from. But as with many inventions, very often they're invented for one thing and they're they end up being useful for something totally different.

Daniel CardosoRight.

Lyssia KatanDo you think that these tools that we're designing with that have become so mainstream are shaping the buildings we live in in a sense of making maybe things feel more similar, copy-paste, or is that not happening? Right.

Daniel CardosoUm yeah, I I think my answer will be like maybe in some cases yes. But I I I I think it'll be a little hard for me to generalize. There's a there's a there's a uh a joke that circulates in uh in in my field or in my subfield, which is that um uh it's a joke that Philip Bernstein, Philip Bernstein, he used to be the vice president for Autodesk, which is the company, the software company that does Revit and Autocad and so on.

Daniel CardosoAnd and he used to joke that he will he would be able to kind of drive around the city and say which version of AutoCAD was used to design each, like what whatever building he was looking at. I think it was probably exaggerating. I think it was a good joke, and and and I think he's he's done it enough times that I've heard it from different places. Um but I I but I think there's a kernel of truth there. I think if you pay attention to the kind of um geometric sort of um modisms, the the kinds of articulation that you see in different buildings, you can maybe make a case for that.

Daniel CardosoYou can start to see like certain kinds of curves appearing in the 90s when kind of spline curve uh drafting became kind of more easily accessible. But but I would also say that I think architecture is a much more complex cultural product that is determined by factors that extend beyond the specific tool that we use to describe it.

Daniel CardosoSo I I would be I like the job, but I I would be cautious about taking it too far. I think I think um we can think of architectural styles and architectural um forms as part of an ongoing conversation that uh that happens uh through time and through culture in ways that include the systems that we use but are not exhausted in the systems themselves.

Lyssia KatanI'm curious to know what are some of those factors that you say influence the architecture? Is it cultural, is it historical, or are there things that maybe we outside the industry have no clue about?

Daniel CardosoYeah, I I think it's a it's a very good question. I think that um I was trying to point out things in general, like the evolution of architectural styles, and people have looked at this into question, it's not my area of expertise and so on, but I think if we had a historian of the Renaissance, right, they would be able to explain in what ways um the forms were the result of as much a technical sort of mastery over certain kinds of materials, as much as uh the product of a culture of cultures of craft and and and building that that produce uh particular kinds of building outcomes.

Daniel CardosoAnd I think today I would say that the way we employ software to organize information over a building, uh the way we use software to uh channel communications and collaborations between different companies and trades and so on, I think that definitely has an impact on on these things. But I think that um when when uh when we look at a building today, we should include the technology that we used to build it as a as an element of the explanation, but I think we can also look at other things.

Daniel CardosoAn evolving cultural environmental kind of consciousness about materials, for example, is definitely shaping how we think about buildings. Um the kind of cultural requirements related to how we use space, I would say that also kind of affects how buildings get designed today. And I think from a technical standpoint, standpoint, thinking about computers, specifically the fact that a lot of public architecture and a lot of um well, a lot of big projects today are created using uh a software paradigm that is often called building information modeling, right?

Daniel CardosoThis is the idea that you have like a data representation of the building that is three-dimensional, that includes information about quantities, about um structural kind of properties and attributes, about budget, etc. This kind of idea that a data representation of the building can codify not just the shape, but also the logistical, mechanical elements of the building. I think that also has uh an impact in how buildings operate, because the way these systems work is that increasingly, as a designer or as a building practitioner, you you have access to libraries of elements that are more or less predefined that you can edit and adjust, and but you are working with a repertoire that is more or less defined.

Daniel CardosoAnd um and I do think that, and you can see that when I drive through the city, I can see what I'm talking about. But uh, but but I think there's a there's an element of that in contemporary architecture where you can start to see certain forms of standardization a little a little more clearly than before. But yeah, sorry, I was rambly rambling a little bit, but um yeah.

Lyssia KatanNo, it's okay. I I I love to hear it because it it kind of it brings us back to the concept of of like, for example, I have been noticing certain spaces have been really changed by the requirements of what we need today. For example, I've seen some doctor's offices that are just stunningly beautiful, these brand new buildings. It's not like a dark, dingy waiting room. It's a beautiful, stunning waiting room, beautiful floor-to-ceiling windows, feels airy, feels clean, lots of technology of, you know, you wave your hand and the door opens, all these elements that not to say we didn't know about them, but our requirements have shifted as a society, as people.

Lyssia KatanNobody wants to sit in a dark room. You know, we're shifting and we can see that I can feel it as these new buildings pop up. We're having new spaces rethought in a way that we had never considered before.

Daniel CardosoYeah, I wanted to touch on the point that you just made. It made me think of something that I don't think had I had thought before, which is that if we think about the impact of software and digital platforms on architecture today, I think maybe 80% of that is on platforms like you know, Instagram or you know, TikTok or like places where our visual taste and culture is cultivated and shaped. And maybe that's another way of thinking of the influence of digital culture and digital systems in in architecture, the way these images circulate and become patterns, become uh things that you come to expect of spaces can be can be another element to consider.

Daniel CardosoWhat exists today that we wouldn't have without computers? Maybe very little, actually. Well, I think I think that computers, like for example, in the context of architectural coordination, right? They have made possible certain kind of to achieving certain levels of complexity in terms of form, but also in terms of systems that I think would be harder to imagine in a pre-digital age.

Daniel CardosoTwo examples. One, and it's it's a common example, the the architecture of Frank Gehry, the um Canadian American architect who well who died actually just uh last year or a couple of years ago. His architecture is often mentioned as an example of the kind of the new territories that architects are able to navigate using computers.

Daniel CardosoAnd in particular, in terms of the geometric kind of exuberance of his buildings. And if you remember buildings like uh the Disney concert hall or the Guggenheim Bilbao, they're buildings that are um quite striking because they are um using materials and geometry in ways that result in very Wow, kind of um free almost almost freeform shapes. And to accomplish that, um Gary and his company, his firm, did a lot of um kind of software development and research.

Daniel CardosoAnd in fact, for a while they uh they co-developed a software package specifically to account for this kind of material because it's quite hard if you think about it. I mean it's not hard to come up with these shapes. Like you can, you can you can you can work as a butt, it is hard, it is beautiful stuff, but it's these are beautiful sculptures, right? That are that are created in a in a in the context of an almost like an atelier kind of style, where people with a strong artistic sensibility come up with beautiful artifacts.

Daniel CardosoIt is in the translation between those beautiful artifacts and a full-scale building that the computational sort of creativity comes into play in the context of Gary's architecture. Why? Because you need to understand um uh in order to make the process of construction efficient, you need to be able to describe very precisely what in the end, what amounts to be like a like a almost like a ridiculous amount of unique parts that are all different, right?

Daniel CardosoSo we're moving away from a context of super standardized um shapes to a context of almost everything is different. And that is very difficult to accomplish unless you have a very sort of detailed process of uh tracking geometry to shape to kind of data. And that's what Gary's firm was able to kind of that's a process that Gary's firm was able to consolidate in the Guggen Handbill Bow and that has become a model for other practices.

Daniel CardosoSo I think in a way the the use of these technologies uh make things like that possible, very exuberant designs that are nevertheless tracked and managed somewhat responsibly because of the of the of the use of these technologies. Another example that is not about the form, it's not about the exuberance of the form, but it's more about the systematic or the systemic complexity of a building.

Daniel CardosoUm it would be like a like a contemporary, like a modern hospital, right? The amount of technology, the amount of systems, the MEP, HVAC, and special technical systems that go into one of these buildings today, it's hard to imagine. Probably like someone in the late 19th century looked at those things, they would, they wouldn't understand what they're seeing, just because of the amount of technology and the amount of um things that go into these buildings. And for similar reasons to the example I gave before, having uh a model that allows you that allows you to track, to coordinate, to uh contrast the different the contribution of the different traits to projects like this becomes really important.

Daniel CardosoAnd I think that's that's one that's one thing that I would say. So yes, that there are so computational tools that have come in support of increasingly complex and large architectural projects, both in terms of their formal complexity and in terms of their systemic complexity.

Lyssia KatanDo you think that there are things that maybe an architect understands by drawing that are much harder to understand by software? Or did software take all of take does top software take everything into consideration? And maybe there are things that architects didn't even think of that software handles?

Daniel CardosoRight, right. Yeah, no, the the one- my my quick answer is yes. I think drawing is uh and I and I don't think drawing will go away anytime soon. I think drawing is a fundamental way we engage with the world, we kind of reflect and think about things. Um it's a fundamental way we we communicate and and uh make sense of information and make sense of the world around us. So I do think drawing has a special place in the landscape of human cognition.

Daniel CardosoI think it's very important. And I don't think software gets close to doing that. And I I I even question the the thing, the sort of the um the expectation that it would, right? Because it's it's these are different to me, it's apples and oranges. They're they're doing the very different things. Um let me try to explain myself. I I think that when we and I share the view by the way, that when we think of drawing as a special place, as a sacred sort of place where creativity and reflection happens, I think there's truth to that, and there are people who have looked at this question through uh a more neuroscientific and cognitive scientific uh lens.

Daniel CardosoAnd so I think that the sciences there, drawing is important. It's uh it's a way to make sense of things that um has value beyond the instrumentality of practice. But but I think what that when we ask the question about drawing in architectural practice, we should also try to remember that 99% of the drawing that happened within architectural firms was not that.

Daniel CardosoThat's not the kind of drawing that is going away. The drawing that the draw that the practices of draftsmanship that became sort of disrupted by digital media were practices of recording, documenting, and detailing, which are not exactly what we associate with design or with creative practice in general, right? Drawing remains central to kind of design practices.

Daniel CardosoIt has definitely receded as a mode of architectural documentation, um, it has receded as a as a means of documenting and recording architectural design decisions. Um but I I think my my view is that drawing remains an important site for architectural and design reflection and so on. I would I would I would I would just kind of make a footnote there that that moment of design reflection and exploration that we associate with hand drawing is like 2% of what architectural practice is about.

Daniel CardosoI guess that's the only point that I was making.

Lyssia KatanThat's interesting because uh someone who is not an architect would never know that. I never knew that. So when you say that 2%, that's the part that when you sit down to design a building or a space, you're sitting down to draft or or what what's the difference? Maybe uh maybe that's what I don't understand, the difference between drafting and designing or drawing.

Daniel CardosoYeah, it's a it's a loaded question, it's a very loaded question. So let me try to see if I can uh approach it well. Well, I think that I think we need to look at this through the lens of what architectural work actually is, right? And I think there's a romantic idea about what architects do that stems from the fact that architecture has evolved in the US at least, or also in Western Europe to an extent, as one of the fine arts, right? So the architects are imagined as sort of artists who are somehow also able to kind of participate in industry and and making buildings that make sense and all of these things.

Daniel CardosoUm there's um uh this uh colleague of mine, um her name will come back to me, but she speaks about architects' double burden, right? That architects are tasked with being artists and architects as also business people. And um because the truth of the architectural industry is that the amount of time and effort that is devoted to the artistic part or to kind of the more quote-unquote creative part is very small.

Daniel CardosoMost of the time you're just trying to kind of manage that you want to take those decisions really quickly in order for the organization to be able to produce and document and communicate that design intent to the kind of building trades, to the contractors, subcontractors, etc. Right. So the shape of the business, the shape of the industry, in a way contradicts the imaginary of the architect as a an artist and as a visual person, right? So and I think that when we use drawing as a as a uh as a sort of site of the as a sacred space of architecture, we are invoking the artistic side of architecture, which again exists and so on.

Daniel CardosoArchitects are trained in that way, right? So I'm not taking that away from them, from us, right? Um, but we are living outside the reality of the industry, which is more complex and some people would say more interesting, right? Because there's there's other constraints of other forms of thinking about the practice of architecture that are less individualistic, that are less focused on the idea of the artistic vision of an individual, and more focused on the capacity of a firm of an organization to collaborate, to participate in the production of a built artifact.

Daniel CardosoAnd um and I think that those questions I think are interesting and they they connect to the questions of technology in an interesting way as well, because then you can start asking questions about what is the role of computational methods, right, in supporting the artistic vision or in supporting the collaboration structure or in supporting uh different aspects of it.

Lyssia KatanWell, that actually kind of leads me to my next question that I think is very becoming more prominent every day. Now that AI is entering architecture, does it feel like things are gonna change almost to the level of when they shifted from hand drawing to CAD? And how how is it how is all of this gonna be impacted, if if at all?

Daniel CardosoIt is really important when we talk about AI to remember, to remind ourselves that AI is not one specific thing, but it's right now we use the word AI to refer to an incredibly broad landscape of things that are happening, you know, like from computer vision to kind of rank statistical methods to reinforcement learning to different things. So I think that um that when uh and if we start looking at each of these technical frameworks more in detail, then we can start looking at specific applications and specific impacts and so on.

Daniel CardosoI do try to be very cautious with the use of the word AI because right now I feel like it's such a culturally loaded term that it carries almost like an assumption that it's something that exists out there. And no, it's it's like a marketing term to kind of to call a lot of a bunch of really different things that often have very little to do with each other. Um that said, yes, I think there are interesting potentials um in uh uh at pretty much every aspect um of the architectural kind of process, uh, from construction to kind of documentation to analysis and simulation.

Daniel CardosoAnd uh part of what we do in our group at Carnegie Mallon is to investigate those potentials and to try to develop uh try to kind of jump ahead and try to kind of steer what we think are uh responsible and humane visions for these techniques to enter design contexts and so on. Um your question was also about will this be like another change similar to the shift from hand to cat?

Daniel CardosoIt's possible. It's possible. I I don't know, it's a it's a big question for me. I study the history of that transition in the in the 20th century uh towards digital methods, and um and I see a lot of resonances with what is happening today. There was also a big effort from industry and from the government at the time to advance an agenda of digitization across industries.

Daniel CardosoIt was, I don't want to say it was top-down because it's always more complex than that, but it was definitely uh uh a project that had to do with the economic policy of the time in the 70s, to kind of try to re-revolutionize industry using computers, uh, manufacturing, construction, everything. So there was a there was a sense of hope in the with the emergence of these technologies, that using these technologies, the United States could retain its prominent kind of place in the global economy and as a military superpower, right?

Daniel CardosoSo it was that that dual thing. And there's something really interesting happening today, which is that if you read what the White House is saying about AI, it's very similar. It's they're saying the same thing. They're saying in order for the US to kind of remain, you know, the global superpower and to win the race of AI, we need to kind of completely transform the economy, right? We need to entirely rethink everything we do, basically, around these technologies. We need to build data centers everywhere because we need uh the compute to kind of uh create this new economy based on data and data processing, right?

Daniel CardosoUm so I see that that resonance and that resonance is fascinating to me. And part of me feels like there is a there is probably like a process of an ambition to have like a like a process of technological and cultural renewal around tech around these technologies, which I think it may be a very US kind of thing. Like to like this is I I see this as part of the long 20th century, right? This is happened with the with digitization, happened with the internet, now it's happening with AI, right?

Daniel CardosoAnd there's always cycles of creative destruction, so to speak, where industries change and some people are kind of um enriched, some people are not. So I think there's a I think we're living at a at a moment of uh inflection, I think, in this in a new cycle of technocultural renewal. Do I think that it will fundamentally change things?

Daniel CardosoAs I mentioned earlier in one of our your previous questions, to me it's interesting that the fundamental structure of the architectural firm didn't actually change that much, that much. It remained extremely hierarchical, perhaps more hierarchical with the introduction of computers. The distance between the architect as the person who kind of wields the pencil and does the sketching and the architectural workers became larger with the introduction of computers. So the process of increasingly hierarchical, larger and more complex, but increasingly hierarchical practices was kind of strengthened or intensified as a result of the introduction of digital systems for design in this in the 80s, 90s, etc.

Daniel CardosoWhat I think my my slightly pessimistic view is that with AI, with what we call AI, which is again so many things, but but one thing that is common is that it may emphasize that dynamic of um increasingly hierarchical, increasingly um uh practices that are increasingly monitored and surveilled, data that is increasingly not residing with you but residing somewhere else, outside your immediate jurisdiction.

Daniel CardosoThere's questions of ownership of that data, there's questions about the environmental sense that it makes or that it doesn't make sense to invest so much on these technologies that haven't really proven to be that useful yet. I think I don't know. I I part of me is hopeful and optimistic and and and see the opportunities. And part of me looks at history and says, hmm, this may be another one of these situations where where people are in the end uh yeah, the visions where there's like a very long distance between the visions and the fantasy of a future without work or a future with kind of assistance, etc., and the practical realities of work that is increasingly precarized and and surveyed.

Lyssia KatanWell, the reason that I asked this is because I mean, I'm sure you've seen it too online. There are all these AI-generated images and videos of these beautiful structures that may not even be realistic to build, or maybe it's just not even something that crossed any of our minds as simple humans, right? And now it's designing spaces that it can design based on insane amounts of data that humans just can't possibly sift through, right?

Lyssia KatanLike when we were first um learning about AI, there was this concept. This was years and years ago. When I think this was in a professor in a class in college, we were talking about there will be a day where AI will be able to analyze on, you know, any, it will be able to take in all these factors that don't have anything to do with one another. For example, like on Mondays when it rains in Pittsburgh, you know, the bakery sells more bread. You know, this like crazy, crazy data points that you would never put together.

Lyssia KatanSo is there a time or a place potentially where AI can really take in all these different variables from different sources and maybe create something that is better than what a human can do, perhaps?

Daniel CardosoRight. Yeah, that's that's that's uh one of the hopes that people invested in the project of AI uh are advancing all the time. I think the idea um by collecting these kind of crazy amounts of data that you will be able to create more more events. But I I think there's two things that come to mind. One is that um from what we see, well, first of all, there's there's a there's there are problematic questions to ask about the provenance of that data.

Daniel CardosoAnd what does it mean for a diffusion model to take on data, scraping data from the internet, from artists, designers, or architects, and then sort of remixing it and kind of returning something quote unquote new based on that, right? So I think there's a question to ask all that. I'm not a copyright kind of proponent or anything like that, but but but I think there's a there's an interesting, an interesting question to ask about the collective nature of the information or the the products that these services are creating, because they are in fact collective products.

Daniel CardosoThey are not, they are they are they are using predictive and they're feeling they're interpolating a data set and creating something that draws from different elements of the data set to create something new. And they do that in very clever and interesting ways, and some of the stuff is fascinating. But um, but I I I think it's worth asking, or it's worth pondering what does it mean that we expect kind of the the universe of visual and sort of cultural production to be sort of um I'm looking for a word here, like swallowed.

Daniel Cardosoby a by a by a machine and then sort of uh uh giving you regurgitating bits and pieces here that look a little different. I I think there's something that is interesting. I think uh interesting people have been uh trying to advance an agenda for for more consensual uses of data to kind of guarantee or or try to ensure that cultural production remains recognized and acknowledged in these systems and and so on.

Daniel CardosoAnd the the other thing is related to that is that when we think about AI and when we think about the systems that take on these systems, I think we need to remind ourselves that the system is doing something for sure, but a lot of the decisions that have an impact on what that AI system is doing are taken by people. Right? People decide what are the inputs of that system.

Daniel CardosoPeople decide if they want to sort of I don't know pirate books or scrape architectural offices websites to pirate information and put it on these are human decisions. These are very much things that that should be within the scope of a conversation about creativity, right? So so I would yeah I would I would push back a little bit against the frame of AI is doing X. No, it's like okay who is doing that who is doing what to whom?

Daniel CardosoAnd um and uh I don't know there's one example one interesting example comes to mind um there's this important Dutch firm called MVRDB they are well they're they're great designers and they're very well established and they um they have an AI kind of office within the office and uh I had the opportunity to to to have a conversation with um with uh one of the people who's who's leading that effort and and they're doing something interesting which is to have kind of uh more constrained models that are trained on the work of the office itself so that that's one way the office um sort of creates a process that is a bit more tractable for themselves.

Daniel CardosoI think it's also interesting it's not the the the the their experiments are not simply producing something based on the universe of available data but they're they're shaping the space of data that the model is using in order to create a design alternative or a design image and so on. And I think that's the kind of question that I would like us to be sort of seeing more and more in architectural firms, in in kind of research programs and so on is like how are we sculpting the spaces of training of data that these models employ in order to produce something else because that's where the creativity is.

Daniel CardosoThat's where the well there's I'm not taking away from the models themselves they're interesting but it's really shaping the inputs sculpting the spaces of possibility for those models to operate and to interpolate is becoming an interesting architectural and design question today.

Lyssia KatanYeah I I I think it's it's very important going back to your note that you said it's human doing this. A human can sit down and say to an AI model, design a hospital for me in the style of Frank Gary and AI will do it and it'll do a very good job. But going back to your point, it's pulling from other sources and sometimes plagiarizing or really pirating as you said where if you're able to educate the AI on the style of on your style on your architectural firm style and the things that you put a lot of weight on when designing a building it can help speed up the workflow.

Lyssia KatanIt can help create ideas, inspiration, but when it's used in a very distinct way where it's not like copying anyone's work or creating duplicates of someone's work in a different way, but educating it on your style and having it help you. So I think that's an important distinction to make and I'm glad you mentioned that because it is a human prompting it at the end of the day. AI is not deciding what it wants to do. It needs a human to to say design this for me.

Daniel CardosoExactly and and what worries me is that the the kinds of systems that we are seeing and obviously because these companies have a financial interest in making their their systems and keeping the systems relatively opaque right so that the provenance of the data that they're using is not transparent, it's not visible. I think there's there's a tension and there is a cultural tension about that. Because the kinds of technologies that would enable a more transparent and a more creative use of data for training models are not easy to produce and they are not the kind of business model that these companies are using.

Daniel CardosoThese companies are opaque in the mostly opaque in the in the way they deal with these things. So I think there's room for innovative platforms to create opportunities for more tractable, more transparent ways of shaping a data set, shaping a universe of possibilities for creative practitioners, architects, designers in different fields to kind of draw from that and and and more deliberately and more intentionally use the machine learning methods behind the systems in order to create something interesting from it.

Lyssia KatanWhen you look at the next 10, 20 years, what do you think will feel the most different about the way we design and build our homes and what do you hope to see in a most optimistic way?

Daniel CardosoWell it's not easy to be optimistic these days sometimes right like it's uh there's so much going on that is um that is difficult to to uh process but these technological projects are uh and I promise I will be optimistic after this but these technological projects are difficult to dissociate from from from a specific approach to to well to to politics to democracy and so on right because they are very much in alignment with certain kind of political projects and so on so so so the question of of AI the question of of architecture and design or is also part of this broader I think question that we have today about how we want to live with each other how do we want to um shape the common the common world um that we share I think so the the my optimistic view would be um would be to somehow find ways for I mentioned in my previous answer something about creative platform design or or uh more speculative ways of allowing folks to use ai systems offering more transparency about the data that they use in order for the data creation and data uh analysis to be kind of more clearly more explicitly part of the creative engagement that users have with this platform i think I think there's something interesting there I think it would be I I see a very hopeful opportunity in building systems that are that are enabling that kind of transparency and that enable um potentially kind of yeah new forms of creative practice I think before we were talking about the the kind of the first wave of digitization right in the 80s and 90s and how it kind of strengthened or intensified a certain kind of hierarchical and and sort of um yeah hierarchical modes of practice I think I wonder one of my hopes is that we somehow crack the code to use a metaphor crack the code of technologies that allow us to do something slightly different right that that is instead of instead of making firms increasingly large and increasingly sort of militarized to use another metaphor to to find technologies that allow smaller teams to collaborate and to and to do more complex projects right so architects designers engineers working together supported by technologies that are transparent um able to tackle projects of greater levels of complexity and and also supporting a more rewarding and happier kind of career right that that's that's also I think what is on my mind we in the end when we're talking about technologies for designing and for making we're also talking about livelihoods and how people spend their time and and um and I think there's a lot of room for for technologies that that that support better ways of spending your time and and more convivial more more and more congenial forms of collaborating and and and working with each other.

Lyssia KatanAnd I think beyond even just livelihood in creating these spaces for humans that can really thrive in them. Absolutely yeah right like at the end of the day why do we have all this technology? Why do we have all all of this? It's to make our lives better.

Daniel CardosoI love that that thought of in the end this is to create things that are uplifting and and uh uh and I'm I'm human. Yeah Daniel you're absolutely fascinating and I've learned so much is there anywhere that listeners can follow your work and see what you're up to thank you Lisha Notwin uh wonderful to chat um thank you for those questions uh I I hope I I was able to kind of touch on them uh um I am easy to find through my kind of academic profile at Carnegie Mellon.

Daniel CardosoUm I also have a kind of uh yeah like a website where I infrequently post my latest publications and projects which reminds me I should update that because it's been a few months.

Lyssia KatanBut no I appreciate the opportunity uh of talking with you and engage with your audience it's uh wonderful to be here yeah is there anything you'd like to leave the audience with on the last note someone listening who may have nothing to do with architecture but simply listening because they're interested.

Daniel CardosoNo I my my warmest wishes to everyone we and we'll get through this that's my last thought.

Lyssia KatanYeah I love the optimism. Daniel thank you so much for your time here today with me you've taught me so much and and I'm so grateful for everything that you're doing and everything that you're continuing to educate on. I'm really grateful for your time. Thank you so much. Thank you Lucy I appreciate it. Thank you so much for spending this time with me on Room to Think. If you enjoyed this episode feel free to follow the show subscribe leave a review and share it with someone who you think would really appreciate a more thoughtful approach to their space.

Lyssia KatanYou can find more Design Meets Psychology insights on social, in our community and definitely in upcoming episodes so you can build a better life by design. Thanks again for listening. I'll see you next time

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Uncomfortable by Design

    Nov 17, 202654 minAudio

    In this episode, Lyssia sits down with architect, researcher, and human-building interaction expert Dr. Helia Taheri to explore how the spaces around us shape the way we think, feel, and behave. Drawing from her...

    Credit the people on this episode

  2. The Art Of An Interesting Home

    Nov 10, 202657 minAudio

    In this episode, Lyssia sits down with designer, developer, and content creator Hans Lorei to explore what makes a home truly interesting. After years of walking through thousands of homes as a real estate agent and...

    Credit the people on this episode

  3. Home As A Mirror

    Oct 27, 20261 hr 5 minAudio

    In this episode, Lyssia sits down with architect, feng shui educator, and author Anjie Cho to explore the connection between our inner world and the spaces around us. Drawing from her background in architecture, feng...

    Credit the people on this episode

  4. How Trauma Shaped Our Cities

    Oct 20, 202648 minAudio

    In this episode, Lyssia sits down with architect, researcher, and author Ann Sussman to explore how the spaces around us shape the way we feel, move, and behave, often before we are consciously aware of it. Drawing from...

    Credit the people on this episode

  5. What Rome Got Right

    Oct 13, 202657 minAudio

    In this episode, Lyssia sits down with architect, leadership coach, and best-selling author Diane Van Buren to explore what Rome can teach us about beauty, architecture, and the way we live today. Drawing from her years...

    Credit the people on this episode