Grow Places

GP 48: Can Architects Stay Relevant? AI and More with Keir Regan Alexander of Arka Works

• Grow Places

Share episode

In this episode of the Grow Places podcast, Tom Larsson sits down with Keir Regan Alexander of Arka Works to ask a big question: Can architects stay relevant in the age of AI?

They explore:

  • Why architects need to move beyond aesthetics to embrace responsibility, coordination, and strategy.
  • How AI tools—from masterplanning software to generative models—are reshaping design practice.
  • The risk of architecture being reduced to “product” and how the profession can push back.
  • The evolving role of the architect as lead consultant, custodian of vision, and human in the loop.
  • What bundling, unbundling, and future industry structures might mean for practice over the next decade.

As Tom reflects, “If architects only define themselves by aesthetics, they risk giving away the real value they bring.”
And as Keir warns, “The danger is treating design like pure science—cities and places are more human than that.”

This wide-ranging conversation looks at where value lies in architecture, how the profession can adapt, and why human judgement, taste, and responsibility remain essential.

Speaker 1:

Hello and welcome to the Grow Places podcast, where we explore the virtuous circle of people, growth and place Brought to you by Grow Places and hosted by our founder, Tom Larson.

Speaker 2:

Hi, keir, how are you doing? I'm good, thank you, yeah, good Thanks for joining us. I'm going to start with a question, keir, before we dive into your intro and other things. So your purpose, according to linkedin, is helping businesses adapt to ai. Why are you doing that?

Speaker 3:

uh, I, most of my clients are architects. I am an architect, um, I I I think a lot about the future and what it would look like and who will be doing the design work, and I really want to make sure that that's architects or people who are trained in design and who have holistic cares about urban placemaking, development, sustainability and so on. I worry about a product landscape, taking that from our profession, a product landscape taking that from our profession, and that was really one of the impetus for me setting up archerworks, which was I was beginning to see a certain breed of product emerge, that um was looking quite proficient in certain things that we would be, would have, would have certainly have considered within our scope, and that was even, you know, three years ago maybe, and it's only gotten more since then. So I think there are a couple of things I looked at, certainly chat, gpt I was in there on the first day on gpt3 when it was released and I've used it pretty much every day since um and another thing that I tested in practice, which was called delve, as a master planning tool, and I sort of put those two things together and thought, um, this is a huge deal.

Speaker 3:

I want to make sure I know where are the architects. I should be involved. Why? Why am I not working for these companies? Um, are our architects going to become product, you know, providers in the future? Whatever that looks like, I want to be part of it and I want to be helping make sure that we're relevant and kept up to date and and adapting, because I think we have a tendency to observe phenomenon and be a little bit um, yeah, we'll see. Where we have a bit of a circumspect, we might wait to see how it plays out. Um, and I think that may be a vulnerability in the current climate, which is we need to overcome that skepticism and and feeling of being quite conservative about change because we've been doing it for a long time or, you know, for whatever reason, and really lean into. Okay, I'm uncomfortable, but there's something here and I need to understand it.

Speaker 2:

I need to adapt yeah, yeah, it's a really interesting term, even architect, isn't it? Because it's used in different um industries in different ways. Yes, so do you have. How do you define architect, then, within the built environment industry?

Speaker 3:

yeah, my, my co-founder, steven, who's a software developer. He says he talks about software architecture a lot and he talks about I need to architect the solution for that, and I mean that's an interesting sort of a verb to some people. I I don't know, I mean I'm not an expert on the history of our profession really, but I understand it was a sort of uh, a privileged gentleman's pursuit at one point and it was the master, a master of building and master of the arts, some sort of combination of the two, and at some point we professionalized it and put a name on it and then created loads of rules around it. When I go practices, I see huge amounts of talent and well-meaning effort being spent on quality and not, yes, there's an understanding of commercial viability, but people really genuinely care about creating places that I'd want to live in and work in and do all kinds of other things. And I think you can only become an architect if you sort of believe in that possibility. You can't have a view, a sort of cynical view, of the world. So I and I do. I mean it does tend to happen like a lot of my I'm an architect, my wife's an architect, a lot of my friends are architects or probably people like hanging out with people who are like themselves. But I generally have a lot of faith in the profession and I and I I think the work that is done in practice is incredibly valuable and I have a huge amount of respect for it and I I enjoy being in practice. I love working with architects and other designers, um, like, I like enjoying what, I enjoy working with lots of different people. But there's something about the collection of people who call themselves architects I have a certain view of the world that's optimistic, that I'm drawn to.

Speaker 3:

The term architect I don't particularly like because it's it's very limiting and and we sort of tend to think of like, oh, you know when you go to architect school and you know because you went to architect school, but there's this sort of like, well, architects do this and architects do that. Well, all these cliches and I don't like any of that stuff and I think it's really actually damaging to the profession because it prohibits an ability to change and it says, well, architects don't think like that or we don't do things like that, and that's really limiting. I don't know why we would limit ourselves in that way. We have a method for solving problems, for listening for and we have an act of drawing, is an abstraction. It's sort of you've given me a problem and, rather than solving it with a sort of maths equation, one by one, by one, I'm going to know the maths equation, but I'm also going to draw a picture. That's an abstraction of the problem and it's going to be the most elegant solution.

Speaker 3:

I mean, that's what design is at the end and that's that's the sort of real pursuit of design. And it's not like wearing black and um, whatever is our arcs, wearing horn ring glasses like this and probably other things that I do that that are like architects, um, and one of the issues with that is with, say, ai. We have to be willing to go explore slightly different domains and we have to be willing to say that software is part of our capability and that product design is something we could do. Maybe we should look at our services and find ways to turn them into products that we control and that leverage our IP.

Speaker 3:

So I think if we don't do that, we might be stuck in an old world that doesn't exist in the same way in the future, and that's a vulnerability. So so I think for me, the term architect is a sort of collection of behaviors, uh, that I respect and admire and think are important for the world. I think the title architect is a bit silly, um, and I think also I you know a lot of people get quite um obsessed with misuse of title and things. I honestly don't really care if someone, I don't really care if someone calls himself an architect, um, and they haven't got their part three, um, it's a way of thinking.

Speaker 3:

That's what makes you an architect it's not a title, and I see lots of people with the title architect who don't practice particularly well.

Speaker 2:

So yeah, yeah, I think the way that you approach problems and think about things is is super important to to value more broadly now in in in everything, not just I don't not limit that to architecture, I mean just just broadly how, how, as humans, do we kind of bring value um to the ecosystem and and I think that is very much through how we think and, yeah, you know um head and heart and you know some of these kind of analog analogies of what humans can do. Maybe machines can't.

Speaker 3:

You trained as an architect and you've gone into a parallel realm a bit like me. What's your feeling about the name architect? What do you think?

Speaker 2:

Yeah, well, I always grew up wanting to be an architect. It was kind of this social thing that I wanted to do and actually, being completely honest with you, I struggled for quite a long time kind of coming out of architecture into development about and like how do I define myself? Because it was it. Yeah, it had quite a weight to it. Yes, for some of those reasons that you're talking about, not actually it's obligation.

Speaker 3:

Right, you feel like I must think of something. I do a certain thing because you have this label yeah exactly um. So why do you want to be an?

Speaker 2:

architect, then um I wanted to be an architect because, um, in all honesty, I enjoyed I didn't have any architects in my family, um, I wanted to. I enjoyed watching grand designs and very kind of simple things like that and, um, I think, actually respect retrospectively, I actually enjoyed um, you know, seeing what it could do for people. You know the way that you could kind of create something and improve someone's quality of life and that kind of in a pithy way, it's kind of transitioned into what we do now, which is kind of you know, growing places that improve quality of life. So it kind of that human aspect about why you're doing something, yeah, um, but then I, I enjoyed design, I enjoyed you know geography doing something, yeah, um, but then I, I enjoyed design I enjoyed.

Speaker 3:

You know geography and so yeah, did you feel that, um, as with the title architect, you couldn't express certain dimensions of your personality or in your, in your curiosity, because they didn't fit within that?

Speaker 2:

yeah, well, I became. I became quite impatient, very impatient, quite quickly, um, with the whole kind of ladder approach. Um, it's very loud. I also remember seeing in meetings when I was practicing in architecture and and asking why, quite a lot in terms of like the brief from the client is this so, for argument's sake, we want this building to be housing, okay, rather than just go, okay, well, how do we make the most beautiful housing? I go why do you want it to be housing versus office, or this type housing versus that? So this kind of like brief layer became quite interesting to me.

Speaker 2:

Um, and seeing how design fits in a wider ecosystem, okay, of finance, of humanity, yeah, kind of social sciences, etc, etc.

Speaker 2:

And I probably couldn't articulate it as that at the time, but I, I, I kind of um, became more interested in design's place in everything else as opposed to just design in and of itself.

Speaker 2:

Yeah, um, and so that was, for me, kind of the frustration, I think that I was pushing on um, and then that's why I think actually development suits me better because it is broader, you are kind of um, touching on all of the different aspects without kind of being deep into any one.

Speaker 2:

Yeah, and that kind of breadth of taking the critical thinking of architecture um, both spatially but also um in terms of looking at problems, as you say, as I think the thing that I was very grateful for live yeah, very grateful, I have studied architecture, I don't regret any of that at all. Yeah, um, and I think actually some of those life skills are really important moving forward. So so for you then, kind of like if you were to summarize for people who don't necessarily know, if you were to sort of say, okay, the, the old process of architecture, if you want to call it that, the traditional practice of architecture, um, with some of these new ways of solving problems through statistical analysis, you know, which is essentially the underlying for llms like that, that way to solve problems. How do you see that going forwards?

Speaker 3:

I think if you were to take a purely data science or sort of computer science approach to design of cities, you would probably come up with some system for measuring quality with numbers, and it would be a kpis based optimization, all these sorts of words, and I think we both would agree that that isn't really what cities are and it's not what places are and that's, it's not a human. And I think our job often as architects is to be in that realm and understand things like daylighting and transport and other measurable things, but also to be expert in the immeasurable things too. And we're constantly jumping in those spheres, between those different spheres as architects. And I think, if you don't, if you just I mean, this goes back to my first response about why did we, why did I set up archaworks. But I think, if you don't, if you just I mean, this goes back to my first response about why did we, why did I set up archways.

Speaker 3:

But I think if you just live in the quantitative realm and believe that there's sort of true, false answers to things always and have certainty, that's quite precarious in design. I think that sense of certainty um makes me a bit nervous. So I I think we deal with a lot of uncertainty and we don't really know what humans will do in certain situations, and we shouldn't pretend we do and when. When architects do that, it tends to go wrong, which is modernism. I love modernism and I love the spirit of it yeah and and and when it works, it's just amazing.

Speaker 3:

But all the examples where it didn't work is where we we tried to treat it to like a pure science, and obviously it's part science, parts arts, parts, other things, parts social science, probably, and behavioral economics and all kinds of other things. So that's why we we do need, so I think, looking ahead, we're going to see a whole host of maybe, products and techniques that presume that they can solve things and they will appear to solve things and will know that is any part of, is any part of it, and and and knowing what is a good solution and having good judgment, which is which is usually grounded in experience, and then having tried things and and failed, um as architects, um is incredibly valuable. So I think, what, if I think about how architects traditionally practice, there's maybe a we could. What can happen is you know, the really hard things about design, I think are actually designing and thinking strategically, but we can think that architecture is like door schedules and Revit models and BIM execution plans and that's sort of the scaffolding of architecture. But if we get, if we dwell too much on those things, those are the things that I suspect are going to be most vulnerable to change, because a lot of those are.

Speaker 3:

If you have a sufficiently developed template and a set of rules, you can teach a system to deliver it very effectively, and that's a lot of what I do now.

Speaker 3:

That doesn't really what we say is we want people to spend time on the other stuff, but, like, basically we need to, we need to get very good at explaining the. The other thing that we do, that's not the bim, execution plans or the draw, literally the production of 100 drawings or whatever is the value, and that's the thing that gets paid for. The sort of the, the technical output, is almost like downstream of the strategy, and we need're strategic thinkers, I think architects, I think that's what we get paid for. So I think that's how I would compare maybe, a view of traditional practice to where am I go? And, um, if, if, if clients could potentially go to some future product to get a similar, like an architect, like service, then in a lot of situations I think they probably will, and we have to be quite alive to that idea and be ready for it. Um, it's, it's and this is true of all professions and it's true of pretty much all knowledge, work and probably when robotics comes true of all physical manual work as well.

Speaker 2:

Yeah, and, and I think over over the passage of time, if you want to oversimplify it, from the master builder to today which I know is an amplification the master builder was kind of hands-on every aspect of the process. Yeah, and you could say that recent architectural kind of practice has actually been kind of hands-on every aspect of the process. Yeah, and you could say that recent architectural kind of practice has actually been kind of hands-off on a lot of things. Yeah, whether that's from a legal perspective or whether that's just from an interest, this scope perspective. So how do you see the role of the sorry, the position of the architect within the kind of consultant team, the supply chain, the you know, architects, engineers, etc. Etc. Etc. Yeah, and you know we need to collaborate, we need cross-disciplinary kind of um conversations and and technologies to kind of create more value in in the process for clients. So how do you see?

Speaker 3:

that I think, um, I think you could you know, if you maybe give it a decade, I think we'll see design teams arranged slightly differently and I think there's probably an opportunity for one of the major consultants to step into a more generalized role. Yeah, can you elaborate how you think yeah might be arranged? And I mean, I remember when I was at morrison company and and you were at stanhope, we, we did, we had meetings, long meetings. Talk about design responsibility. Design responsibility is a major challenge. I think, and you're right, I think architects have gradually become resistant to taking on more responsibility, not least because, I would say, our fees are often being driven. They aren't, they don't correlate to that responsibility. On the counterpoint, I hear clients say this is all great, you want to get paid a value-based fees, percentage fees, but you need to take more, more responsibility, or or whatever it is. There's frustrations there. Um, in the future, I suspect there's a potential for the architect to generalize again. So and this is an optimistic view which will tend to get from me but, um, rather than narrowing further, narrowing and disseminating more skills and knowledge, either to manufacturers or product suppliers or to other consultants who will pop up here and there, I would be very keen to see architects expand their role and generalize more sufficiently well set up with with systems to deal with it. I think some a technology like ai can absolutely make you very capable of leading on that and providing the level of reporting that's required, um and the level of services that require is required by the legislation, and so that's an example of we can take that. We shouldn't give it away, we shouldn't wait for some other consultant to pop up. That's us now. Now it should be us, we should take it, and it's also protected by law. So that's really good because that'll be good for architects to have more services. That architects, that they have to be paid for and they're taking real responsibility and risk for that service and that's why they should be paid for it, risk for that service and that's why they should be paid for it.

Speaker 3:

If you look at other things like um, structural engineering, which is very specialized, you know it used to be. A lot of engineering did used to be done by architects. If we went back to the 60s facade engineering, um, quantity surveying, project management these are all exist. These are all roles that exist now that didn't used to exist, in my knowledge at least in the architect, used to deliver a lot of them through different specialists in their practice, and I generally think we could see those roles change over time and I think there's probably a if we, if we could describe like a um a new definition for a lead consultant, that that that's a void that I think could be filled by various people.

Speaker 3:

So I do think architects are very well placed to be there because we've got a very good overall, holistic and strategic understanding of every project. We we're in the cost, we're in the planning, we're in the technical delivery, we're with the client on the brief. We see it all um. I think engineers probably have a good position on that as well. You could easily see an engineer say well, we take, we're taking on technical responsibility and we're capable, um, actually we're going to use ai to deliver design and project lead services, which makes me feel nervous, but I could totally imagine them saying that, and so I think we're sort of I would say we're in a position where we should be jostling to get into that new position and I think, if you give it a decade, I think we won't have design teams set up the same way. Um, and I'm interested to hear what you think from the client's perspective about that I do.

Speaker 2:

I agree. I don't think we will have um them set up the same way. I think it goes back to kind of you know why? Why is each element of value in the chain and there's a kind of um. If you look back over history, typically industries kind of have tribes, either an unbundling or a bundling of things together. Yeah, so you could say unbundling is basically splitting things out into different roles, which is kind of what we've gone, we're describing.

Speaker 2:

That's what we're describing yeah, from 1960 to 20, yeah the opposite is then a bundling, things kind of coming back together, yeah, and naturally over time things team economically tend to unbundle and then bundle yeah. So if you look at, if you look at that, but the reason for saying that is I think, yeah, there probably is a natural kind of tendency now to think there might be some kind of, you know, bundling or kind of consolidating. I think the counter to that is the um is it's kind of where does the value lie? And I think actually, personally, I that you know the value of the architect I don't think lies in aesthetics and taste. I think that's part of the value of the architecture.

Speaker 2:

Okay, I actually think that going forward, that shouldn't underlie the's part of the value of the architecture. Okay, I actually think that going forward, that shouldn't underlie the value proposition of the architects. Um, I think it needs to be part of it. But often you know, um, that's so subjective um, for a start, it's harder to attribute value to um, yeah, because, um, it should come from effectively much more kind of customer analysis, if you want to kind of, yeah, call it a sort of that's the kind of product design product design is the product desirable at the end?

Speaker 2:

yeah, it's like who's it, yeah, and and is that kind of coming from a lineage of kind of architectural progression or is that coming out of sentiments, analysis, surveys, talking to people? Okay, local. So so who is, who is the kind of the, the custodian of that, or the? Or the overall decision maker in what is taste, what is aesthetics? I think that's a real question. Yeah, around that, and I think that's partly where the clients are questioning the value of the architect. Right, if the value of the architect is attributed only to that which architecture team seems to tend to say, yeah, that's what we tend to think, so I think there's a fundamental issue there. Okay, um, I actually think the value of the architect is in the stuff that the architect is wanting to kind of get rid of. It's in that technical competency, it's in that compliance, it's in um, it's in um delivering, delivering and dealing with regulation and other things.

Speaker 2:

Because if you, if you think about the value chain, if you oversimplify it, you have, you have the money, and the money is broken up because if you're an investment manager, yeah, you're actually investing pension funds money, which is actually everyday people's money, but let's bundle that up for the purpose of this into money. Money doesn't want to get involved in delivery. Yeah, it wants to be involved in money because that's where they make their money and that's their interest and their skill set. Yeah, um, if you're a developer, um, within the chain, you don't actually really want to get involved in designed or in compliance. Um, you're looking to kind of, you know, offload that risk, offload that liability onto someone else. Yes, so, and everything, everything flows downhill. Should we say so? Liability does want to flow down downhill, yeah, so, um, I actually think that the, the underlying value, is being eroded by giving away all of that to other people.

Speaker 2:

Because if it's only about taste and people are questioning a lot of times whether it is good or the taste is right and people go well, why do I need this? And I think that is a drastic oversimplification of the role of the architect and the value of the architect, given what I know being from that persuasion. Yeah, but I don't think necessarily that translates to other people. So I think, um, you know it, we are going to get much more digitally integrated as an industry. You know that is essential. It has to happen.

Speaker 2:

Yeah, but no tech company, um, or product or piece of software is ever going to really want to take responsibility, okay, for design of a building, compliance of a building to that level. Yeah, someone has to always do that. Yeah, and that should, in theory, be the, the lead designer, the master of the design team, should we say the lead master builder, however you want to frame it, yeah, and, and so I think a doubling down on that actually is where the value is for it, for architects, and then it's okay. We're saying how do we get there?

Speaker 3:

um, so you, you make you agreeing, then, I think, with the observation that we, we can what did you say?

Speaker 2:

bundle? Yeah, I think there is an element of that, but but that that be desirable? Yes, but bundling the right things? Okay, so I'm doing the things that are of value. You're procuring architects.

Speaker 3:

What would you like to see bundled? Well, I think I like this bundling.

Speaker 2:

Yeah. So I think don't get me wrong I'm not saying there isn't any value in architects for design, because I think clearly there is, yeah, but I think that the value has to be in the lead coordinator role, okay, um, because the architect does have a sensibility to to look across the hole and keep the vision in mind, yeah, and know how the different bits fit into that vision, yeah, and to hold all those different people accountable for it?

Speaker 3:

what? They're giving and whether it's whether it works with it.

Speaker 2:

Yeah yeah, so there's that on the technical.

Speaker 3:

So the lead consultant you're describing rather than lead designer?

Speaker 2:

yeah, lead consultant. So I think that there is a role there that that needs to be filled and it's for architects to want to fill. It is the kind of question, but I think there's that, and then I actually think, almost flipping completely on its head, what architects think about themselves and it's actually saying, okay, no, the design side, we're really open. We're super open, super flexible. We're led by sentiment, we're led by input, we're led by consultation. We're designing in a positive sense by committee, but through a really smart way of doing that, this engaging. You could say co-design, but but not, I don't, I'm not just saying just with communities. We've, with inputs from, um, best examples from around the world, or from what, what? Um, yeah, people in this customer base actually really think so it's kind of informed design that way. Yeah, the architect is still the custodian, but they are responding to it as opposed to that curatorial thing I think is done very skillfully by great architects.

Speaker 3:

I, I get a bit nervous about that description of design. I have to say maybe it's the architect, yeah, so the title or whatever. I I think the taste thing maybe, and I agree it's completely subjective. But then at the same time you know, I'm just going to take it to the product side imagine you're looking at the first iphone, yeah, which is sort of such a brilliantly designed object and, yes, there's all kinds of like objective reasons why it's good. But it's also just it's just on a subjective, uh, qualitative, experiential level just a beautiful product to use, feel, have in your hand, and so on.

Speaker 3:

And I think steve jobs would say if you, you know, you know there's like a Ford quote about. You know, if you ask people what they want, they just say more horses or whatever, yeah, or like faster horses or more horses rather than a car, cause they don't necessarily like. I think our job is often to come up with the car in the situation, which is maybe the answer people didn't know they needed. But you're listening and I think that curation and and and and sticking your neck out and taking a risk is very important. And if we didn't do that, I think we could have a very we we would have quite a sort of flat, underwhelming experience of design, delivery and placemaking where people aren't taking any risks, which would be sad.

Speaker 3:

So I I do see what you mean and I think I think great architects are very good at hopefully um, triangulating, like so many different needs at once and and hopefully listening as well and not just being kind of bull-nosed or bullheaded or whatever whatever the right phrase is for that. But there is also that sort of. For me, I think, like one of the key things and I don't hire architects, so maybe, maybe it isn't valuable, but certainly the sense of why I appreciate among architects is the ability to say yes, but for this particular situation we've got to do it like this and then, and when it's delivered, it just lands and it just makes sense and there's some risks. I think all the great projects I'm thinking of have taken some kind of risk that probably wouldn't have come out from someone's opinion, because they don't. They're also not expert in design. They're. They're expert in their needs. I suppose they're expert in their brief yeah, so I completely, I agree.

Speaker 2:

I think, as I say, the vision part is still the custodian. That should still be the architect, in my view. So it, it's not just the, it's just not the technical lead designer, it's the, it's the, the, the vision aesthetically designed as well, but, but we, I just think it needs to be more overplaying that.

Speaker 3:

I think there's a responsibility, like actually that's yes, that's great, but also the thing I really care about is that you are taking responsibility for things that are really important to us yeah, I would say that on that perspective.

Speaker 2:

And then a question to the industry is how are you, how are you, justifying the gut feel around the, the vision piece, if you want to call it that? So what are the inputs into that? You know, because it are the inputs into that, just um, an architectural sensibility? Yeah, or are they social sciences? Are they other things that are validating why it should look and feel this way? Yeah, so, so the inputs into that process is what I'm saying. I'm not saying that there shouldn't be the guardian. Yeah, the guardian should still be the guardian. Yeah, um, but I would. I would also kind of propose that that vision is super important and that pushing the boundaries is super important. But I wouldn. I would also kind of propose that that vision is super important and that pushing the boundaries is super important, but I wouldn't necessarily say that architects are naturally always that kind of forward looking. In that, actually, anyway, I think there's quite a lot of kind of looking backwards, yeah, and that's which is important.

Speaker 3:

You know it's got to look a certain way because, like you know, you want to live in a certain design school or whatever it is. Yeah, and we want to get published, yeah, in a certain which is totally important as well.

Speaker 2:

Um, but yeah, there's definitely, I think, a question of who is. Who is the the architect talking to? Is the architect talking to the client? Is the architect talking to the local people? Is the architect talking to other architects? You, you know, and some people do that better than others, should we say yeah, I was going to ask you about, kind of, what are the common questions or fears which come to the surface, and one of them, I would imagine, if you look at kind of you know, elton John, quite publicly recently, is kind of talking about ownership, copyright. Yeah, um, within creative pursuits, obviously, that's particularly music, but I know there's the same narrative goes on in architecture. You know, if you use an llm or mid journey and it spits up an image, how do you know that it hasn't referenced five morrison company or big projects or whatever that might be? Yeah, so how do you think about?

Speaker 3:

that well, in a small way, it has um, and you can find out what's in an image model. There's a, there's a really good website called um haveibeentrainedcom, and it's it's the layon data set, which is nearly six billion images and you, that's just the base data set that probably all of these models have trained on. Plus, they've got proprietary data on top that users have given them and and the user agreements say we can take your images and you know, mid journey is a closed source tool which will take anything you put in. So, in some statistical way, um, a photograph I took on a road trip in europe has contributed um to any image that someone pulls out of a building in that region of, I don't know, switzerland or wherever it was I was, or a design that's similar. The way that these models work is not to copy and paste. They are essentially reverse labeling, which is taking a series of words and trying to imagine the picture from the words based on the learning that they've done. So one way to avoid infringing on someone's IP is just don't use an architect's name and don't use their image as an input, and use your own image or your own drawing and your own description of what you would like to see. And then I think, in terms of reducing the risk of infringing on others, that's a really that feels to me like an honorable way to do it and a way that is.

Speaker 3:

These models couldn't exist without training on tons of images, and if there was a law that said that's illegal, then immediately no one would use any image models. But at the moment there's this big gray area, which is that the AI companies are arguing that the training itself and the prompting process is not illegal. Um, it's akin to me going to a library and reading a load of books and being able to reproduce ideas from the books, and I am sympathetic to that view as well, but I also think publishers should be paid fairly. So I think, unfortunately, it's a gray area and I think it's one that we're going to be navigating with with care. Um, the there is a case disney have just opened a court case against mid journey, um, which will be really important, and there's also another one that's ongoing between the new york times and open. Ai and and New York Times are claiming a loss of revenue because ChatGPT is able to verbatim quote paywalled material to users, which clearly shows that it's been weighted highly in the training data and also it's taking traffic from their website.

Speaker 3:

And I think these source of court cases are likely to lead to publishing deals and I imagine the game theory behind all of this from the I companies is some sort of policy, uh, some sort of publishing deal arrangement. Some companies, like perplexity, have done publishing deals first, so they're actually, when they're serving up news, they're paying the news publisher for it and that feels like a fair and reasonable thing to do and I suspect that that's where it will go and that's probably what should happen, um, if insofar as it's possible to attribute um exact authorship on a given thing and to pay the publisher for it. I think if you had a technical conversation about it with with a software developer or a computer science specialist in ai, they would say it's very hard to unscramble the egg. The idea that you could take a given image and then apportion the, the credits to all of the many millions of people whose images were we were ultimately called upon in the training is is probably unrealistic yeah, there's no such thing as true inspiration.

Speaker 2:

Isn't there's, the saying goes. It's like where's the seed of anything kind of come from?

Speaker 3:

yeah, you can't know anything without having seen things. Yeah, and and I and I wouldn't presume to, I don't know. I'm a designer and I I'm hugely inspired by other people and I think we all know we have this invisible line which is I'm inspired and I'm riffing and I'm coming up with something genuinely new and I'm excited. And then this other thing which is well, you just, you just copied that. You didn't really make any effort at all. That feels very cynical. Yeah, and we constantly flirt with this boundary, um, and I I think we know it when we see it yeah, no, it's quite interesting actually and I hadn't really thought about this until this conversation.

Speaker 2:

But what we were talking about earlier in the um the structure of the industry point and that point about as a client kind of wanting a bit more kind of understanding of of the process behind how you've kind of got to any one outcome, whether that's a technical process or whether it's kind of aesthetic process. And actually in a lot of what we're talking about here, it's actually like the value is in the process rather than the output.

Speaker 3:

Oh yeah, totally and it's not um the. The value proposition of AI a lot of the time is in a transformation or a synthesis of something. Going back to your earlier point about engagement with local people and working out exactly what the right thing is, the right fit for that place is, I suspect AI is going to really lend itself well to that. Sentiment analysis. Um, using ai in engagement settings in both text and image world, and maybe even video too, to bring ideas to life in the room so you can get a reaction. Um is the more sort of would be on the on the more adventurous side of what I've seen, I suspect well we're, we're batched that value proposition for the architect point.

Speaker 2:

Yeah, so it won't be too long before on a grow places project yeah our bridge is going to include a lot more of that. Yes, so when we're, when the architect is presenting something to us.

Speaker 2:

Um, our critique of that is not just going to be well, what's the risk? Or do we like it? Or is the planning going to like it? Yeah, it's actually going to be. No, no, this is what we think the sentiment says. So I'm just sort of checking myself.

Speaker 2:

Really, the way that I framed that and I think maybe the way that a lot of this conversation today has been framed as felt not between us but between the way we've talked about the roles of the industry, is quite adversarial, quite kind of like combative. It's like well, you take the responsibility or you prove to me that that's why you're doing that. I actually think that that's not the way it needs to work at all. I think actually it's a much more positive view that if the you know, the client or one of the entities is coming forward to a process saying, look at all these amazing insights we've got from the process, that shouldn't be seen as something threatening to a given actor within the chain. It should all be kind of enriching to the process to get to a better answer. But that's not where we are.

Speaker 3:

I yeah, and I, I agree like 100, but at the same time, people are human and and if you you're what you're describing is a shift where where you used to get this type of information and now you get this type of information in a brief and they're going to be. That's interesting you're now communicating with me in a different way. Your expectations are shifting and also you might becoming equipped with information you didn't have last time yeah and with tools that you didn't have last time.

Speaker 3:

So suddenly, when you're commenting, the thing I'm interested in seeing how it will play out with with is when, when a client can say this is, you're at this stage with a design. This is what I think we need to do, and I have an opinion and I'm able to express it with an image back to you, and it's not just words. That will become very strange because they've never had that ability before, because it's all been kept behind a bin model and many years of technical training, essentially to be able to produce a sophisticated render or visualization. You can now take an image and, you know, very, very soon be able to say you know, put a setback on it, add a story, um, change the entrance arrangement, and it'll just rearrange it on screen, um, and you'll be able to send that back. So that will be a complete shift in culture. When it happens and I agree, the spirit of collaboration 100, it also relies on good faith actors oh yeah, this could go very wrong.

Speaker 2:

It could go very right. Yeah, there's, there's definitely that, and I think actually that's where the architect does have a lot of of value, because, you know, wisdom, as you say, is kind of learn over time and that sensibility that we talk about could also be defined as kind of wisdom and experience of not just um you as the architect over your career, but the kind of collective and experience of not just um you as the architect over your career, but the kind of collective wisdom of architects that gets filtered down into architectural training, that then it's kind of viewed into you as a student.

Speaker 2:

So I think that is that is definitely really, really valuable yeah but, um, you know, I hadn't thought about it like this, but maybe actually there's a there's a fundamental kind of shift happened where historically, the architect was seen as the person who can visualize things. Yeah, and no one else can. Yeah, that's that is. That's what we're kind of talking about here. Yeah, that's changed.

Speaker 3:

And so if, if, if the architect is putting all of its value in the visualization and the aesthetic that's that's kind of what you're talking about, yeah yeah, and it's the same if you're, if you're a lawyer, yes, yeah, it's not just you're going to be fast, you're going to be sick of people sending you clauses that they've yes, that they've had written in claude or whatever.

Speaker 2:

Yeah, um and on that analogy, yeah, like the difference between a clause that is legally binding and not might be one word, yeah, and and the difference between an image that the client gets sent and that could be so subtle, but it's everything to it being a successful building or place. That nuance of sensibility is is really important yeah, yeah, um.

Speaker 3:

Well, it's, like you said, about responsibility. So you can come to a doctor with an opinion about what's wrong with you and you might be right, but he that you know, that doctor, yeah, is taking responsibility for the diagnosis and the treatment and and it's pretty um, what's the word? It's a pretty sort of uh, it's pretty serious responsibility yeah um, and I think that's true in our realm too.

Speaker 3:

I have heard software people say oh no, products will solve the responsibility problem too. Somehow you, as in a client, will be able to get some sort of indemnity and they'll work out commercially how to do that. It's an interesting idea that somehow there has to be a human entity or a business that is chartered as a professional in order to take sufficiently to accept a bundle of indemnity or not. I think a lot would order to take sufficiently to accept the sort of bundle of indemnity.

Speaker 2:

Um, or no. I think a lot would have to shift yeah, you know, in the whole system for that to occur yes, I think that's right and I don't want that to happen because I think I we want.

Speaker 3:

It's very important that the humans are responsible and that they are yeah, I think in every aspect yeah, and throughout an ai process process, we have this phrase human in the loop.

Speaker 3:

It's very important that if you're doing, you're working with, say, an agent and it's going off and doing tasks and coming back to you and then going off and doing other things, that you're curating every step and you're checking what comes back and, you know, taking responsibility for it, and that's yeah. If you're not using it right, you will have all kinds of errors and all kinds of risks and, um, that's that'd be very bad, yeah yeah, I do.

Speaker 2:

I do see a future of fully agentic companies. Yeah, um, dealing with fully agentic companies on blockchain or or other things in their own language, I think. I think that is definitely going to happen, yeah, um, but I think for what we're describing like the human in the loop thing is fundamental really.

Speaker 3:

Yeah, certainly for the professional realm, which I think we somehow maybe it's old-fashioned, but we kind of do we look towards institutions and things to like capture and validate that someone are capable of taking the responsibility, because anyone can say, oh, I'll take it yeah but that that kind of um, oversimplification, but the the kind of technology mindset would be.

Speaker 2:

You know well, facebook as a platform isn't responsible for what goes on. What goes on on the platform, yeah, so so it would be quite a change for for those kind of entities to suddenly go. No, we are going to take all the responsibility as the one goes on, because that's essentially what you'd be asking. Well, it's also responsibility for the outcome that's the issue, exactly. Yeah, so you, yeah, yeah it doesn't feel like that's kind of the approach.

Speaker 3:

Yeah someone that some definitely have to do, that startups looking at, um, technical documentation, which, um, I'm somewhat skeptical about because I think it ends up being template buildings, yeah, um, every single time which maybe is fine for certain types of buildings um, where you're?

Speaker 3:

You're in a world where you say, you know, I've got a set of row houses, here's the site, create the stage four level of information and it just goes off and does it and somehow that that entity is capable of taking responsibility for the design as well. Um, the other thing is in our favor in our whole, and and this is true of developers too, like the, the the final thing is happens in the real world. It's not a digital service architect like, it's physical and, um, it's not as physical as we'd like often, because we work on projects that never get built, but it doesn't happen in the drawings, it's, and you know, yes, robotics will probably change labor markets and I'll probably change manufacturing in the longer term, but that problem has been quite difficult to solve so far. So prefabrication has shown promise, but it's not reached sort of a mass audience yet. Going to site is chaos still. It's people digging holes, it's it's people shouting and, uh, you know, instructions to each other on cranes.

Speaker 3:

It's a really high risk, high reward type of environment, yeah, and so we should be on site learning about all that stuff and being real experts in it and I worry a little bit about the current professional paradigm where we've we have we've got sort of like an increasing sort of bifurcation of design architects and technical architects, and probably because of the technical difficulty of delivering everything at all stages to the to the sufficient level of of quality.

Speaker 3:

But, um, we don't want to separate ourselves from delivery and being on site and taking responsibility for details and and learning on site how to feed it back in at the beginning. If we do that, then I think we're more vulnerable again, um, to these sorts of things. So it's, theperson-ness of our job is very important and that includes engagement with people, being on site, being a you know it's the diplomacy of the act of architecture and you know it's a very in-person, physical phenomenon, like pitching to a client or explaining your design to a local planning authority or a stakeholder engagement meeting, being shouted at by someone who hates all kinds of development and then having a debate with someone on site about the best way to solve a particular membrane detail or something.

Speaker 2:

Yeah.

Speaker 3:

We have to do them all somehow. And those, the physicality of those is good because it's harder to alienate from us. Those is good, yeah, because it's harder to alienate from us, whereas if we live completely in a metaverse then that's not very that's, that's much easier to take.

Speaker 2:

Yeah, somehow it's not a moat, agreed, yeah, agreed, and I do think although it's probably for another conversation, but I do think the whole robotic space is fascinating and and is going to become super important to our industry and hopefully, over time, help with some of the productivity issues we're facing, labour issues and other things like that. Yeah, but maybe just to end and I've been told by other guests, I tend to do this I try to end on a question which is so nebulous and big picture that it's very difficult to answer. All right, but maybe if you were to frame kind of everything we've talked about today and then, um, not just within architecture and design team, but within the industry as a whole, you was looking kind of 10 years time, yeah, um, yeah, do you have a feel for, like, what that would look like in terms of the number of actors, the different kind of um?

Speaker 3:

I think this is really hard, I think um future gazing really hard, particularly at the moment, and I find it hard to look more than about two years ahead. And I think it would almost be unwise to make strategies for longer term, because we couldn't have known we'd be where we are today in terms of the state of the art in ai. Even two years ago, yeah, you know, we had the signal that things were going out direction and you know, people like myself have rushed towards it um with great curiosity and that so far has has has paid dividends the because the models just got better. So you know, for example, if you know we're building a workflow or something in omni chat, we build it once when the foundation model improves and we go from core 3.7 to core4, that thing that you spent time on just got better and and it's just going to get better over time. And and the question is how much better it can get. And you know we talked about agents and agentic work places and workflows. That's going to be massive. But there's this there's a big jump between where we are now and getting to that, because we can talk about agents, but actually building, maintaining and getting them working in a robust way is extremely hard.

Speaker 3:

Um, so I think we might hit a ceiling in terms of human behavior and human organizational capacity to actually accept more technology. That's kind of where I think we're at, which is there's like there's this just ballooning amount of tech and there's just too many things to look at, almost, and that's part of what my thing is trying to help people work out where, to put it. The bottleneck at the moment is cultural and it's it's organizational, which is, do we have even a strategy or a policy about why we, why we should use it like why would this company want to do a certain thing with ai? But I think, for people who get the policy and the strategy right and they start just moving in that direction methodically, they will become different businesses over time and they will change what they're selling. Um, so it might be that in and I don't want to speculate on five and ten years, but I do think you'll probably have a divergence of different kinds of organizations and you could imagine architects becoming more.

Speaker 3:

What I try and do when I talk to practices, about strategy, which is what is your super niche and what is your super strength in a market that's very, very competitive, and are there ways in which we can adopt AI and deploy it that make you even more super niche and more specialized, and that's going to be a way of protecting yourself, I think, and doing better commercially.

Speaker 3:

So maybe we'll see that you know which is further niching down, that, uh, you know which is further niching down, uh, um, but I, I, I almost, uh, you know, uh, warmly reject your question, which is, I could probably speculate on two years. I think five and ten is just almost impossible, okay, um, and what we can do is we can focus on where we are right now and our actions today and tomorrow, in the next year, and that's what we should be doing, because we can control those things and it's, you know, if you've dwelled too much on five and ten years, I think you won't get anything done and you'll just you'll be stressed and and actually just being try not to worry about stuff you can't control. Just focus on what's in front of you. Make the next step. That would be my advice agreed, and what?

Speaker 2:

what, though, over that longer period, would make you optimistic about where you think we can go with all of this um.

Speaker 3:

I think we want to see quality go up and I do agree with the productivity um critique, which is people are worried about lost product or lost human productivity being replaced by ai I. What I would be, what would make me hopeful, is if we, if we see what we've seen with every other technology and every other shift, industrial shift which is that we use it the increased capacity, to do more and to create more output and hopefully in a way that is sustainable as well. But we that we that there's enhanced productivity, that there's more commercial output, there's more building output, that we're building more and that we are able to do that because we've been enabled to get things done in in new ways here.

Speaker 2:

Thank you very much for your time. It's been an amazing conversation. Thanks for having me.

Speaker 1:

It was really fun thank you for listening to the grow places podcast. For more information, visit growplacescom and follow us at. We grow places across all social channels. See you next time.