May 20, 2025 • 53 mins
We sit down with Keir-Regan Alexander, founder of Arka Works, to explore how AI is transforming architecture.
With over 15 years of design experience in top London practices, Keir shares his journey from architect to AI consultant, revealing how AI is changing the way architects work.
We cover everything from early career influences and design school experiences at Liverpool and Westminster to the pressures of running a practice and the challenges of embracing new tech.
00:00 - Intro01:23 - Why architecture?03:21 - Early influences05:13 - Architecture school: Liverpool vs Westminster08:28 - Early career at AHMM and Fielden Fowles12:43 - Transition to Morris+Company15:58 - Navigating the split at Duggan Morris18:51 - What are your leadership values?20:52 - Leaving Morris+Company for AI consultancy23:51 - The AI awakening: early recognition of potential27:19 - AI in architecture practices: two buckets approach30:06 - Practical applications of AI in architecture33:44 - Challenges and opportunities with AI36:37 - Omnihat: a new AI startup for architects41:05 - Using AI to optimise architecture workflows43:54 - Future of architecture and AI45:28 - Advice for young architects in the AI era50:04 - The economics of architecture education today
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Anyone that tells you that they're an expert in generative
AI before 2022, before they'd even seen ChatGPT, they weren't.
Because yeah, there's phenomenaltheory around it, but in terms
of applying it to real world domain expertise and use cases,
that's happening right now. Keir Regan Alexander is an
experienced architect and founder of Archaworks, an AI
(00:20):
consultancy for the built environment.
He spent the last 15 years designing key projects in London
for a number of prominent architecture practices,
including Morris and Company. Keir was an early adopter and
advocate for artificial intelligence and generative
design within the profession andwith Archaworks he is supporting
architects and implementing AI tools, evolving their processes
for these new and growing opportunities.
(00:41):
What use cases are they finding for this and how are they
adapting their practices? I talk about the work I do in
two big buckets. There's images and video, and
then there's text and data. And those two realms have
different implications. You know that meme of like, I
want AI to do the dishes. I don't want it to write poetry
or whatever. Yeah, I think of a lot of the
text work as the dishes. Maybe the image models are more
(01:02):
like creative stuff that we'd rather do ourselves.
If I were going into practice right now, I would be trying to
find ways to become essential. But then how do you do that now?
Well, it's probably different towhat it was five years ago.
So that combination of working out what's essential and also
being flexible, it's the idea that probably the role of
architect is going to change a lot and that we need to be ready
for that. So Keir, thanks for joining me.
(01:24):
To start off a question that I ask most of the guests.
Why architecture? What is it about architecture
that keeps you going, keeps you coming back for more?
You can't switch it off, can you?
Once you have the architecture brain?
So like walking into this room, I was like registering all of
(01:46):
the finishes, all of the furniture and the various
things, and the same with the building.
And then while we were talking earlier, I was noticing the
atrium detailing and the ring beam and all of that good stuff.
So I think as soon as I got an eye for it, it's like it became
a bit of an obsession. So yeah, I just, I like all
(02:07):
kinds of visual things, all kinds of aesthetic things, all
kinds of design. But there's something about
buildings and being in them thatmakes you feel different.
I like the idea of making thingsthat make people feel different.
That's a really nice summary. Is that an unusual take?
No, no, I, I think it's for, forme, it's an interesting one
because I always had these conversations with my wife about
how other people experience, youknow, space and, and, and what
(02:29):
people see sort of thing. Because she is in no way
architectural and has, you know,no interest in most of the
things that I say. So OK, it's, it's quite
interesting. I always find to understand what
we as architects see and how we perceive things versus the
general public. And yeah, I, I think that's
always kind of an interesting conversation and, and how much
(02:49):
we as architects put into details and things like that and
how much that's appreciated by people sort of thing, what they
feel in the space. Yeah, I've, I so my wife is an
architect too. And so we can just, we just talk
about it all the time. And unfortunately our children,
I fear, are afflicted with a similar problem.
So like Theo is extremely good at Minecraft and his like 3
(03:11):
dimensional reasoning is like slightly worryingly.
I mean I fear for him a little bit, but we'll see.
Don't know if I'll be encouraging it or dissuading it.
We'll see. So just with that in mind, what
was your kind of early, you know, idea into architecture?
Why, why did you get into it? Are there any sort of early
memories or early experiences that kind of pushed you that
(03:32):
way? I do remember being taken to
certain places by my dad in particular, because he's taking
me to kind of cultural places like galleries or theaters and
things. And I remember feeling different
in those sorts of buildings. I didn't really think it was.
I didn't consider it was for me.I was sort of into music and art
and then I did quite well at science at school.
(03:53):
And then my dad, who's a doctor,sort of said you should be a
doctor. And so that's was my plan for a
couple of years. And then I wasn't very good at
chemistry. And I also, I don't think my
heart was in the idea either. And then my, I remember my
sister saying in the car one day, oh, she wanted to be an
architect and I didn't really know what it was.
And then she said, you know, like Tom Selleck from Three Men
(04:16):
and a Baby. And I was like, Oh yeah, that
does seem like a cool job. And then, and then my career
advisor said something similar. He's looking at what I was good
at. I was very into graphic design
as well and, and, and things like that.
And then I got, because I did all science at A level, I got
into engineering and then and then I had a second thoughts
(04:37):
about that and I took a gap year.
And then in the gap year I spent, I basically decided I'm
just going for architecture. So I went around the schools, My
mom came to all the schools withme and I didn't look back.
To be honest. It was, I felt like once I got
into undergrad, it was like the first time I was doing a subject
(04:57):
that was designed for me. As in, I felt finally, like all
of the other subjects had done, there was somewhat a match.
But then with architecture at 5,that was fully a match.
I was really obsessed with it and it was, yeah, a total
pleasure to study and to do. And I know you went, you did
your Part 1 at Liverpool and then you were Part 2 at
Westminster. That's right.
(05:18):
Yeah. What's the, what was the
difference between those experiences?
I know you mentioned previously when we were speaking about the
the appreciation of the studio space at Liverpool, for example,
and I think it's kind of an interesting thing when you jump
between Part 1 and Part 2 in those sort of experiences.
Yeah. So Liverpool was, I think quite
traditional in the sense of we did a lot of like technical
theory with a lot of history of architecture.
(05:40):
I had an amazing professor, David Dunster and he he was very
inspirational. Like we studied the modernist.
We were really, there was a realsort of rigor to it.
When I went to Westminster it was way more experimental and
that was what I wanted. So I think I got a good balance
(06:01):
of quite good technical background at Liverpool.
It was, it very much looked and felt like I thought an
architecture education should, the studios, culture, drawing
boards, making models fantastic.It it, it and great friends
there as well. And then at Westminster, I'd
done my Part 1 in in London and I wanted to get to a London
(06:22):
School. I wanted that.
They're much more avant-garde actually in terms of design
culture. I think, I mean, at that time, I
think like the President's medals were being won in that in
at that school. And it was, it was wild.
Like we had 6 or 7 studios. The studios all had very
different remits and different styles and you could pick, you
(06:43):
could be postmodern for, you know, a year and then you could
go parametric, you know, and they had every kind of design
school in a way in that in that school microcosm.
Sort of thing. Yeah, exactly.
So you sort of picked your tribeand I was in Gene Sillet's unit
there and we did. It was, I would say it was very
(07:03):
place based history, history based, deep research,
phenomenological, very much about atmosphere and like the
tectonic feeling of spaces. And we did, we did one project,
you know, we did the whole year was 1 project and then the whole
year of fifth year was 1 projectpractically.
And it was, it was great. They were both brilliant for
(07:26):
different reasons. I'm, I'm quite glad I got a mix
and didn't do the, I actually went back to Liverpool and I
remember David Dunst was there and he, he, he saw me in
reception. He was like, what the F are you
doing here? As in he was like, you need to
go do something else. And it was such good advice
because I wanted to go back and work with him again.
And, and his advice was no challenges like as not that it
(07:49):
wouldn't have been a challenge. It was more like do something
different. You've got an opportunity here.
You know, he actually told me togo to Switzerland and I did get
a place and that I found out toolate.
And anyway, I'm very glad I wentto Westminster.
It was good. I think that is the fascinating
thing, isn't it? Like the older you get, the more
you kind of reminisce and think about the sort of early days of
when those experimentation things can happen and, you know,
(08:10):
all the different opportunities that, yeah, I mean, you know,
the amount of people that I talked to about pushing them
into different things when you are in that early career, you
know, point of just trying everything and, you know, no
experience is kind of negative. It's just about getting a
different type of experience in architecture with that sort of
in mind, your your early career was at HMM and field and Fowles,
(08:32):
obviously quite two well known and and very well regarded
practices, but two very different sizes and work, you
know, types of work. I wonder if we could just spend
a little time just kind of looking at the differences
between those two and your differences in their sort of
experiences. Yeah, I had a HMM, something of
a baptism of fire. I went in Part 1.
I was definitely the youngest inthe whole practice.
(08:54):
I think there were two-part onesthat year.
I'm still friends with James, who was the other Part 1.
It was, HMM, very much that timewas like, I think, I think there
was 75 when I joined and then they took quickly.
It was 100 and it was a time of great growth.
It was incredibly exciting. And they put me on this amazing
(09:15):
office building. It's called the yellow building
in Latimer Rd. It was Monsoon Accessories
headquarters and it was, they were topping out when I joined
and I basically got to work on it on site.
I did, you know, doing various in the packages, package
information, subcontractor information, like all that real
nitty gritty. It just went straight in for it
(09:37):
and it was just brilliant. And then I think my take away
from working with AHMAM is that they are such a well run
machine. Like it's such an effective
practice, design practice. And they have at that time, like
my big impression it's still true.
Is, is there was a sort of system for everything like they
(09:57):
had, they'd worked it out. They were and they didn't, they
were obsessed with like taking something that they did well and
then making sure they didn't forget that in the future.
So everything you did was exemplary, as in you would try.
I mean, and I had a very good project architect called Sarah
and she was obsessive about detail and really used to pick
(10:18):
me up on stuff and red line things.
And, you know, immediately you'll sort of, you have to
really be on it. So I think the thing I got from
H Man was like really good background and sort of thinking
about the best way to do things in practice and systems and make
sure things get better over time.
And I still obsessed. I'm still obsessed with that
now. And that's sort of a lot of what
(10:39):
I do now is systems thinking with practices working out.
OK. Your practices is an engine.
Like how do you make sure it's running well?
How do you take, yeah, change something and make it run
better? That's very much what I'm doing
now. And then we filled in files.
I wanted to do something very different from H&M and they were
great. I I worked at H&M through my
(10:59):
part too. They were very supportive and
I'm still good friends with themand my wife works there still
now field and files. I knew Fergus personal friend
and I was very excited by what they did.
They did T Prem and it was yeah,very early days.
Like we were in this like very cold warehouse, freezing cold
warehouse in, in Hackney. And he, he let me come in and do
(11:22):
a couple of things and then theythey hired me as a Part 2.
And at that time we didn't have a Part 3 in the whole practice.
We were all part twos. Yeah, incredibly young practice.
Wasn't it so young? Yeah, they.
And because they just had the attitude of like, just just do
it, you know, we'll work it out.It'd be fine.
And. And it was, yeah.
It wasn't without risk. And so without stress, it was
(11:43):
very, very much, it was very stressful and hard work.
But because I went in to a smallpractice, I, you become, you
know, it's just everything, every, you have to do everything
and you become quite essential quite fast.
And then I, I was able to get projects which were way really
beyond my years and thankfully managed to deliver them well.
(12:07):
And so I did my part three thereand then it was, it was just
something about the kind of justgot to get on with it and do it
all and and we'll work out and we'll make it work.
And they and the work was very high quality and it's continued
to be high quality. And naturally, I always joke
that like the best thing that happened to field and files was
me leaving because they've just gone from strength to strength
(12:29):
ever since. Fabulous practice.
Yeah, I mean, it's even in, in just two experiences.
It's it's just such a variety oflike you say, experiences and,
and, you know, responsibilities and, and, and things like that.
When obviously your next move is, is Doug and Morris or, you
know, now Morris and company. I I mean, obviously the kind of
(12:49):
interesting thing is, is Joe is,is XAHMM.
And at the time they were probably, you know, quite a
young practice as well. So it, it feels like, you know,
the field and powers and the H&Mtogether made, made a Doug and
Morris sort of thing. I wonder what that sort of
experience then going into that practice.
And I know on your CV you talk about being employee number 20
(13:11):
and, you know, leaving that practice 10 years later as a, as
a director. Yeah, so.
It was a great experience. Yes, it was very much so.
I did big, then I did small and then and then I guess medium, I
don't know, it's 2020 is sort ofsmall to medium, isn't it?
And then we by the time I left, we were sort of more, we were
practicing up at like 50 or 60 ish kind of kind of range.
(13:35):
So it very and yeah, so I I do think Joe and I worked well
immediately together because probably because we had that
shared history of HMM, like we Ithink immediately he kind of
like I recognized the filing system when I went in, you know,
and and so and like we in those days, we were micro station as
(13:55):
well. It was, you know, it was dreamy.
And so you just sort of, I just kind of it felt like, and we had
a lot of like shared staff members.
I don't know, I just felt like very familiar.
And he at that point at 20, theywere starting to go from lots
of, they'd done lots of like private houses, lovely small
jobs and got a lot of recognition.
(14:16):
And then now they were starting to be courted by bigger clients,
PLCS and bigger organizations. And you needed that sort of
understanding of what it meant to work at more of a commercial
scale and with large clients andwith more risk.
And so I think I came in and I was given sort of attended to
work on the bigger jobs I would say in the earlier day it like
(14:37):
in my early days there. And the first project I did is
it was delivered last year. I mean, it took 10 years to
build practically. That's the architecture.
That's the way it goes. Yeah, Yeah.
So it was a good blend. And you know, if you know, in in
terms of advice to sort of younger architects, I do think
it's good to have a sense of different, there are different
types of practice and they have very different cultures and it's
(14:59):
good to just understand what they're all about.
And I wanted to be part of something that was changing and
fast and like at field and Fowles.
I mean, it was, it was I left like, you know, with, with very,
very fondly, but I, I basically,I'd seen Doug and Morris's Otis
building in Denmark Hill and I was like, oh man, I just have to
(15:23):
go work with that practice. I really wanted to go work with
them. And because I had the feeling
that they were just going to go on and like suddenly sort of
kick on to the next level. And I wanted to be part of
something that where that feeling of like rapid sort of
growth, I suppose and and experience and that was that was
a good decision. Yeah, I mean, it does feel like
(15:45):
they're, they're one of those practices of that generation
that have been able to sort of navigate the commercial aspect
whilst retaining that sort of design quality that really, you
know, stays with the projects ofspace.
Honestly, they they went throughthat sort of difficult period
and and, you know, the breaking up of the band sort of thing.
And yeah, obviously that as a, as an individual and as a, as a
(16:05):
person in that company, that must have been a really kind of
strange thing to experience and navigate, I suppose.
Yeah, it was really stressful. It was really, really difficult
period for everyone involved. John marry and like, you know, I
was in me, David Miranda and Charlotte at the time, Co
directors and we were we'd sort of we were tasked with, I
(16:31):
suppose, helping the practice through that period and sort of
navigating it and making sure that it was quite a disruptive
time for obvious reasons. And it was a lot of difficult
meetings, frankly, And that's but you know, we came through it
and it was everyone came throughit and it's just you just stuff
(16:52):
like that happens in practice. It's not it's not a it's not
always a walk in a park, you know, and you've good score is
be like, we had that sort of like, well, OK, what's next,
what's next? And you just keep working
through it. And then we came out on the
other side. Mary came out with a great
practice and she's done wonderful work.
And you know, we're still we're in touch.
(17:13):
And I think probably it was, youknow, with the benefit of
hindsight, the practices went indifferent directions.
Like you can see what Mary's done and Joe and the stuff that
we did with Joe afterwards. It's sort of, yeah, I think it
was the the founding of MorrisonCompany, which I was sort of
pivotal in along with the others.
(17:36):
It was a repositioning and sort of we were trying to be be clear
about the sort of what we wantedto do and the way we wanted to
do it basically. And it was, yeah, it was an
interesting, an interesting timefor sure.
And was was the point of I mean,obviously with the naming of the
company, Morrison company sort of thing and and opening that up
(17:56):
to a sort of, you know, was the intention was that it it became
more about the OpenTable with the rest of the directors, I
suppose. Yeah, so the and company thing
was a big deal. We actually talked.
It was the last idea we had. So there was AI won't name the
other names that we had. There was some funny ones,
someones I'm really glad we didn't do and we were called
(18:17):
something else for a few weeks. I definitely won't say that, but
we were trying to not do a name name.
And then I think Miranda said, hang on a SEC, like, what are we
doing? Like there's a reason Doug and
Morris is, has has a value. I think she, I remember she used
the word cache or and, and, and there's a reputation there.
And people do come to us becauseof Joe.
And so the thing was like, OK, Joe is the sort of spiritual
(18:42):
leader of it. And then we're the company.
And it was very much about shared voices, trying to very
much share the culture of leadership across the practice.
What were the sort of values that, you know, you as a
director really tried to sort ofhone in your role and, and
amongst your sort of staff and the sort of, you know, the
values you try to push amongst your people to get the best out
(19:04):
of that? I get I have a thing, which is
what you do is more important than what you say.
So I think the way turning up and like leading by example is
important. So if things go wrong, like how
do you take responsibility for them yourself rather than, you
know, blaming other people? I think another thing, like I
definitely made mistakes earlieron when I got to the point where
(19:26):
I was maybe overseeing other people or starting to run
projects where I definitely micromanaged people and made
them. You know, I could see that it
was not an effective approach. And it was coming.
It was coming out of probably because, yeah, when you're
stressed, that's what you end updoing.
And if you've not understood howto do it properly, like that's
(19:46):
the first major error that people make when they go, when
they reach more management or senior level or leadership level
is they think they have to control every detail and they
try and and they try and, and they end up micromanaging.
I don't, I don't, I mean, you'd have to ask my teams, but I
don't think I did that. I was very much always about
like I just trusting. We had amazing talent, talented
(20:08):
people there. So just trusting people and
being available if they needed support.
I think so, yeah. Having a culture of like doing,
not saying I think is important and I, you know, I genuinely
just really liked all the teams that I worked with.
We, we, we were lucky to have amazingly talented people.
So it does make it a lot easier when does.
(20:30):
Yes, you know, so eventually youleave Morris and company.
Yeah, onto. Different things.
Different, different things, let's say.
Was it, well, I suppose we'll, we'll come on to the sort of
incredibly exciting AI stuff andconsultancy and, and software.
(20:52):
But the thing I'm kind of interested in what was it that
allowed you to have your kind offinger on the pulse with that
and, and take that kind of opportunity because it's, it's
something that, you know, so many of us are trying to grapple
with. Yeah.
And I suppose what is it that you saw as probably an early
(21:14):
adopter to, you know, those sorts of things, those sorts of
conversations, those sorts of rumblings that made you kind of,
you know, stop and, you know, stop practicing as a architect
and jump into this kind of new field?
My, in my family, I've got people that are in tech and
people that are in AI. And so we've been talking about
(21:35):
it for about over a decade aboutwhat what it would all mean.
And I kept thinking, when's thisgoing to come to architecture?
When's it going to start impacting us?
And I would say 2022 was the year when we saw that really
happen. And I'd spent probably the five
years before the, if I had sort of, if you could describe them
(21:56):
as side hustles, I'm not sure. But like hobbies or whatever it
was that all the hobbies were around tech and software.
So I was doing sort of evening classes in Python And product
management and listening to and reading about tech entrepreneurs
and just getting kind of just following my nose down that
(22:17):
rabbit hole. Like I was, I, I'm quite an
obsessive character and I was kind of a bit obsessed in that
realm. And I did sort of several things
beforehand. Like I had this, I did a thing
with a family member on like algorithmic trading, like high
frequency trading with, with using Python to do that.
(22:41):
And then I did another, which was like a statistical method to
predict the next Michelin star restaurants in London.
Like so it was like constantly coming up with these things.
Then we did a passive house retrofit passive house tool.
I did that with two, two other guys, one of whom's gone on to
found he's got a start up that'sthat's doing that.
So there basically there were rumblings, people talk about
(23:03):
finding your niche. And my brother-in-law had said
to me around the time we were tricking around with things
together, well, your niches mustbe architecture plus like
software and tech or whatever. And I was like, oh grace, that
sounds awful. And I just couldn't really see
what that would look like. And because then I was like prop
tech, like not really interestedin prop tech per SE.
(23:23):
And then in, yeah, in 2022, we, me and Mateo, who's the manager
at Morris and Company, we were looking at thing called Delve,
which is a Google, originally Google owned product.
And it's like a master planning generative design tool.
And we'd, we'd gotten a trial for it.
And I remember two things. So the, the, the first thing was
(23:43):
it, it didn't really work. It wasn't robust.
It didn't at all do what I wanted.
And I was annoyed that I felt like the people who designed it
hadn't really understood the anatomy of buildings and of
places and everything was numerical.
They'd been turned into a statistical assessment.
And I felt that was very utilitarian and not really what
the world needed. And the other thing was they
(24:04):
were pricing it very highly. So it was, it was a yeah, yeah,
it was, it was this, it was a sort of full time salary type of
price. And I just thought I kind of
came away from that experience thinking, OK, that feels like
quite important as a moment. And then I went home and I, it
was like, say it didn't really work, but then it also did also.
I was impressed. I was very impressed.
(24:24):
I was like, oh, wow. And then I just thought, we'll
give this five years. This is going to be a huge
thing. And maybe it won't be that
product. I'm sure it'll be other things.
And I suddenly start to think, Ithink we should be, I, I don't
understand where the architects are in this like, and I went on
their website and I started looking at the job section.
And so I'm, and then I realised I'm, I'm thinking I'd quite like
(24:45):
to work on, on this product. And that's a strange thing to
think if you're a director in a company and I was doing
literally my dream role. I had a direct ship in a great
design practice as delivering projects.
We probably finished 5 or 6 projects that I'd been, you
know, pivotal in, in my last year.
And I'm going online like seeingif I can get a job in like New
York at this tech company, whichis ridiculous.
(25:07):
But it was a bit of a signal. And then very, very soon after
that, the first release of ChatGPT happened.
And I, because we'd had this narrative and discussions for so
long, you know, I was the first,I was like on the first day
logging and using it. And Dali was happening around
that time as well. And I guess I sort of looked at
(25:29):
it and I've had a few things in life where you know, when you
get these nudges and you see things and then you ignore it
and then two years later you seeit again and you ignore it.
And then a couple of years later, it's like a huge thing.
And you just think, why didn't Ijust act on the 1st impulse?
And, and, and then I started playing in my head, well, what
would that look like if I just acted on this impulse that I
have? Like I'm spending all my time on
this, why don't I do this? And so I think I was also at a
(25:54):
moment in Morrison Company wheremy major projects were coming to
a close and I was looking not atthe next year or two years, I
was looking at the next 10 years.
And we were also appointed as a sort of leadership team where,
you know, we were looking at very longer term plans in terms
of, you know, sharing company and all the other good stuff.
So it was all in. It was all in it was like
(26:15):
you're, you know, basically see you at retirement or do
something else. And I, I just decided to, I
think I, I, I said something like, I'm not sure if this is
the fish I'm going to catch. I think I need to go catch a
different one, which is mental actually, in hindsight.
And it, it really was does feel still a bit mental and it, and
(26:35):
it, but it's been an amazing pleasure the whole thing since
then. And thankfully it's worked out.
And I, you know, if it hadn't worked out, I probably would
have just, you know, gone and tried to get a job again and
carried on being just a traditional.
No, no bad thing. So I've yeah.
And then it's interesting once you spend for when you work on
something full time, you, you get, you know, quite good and
(26:56):
quite obsessed with it quite fast.
And then the other thing I thinkabout this whole space is that
really anyone that tells you that they're an expert in
generative AI before 2022, before they'd even seen ChatGPT,
they weren't because, yeah, there's phenomenal theory around
it. But in terms of applying it to
real world domain expertise and and use cases that's happening
(27:18):
right now, you know, and it's we're two or three years in
basically. So with that in mind, obviously
your company ARC works, your consultancy, you work with other
architects, other practices navigating this whole thing.
Yeah. What is the industry?
You know, how are the people that you're working with, what
(27:41):
use cases are they finding for this?
And I suppose how are they adapting their practices to
these things? And I suppose at a time when so
many of us feel like, you know, we want to get on this train and
and we don't want to miss it sort of thing, I guess it's kind
of like understanding what thoseopportunities are.
So I talk about the work I do intwo big buckets.
(28:03):
There's images and video, that'sone bucket.
And then there's text and data. And for the images and video,
we're talking about diffusion models mostly things like mid
journey, stable diffusion, flux.In the text world, we're talking
about large language models essentially.
And those two realms have different implications for
practice. So the the image realm is I
(28:26):
guess there's that, you know, that meme of like, I want AI to
do the dishes. I don't want him to write poetry
or whatever. I suppose I think of a lot of
the text work as the dishes. It's the more the administrative
supportive tasks, project support tasks and maybe the
image models are more like the not dishes, whatever that is,
the creative stuff that we'd rather do ourselves.
(28:47):
What I find is every practice falls in a different place.
So some people want to do everything and they are all in
on let's change the way we process all written information
in the practice. And also at the same time, let's
like completely revolutionize our design process.
And it does go to the core of your design culture if you want
(29:08):
to do it. And then other practices, you
know, I'm very lucky to work with a lot of amazing practices
with, you know, international reputations where they are more
circumspect, particularly about the design side and which I
completely understand. And that's kind of one of the
reasons I set up Archer works, which is I could see this thing
was going to go straight, straight to the core of our
(29:29):
practice culture. And they are not really, some of
those are just saying we're justnot going to touch image models
because we don't really want, wedon't really want to bring AI
into that part of our world, butwe'd love to have help with.
You know, the stuff we don't enjoy.
The stuff we don't enjoy and thestuff we don't enjoy would
include things like regulatory audits, briefing trackers,
(29:52):
meeting minutes, condition trackers, all kinds of
administrative project management tools which are
hugely aided. Now if you if you can learn how
to do it in the right way by theuse of these phenomenal
particularly, yeah the latest reasoning models that are coming
out. If you use them in the right way
(30:12):
with the right setup and you andyou know how to, how to avoid
certain risks, they are astoundingly effective.
And I'm, I continue everyday I'mshocked by what I see is
possible. And one of the reasons I love
doing workshops and, and meetinglots of different practices is
(30:33):
in every single workshop, someone comes up with a
different thought and they'll say, oh, could I try this?
I've just been trying to do thison a project and suddenly you're
taking something that exists in the world, technology and you're
actually applying it to a domainwith, and you're applying your
expertise and you know your field.
So I know what I understand the risks of being on an
(30:54):
architecture project. I understand the sort of
imperatives and the stresses of it and also the opportunities.
So your, your job then is how doI take this thing that's out
there and fit it into that worldin a way that's actually not a
pain in the arse and, and actually something that people
are so grateful for. So when it's working well,
people genuine, you know, you get texts like people just
(31:15):
saying, Oh my God, this is unbelievable.
I'm so delighted with this thing.
Or they're really happy to show you something that they made.
And so that's what I'm saying. I think I talk about heatmap and
AI heatmap for every organization.
And the idea is that your particular sensitivity to this
technology depends on the culture you're practicing, how
you operate already. And, and so you can look at a
(31:37):
practice and you can say, where are the opportunities?
Where do we, and then where, where do we have like where do
we want to be really aggressive and where do we want to be in
more circumspect? And you'll, what will emerge is
a picture of opportunities. And I call it a heat map.
And essentially you just work. There's a way of looking at it,
analysing what you'd like to do 1st and you just go straight
into those things. And you, you, you have a quite a
(31:59):
sort of objective way of saying,well, there's 100 things we
could do, but what are the five things we should do in the next
6 months or so? And you just work at it like
that basically. So with the, with the pressures
of the industry and we all know this, this idea of racing to the
bottom in terms of our, our feesand our value.
Do you see the AI opportunity asa, as an option for efficiency
(32:21):
and, you know, the idea of tackling these problems or I
suppose with the rest of the world, you know, speeding along
on that same thing, Are we just going to be in the same position
that we were all along? I wrote something last year
which was we need to talk about fees and we do need to talk
about fees in general, like outside of AI there is, I don't
(32:44):
like talking about, I find the word efficiency slightly crude
when we talk about AI because it's so 1 dimensional.
I think I want practices to become more effective and I also
want to make sure that any benefits that they get that they
actually feel themselves and that they don't just pass them
back to the market because I think we've been giving value
(33:05):
away to the market for too long.I'm worried if we just are
particularly worried about the idea of time saving, if we are
moving our fee structures to a time standard, time based
standard. I in the 15 years or so since I
sort of did my part to the culture of fees has shifted from
percentage based value based fees to lump sum fees derived
(33:30):
from resource plans or and, and unfortunately those are usually
underestimated. And I see that as being a major
problem for undercutting as well.
And the problem with that is if you make a time saving and you
use it as a competitive measure against another practice, then
what you will do is you will drive our fees into the ground.
And that's how we lose at this whole thing.
(33:53):
But we, we are uniquely positioned to discharge the
responsibilities and duties of architects, particularly with
the Building Safety Act. We're, we are essential.
I think I still believe that we need to understand our value.
And with the, with the rise of AI, we need to be the best at
using it and not, and we need tonot become at risk from some
(34:14):
product that pops up that promises to do what we do.
We should be better. We should be way better at that
at, at any product than that. And we should be paid well for
it. And so I, I sort of push back.
I try not to talk about time saving.
Of course there is time saving involved.
(34:34):
If you're doing it the right way, you should find the quality
goes up and you should also findthat your enjoyment of the tasks
way goes up. So you, you get great pleasure
and satisfaction when from say sitting in a meet.
You know, the big, the big one that I'm enjoying at the moment
is transcriptions are just so phenomenal.
So it and there's ways of takingtranscriptions in person that
(34:56):
that recognises different voices.
So you know who's speaking and when and taking those and
basically producing your, your written records, your reports,
updating trackers, doing all kinds of things with that raw
material. So the meeting becomes
incredibly important and what you say in the meeting is
consequential. So you start, you're more
declarative when you're communicating.
(35:16):
You start making sure, OK, so what's the action?
And what have we decided or whatdidn't we agree?
And you start saying these things out loud because you know
it's going to get picked up and it's going to end up in the
tracker. It's going to end up in the
minutes or it's going to end up in the planning statement or
whatever it is that this thing'sgetting put into.
And that means that you work better because you, you, you
work. You have better meeting hygiene.
(35:37):
You have better communication. What fascinating feedback loop
to say that we're changing our behaviour because of that.
Oh yeah, we definitely will do. Yeah.
I mean, I'm seeing it already and and I actually in that case
I think it's a positive. Change.
Yeah, completely. There is an interesting sort of
counterbalance to that, which isif if an AI is listening, then
we're we could be on the hook for all kinds of things in terms
of indemnity. So, and if you're a cynical
(36:03):
project manager or, or contractor, then you know, you,
you might want to use transcriptions.
And I know that they're already thinking about doing this and
essentially to pursue architectsand make claims more
effectively. So maybe what happens is the
culture of meetings, technical meetings become extremely,
they're just procedural and no one says anything of any
consequence. And then what happens is you'll,
(36:25):
you know, I'll take that away iswhat people say.
I'll just have to confirm that. And of course, the real
conversation happens somewhere else, so we'll see.
With with with the whole, you know, changing attitudes around
AI and how practice is using it.I wonder if we talk about your
sort of new start up Omni chat, if you can give me a kind of a
(36:45):
brief intro into what the origins of that are and and why
you decided to go down that route.
Yeah, I mean I'm incredibly excited about it.
So I met Stephen, who's my Co founder on a on an AI panel at
Arab an event and we, we, we sort of clicked.
We immediately went looking for innovate UK funding for for a
(37:08):
project. We didn't actually get the
funding, but it's interesting the idea we had, or that's
really Steven had, I shouldn't take credit for it, is what
we're going on. We're going to probably be
implementing in the next year orso.
And that's a more agentic approach to an API based agentic
approach to building different automated workflows for
practices. But he sort of said to me,
(37:30):
we're, we're working in a similar space.
So he's, he's software development background AI
specialist and he's working withsome, some practices.
I know. And he's, he said to me, well,
it's like your biggest challengeat the moment.
And I said, you know, I, I thinkthere's huge potential in large
language models, but I'm worriedabout privacy and I'm worried
(37:50):
about access because what I was,what I see is the private tools
you have to pay for. When I say privacy, I mean
protection from training. You still have to trust third
parties. These are not private like
airlocked LLMS. But you're at the moment more
than if you do a survey in a practice, you will find more
than 50% of people using free chat BT and they are just giving
(38:11):
all the data away for training. So they're making the models
better. All of the stuff you put into
ChatGPT, all the attachments, whether you own them or not.
And you probably don't own a lotof them.
Probably they belong to other people.
You're giving them away. They might be confidential, they
might be GDPR sensitive. And you know, there's, so
there's a, there's a data issue there.
There's also, there's a, there'sa quality issue.
(38:34):
I I did a test last week, just the way you prompt will
massively change accuracy. So if you drop an attachment
into freech HBT, if you drop theEU act into freech HBT and you
ask it to retrieve all of the articles it does about it gets
about 4% accuracy. Which is shocking.
If you do the same in GBT team using attachment methods, you
(38:57):
get about 16% it, you know, getsa few of the articles doesn't do
a good enough job. But if you do this method we
endorse could unabridge prompting you get it.
It retrieves every single article 100% accurate.
And that's using the same model.It's just interacting with it in
a different way. So I was talking to Stephen
about this and he said, well, I've got this prototype and
basically as soon as I saw it, Isaid like, this is what everyone
needs. So we implemented it first for a
(39:19):
big practice in London and now it's the main tool I use for all
use for all of my training. And the main idea is it's we've
got all of the language models that we think are any good
connected up inside it. OK, So it's one interface that
you learn and you can flip the models at any point.
So you can do work in Claude andthen you can have GBT come in
(39:41):
and check the check the homeworkof Claude and see how it did, is
it accurate? Did it make any mistakes?
And then you can maybe use a different model for data
analytics or whatever. So you've, you sort of have
configuration set up for different tasks and you, you get
to know when you get very into it, you get to know which ones
are good at what and they have different personalities.
They're sort of like your team and you have to decide who in
(40:01):
your team you'd like to pick forcertain tasks.
So I think of GBT 4 O as like a very, almost like an engineer.
It's like very sort of analytical, very quite
utilitarian. Whereas clawed, the clawed
personality is much more able tohandle nuance.
And to, to, to deal with uncertainty and to use and to
(40:24):
adopt style of language. So you would, you would use
those two types of characters for different things.
So that's why it's called Omni chat because we have these,
these Omni models basically. So it's multi model.
And then the other thing that it's allowed us to do is to
standardize workflow. So when we're working with
practices, I'll build them a solution for, I don't know,
getting basically taking something from one position into
(40:45):
another position and and outputting material that's
essentially usable. And we we set those up as pre
made templates. So you just click a button and
they run. And yeah, we're starting now to
implement agents. I don't know if you want to talk
about agents, but agents can do sort of end to end tasks now and
link different tools up with AP is so that's also going to sit
(41:08):
in Omni. So that's becoming extremely
busy. We've got clients in Denmark,
Ireland, Germany and the UK and.So that's almost like having a
staff member that will go and doa task for you and put the
pieces together. It's exactly, it's well, you
still have to say. What I like to encourage is that
(41:28):
people learn to prompt really well.
And then they essentially have asuite of workflows that they
have, they've built themselves using this method.
And it'll be like, it might be quite a big job that the thing
does, or it might just be a small section of a job that you
do regularly, but it just reallyhelps you with that part of it.
And rather than doing it in thisad hoc way, I describe it like
(41:49):
what most people do in GPT is they kind of open the fridge and
decide what to cook for dinner. And it's like, oh, I've got this
and this and I'll just throw them together.
With this method, it's much morelike you've got a recipe and it
executes the same and you're just changing the ingredients
each time a little bit. So there's a couple of things
that change, but most of it stays the same.
And so you're looking around thepractice constantly for like,
what are these things we do all the time that actually, yeah,
(42:12):
they're different. But there, there is a structure,
there is a template, there is a method that we follow.
Those things are perfect for this.
So there's, I think we worked out there's something like 80
typical, there's, there's about 80 things you can do across a
whole practice. And that's just as a starting
point that we, that we've started to do and that.
(42:33):
So a lot of what we do is that, and that's becoming more and
more of my efforts and, and focus.
And it's also made me become waymore specialised on large
language models because I'm working with Stephen and Stephen
reads every single research paper.
He's every new model release, he's trying it on the day and
we're able to connect new modelsvery quickly.
And that whole, the process of working with Stephen has been
(42:57):
just the most amazing education for me as well.
And I'm, you know, a lot of my clients benefit from that
directly. So yeah, I, I probably, in terms
of my, where I see opportunity, I love the image realm.
I'd, I find it, it this sort of,it kind of interests a different
part of my brain. It activates the sort of the
part of the brain that you enjoyusing for sketching and that
(43:18):
sort of part. But the thing I think is really
transformative is probably the language models work, yeah.
Yeah. I think just sitting here
listening to you talking about the processes and understanding
the different processes, you know, an architecture practice
goes into, you know, I am thinking about those early days
at HMM that you're talking aboutwith the systems and stuff.
And that's right. I think it's a really
(43:39):
fascinating thing to you. Know Go go.
Full circle with that and and you to be looking at new
technologies but solving the same problems that you were
thinking about. I like that you closed that loop
because I, I, I think that is about right.
Actually, I think there's a, there's a method of thinking
about businesses as systems and,and the within a system, if you
(44:00):
can observe, can you observe it objectively and say, can you
describe it to someone else? Like how do we do bids?
Can you describe it or is it totally made-up every time?
If you can't describe it, it's not a system and you need to be
able to draw it. And then you need to say, OK,
well, there's 10 parts to that process or 20 parts to it,
whatever it is that you're doing.
Like maybe you're bidding for awards or whatever it is, though
(44:21):
you could probably describe it and draw that process.
And then you should, then you can become obsessed about
improving the system. And then it's almost like
you're, you're constantly refined, refining this overall
vehicle or engine or whatever, whatever the right metaphor is.
And that's how I think about practice.
And I think maybe it did come from HMM.
Or maybe it's just the way I tend to think about problems.
(44:43):
I don't know. With with that in mind, just to
kind of close up the conversation, what do you think
people can kind of take away from your career journey?
Considering, considering that like we've just kind of alluded
to you, you've kind of come, youknow, you've gone full circle
and then, but now are, you know,incredibly sort of future
(45:03):
looking and, and trying to solveage-old problems.
But in a, in a, you know, clearly sort of contemporary
manner. What is, you know, if, if you're
a young architect starting in practice or, you know, even, you
know, you're a young person starting into architecture, you
know, what is that? What is that journey going to be
looking like in the future? And you know, what are those
(45:26):
sort of takeaways that we can get from your journey, I
suppose? I find this a bit hard actually.
I run it really hard to think about the future right now in
terms of career, career advice for other people.
But I think maybe what we can besure of is there's going to be
lots more change. And I think there's maybe a way
of thinking about your career ina more flexible way.
(45:50):
I've probably had a problem withthe title architect for some
time, as in I'm very proud of the fact I'm an architect.
But it's also, it's something ofa straight jacket because there
is a certain way that we think about what we must do each day
and how we must do it. And that is, it's not, I don't
(46:14):
know that that's, I think that that's sort of, I'm an architect
and that's my ego and it's, and I define myself in that way.
I find that to be limiting and not ready for the what needs to
be much more flexible approach. So I, I think the skill set for
architecture is phenomenal. What we're really good is like
(46:35):
quite abstract problem solving, like we are taking very complex
things and weaving an answer andalso communicating that story
from beginning to end to clients.
And those sorts of skills are universal and those are the
things they're not going to be alienated from us.
Like if we can communicate with clients, connect, understand
their problems, find solutions to them.
(46:58):
Building's incredibly stressful.Also, it happens in the real
world, like with like in with like muddy holes and people
digging stuff on site. Like it's a, it's a messy
business and we need to lean into that sort of those parts if
we want to make in terms of AI and disruption.
We have that's where we bring value.
I think is, is our, the, the ability to come in and, and see
(47:21):
the problem, identify and, and be able to weave a solution and
then communicate the solution. And then you're, we're almost
like producers rather than like artists.
Like we, we sort of we part produce a part director or
something and we're building a vision and we need to bring
people in at the right time. But I also think like, I don't
that we've got this thing where maybe it's our training, but we,
(47:42):
we sort of feel like we need to be like the best.
We have to be totally like all rounders at everything.
And I also think that's a bit ofa problem.
So it just in terms of if you, if I were going into practice
right now, I would be trying to find ways to become essential.
And I think that's probably, that's probably true of any time
in, in, in human history. If you want to do well, you
(48:05):
basically have to work out in any situation, how do I become
essential and indispensable? And that's that's true now.
But then how do you do that now?Well, it's probably different to
what it was five years ago. And I think so that combination
of working out what's essential and also being flexible to the
idea that probably the role of architect is going to change a
(48:26):
lot. And that we need to be ready for
that in order to have a flourishing and happy career.
And not be too rigid about like,I don't know if you have the
same thing. I have this sort of monograph.
It's my joke is like in my head,I have a monograph and it's like
the, it's the like career architect name on a book.
(48:48):
Here's My Portfolio. I've delivered these projects.
Aren't they amazing plan sections, elevations and who am
I trying to impress? Like mostly other architects.
Like that's all of the things I just said.
Most architects that I know, I think are vulnerable to that
fantasy and I definitely am, butit doesn't serve us well and
it's very limiting because actually there's all kinds of
(49:10):
other opportunities. If you just say, well, how can I
be essential and where can I provide value?
So like my, it may be my thing is, you know, when you're
looking at the next, your next business opportunity or your
next move, it's like you could think, well, what do I want for
myself? Or you can think?
And in my situation, I probably would have come away from Morris
(49:30):
company thinking I should set upa practice and, you know, set up
my own practice, architecture practice.
Maybe it's AI driven. And then I thought to myself,
I'm not sure we need another practice right now.
It feels like there's a lot of really good practices.
So what maybe the world needs isfor me, or what I could do that
would actually be more valuable is go help focus my time on
(49:50):
becoming really good at something and then go help lots
of other practices with that. And that's a different way of
looking at a problem is like, rather than what am I trying to
impose on it? It's like, well, what does, what
do they, where could I be valuable?
So that's I would say that to young, young up and coming
architects. But I do think it's a funny time
and I I'm, I'm, I'm slightly, I don't.
(50:13):
What do you think about the future of the what would you be
saying to architects coming up now?
I mean, you're seeing all this stuff too.
Yeah, I mean, like you say, it'sit's something that we're all
trying to navigate and I think it's, it's an incredibly
difficult time and, and particularly, you know, even
just choosing to study architecture at the moment is a,
you know, such a, a big, you know, decision to make and, and
(50:33):
things like that. So, yeah, it's, it's just
complicated, isn't it? It's.
A huge undertaking financially. Crazy undertaking.
It doesn't make any financial sense anymore.
And that's, that's something that's changed in my career.
You know, I, I, I was, I met, I went, I went to the Arsenal last
night and I met with a friend from Part 1 and he went and did
(50:54):
contracting. And lots of my friends actually
have ended up going and doing different things later in their
careers. But we were reflecting on our
starting salaries at Part 1 and how and how proud we were.
And you know, we look at the numbers now and they're not that
different. And yet we know everything costs
about 10 times more. So I, I find that to be
incredibly unjust and upsetting for, for up and coming
(51:15):
architects. They need to be angry about it
as well. Like it's not acceptable.
And I think the, the slight problem, I mean, it's a, it's
fee driven. I mean, practices are not making
our fees, they're not making ourprofit.
I think there's maybe a sense of, I don't know that those
things are felt by incumbent owners and leaders and practices
(51:36):
when they've been running them for many years, maybe maybe many
decades and they are fairly welloff and they're probably doing
it for the pleasure of doing it and it's less essential for
them. And I think that it got worse
during COVID in just in terms ofthe disparity of perspective.
And it's a difficult time. We need to sort it out.
I don't know how we sort it out.It would, it would really help
(51:59):
if, you know, our professional institutions were more direct on
this and took a lead. I think they've been a bit
agnostic about it for too long. And I think I don't know what
happens. I mean, basically the only logic
at the moment, you need to be pretty privileged, I think to,
to look at the economic equationfor architecture and think that
(52:20):
it makes any sense to and I probably am, I'm probably lucky
to have come. I had like my parents were
always like, you know, do what you love and, and which was
lovely. But that's the sort of that's
like generational advice. I'm not sure that I'm not sure
that's so appropriate. Not anymore.
(52:40):
In a way, it's like it's, it's, that might be quite an
irrational thing to do now, you know?
Completely and and you know, particularly given now the
government is talking about moving, you know, the
apprenticeship schemes away, youknow, the funding away from
apprenticeship schemes, which again, it was quite a hopeful
thing for. Obviously it was never going to
solve the problem, but it was a hopeful kind of thing.
So yeah, it's a difficult time and I don't think anyone's quite
(53:02):
got the answers. But like you say, at least our
professional bodies pushing and and having those conversations
might help. They really need to, yeah.
Brilliant, absolutely fascinating conversation.
Thank you so much for joining me.
Thanks for having me, it was very fun.
Anyone that tells you that they're an expert in generative
AI before 2022, before they'd even seen ChatGPT, they weren't.
Because yeah, there's phenomenaltheory around it, but in terms
of applying it to real world domain expertise and use cases,
that's happening right now. Keir Regan Alexander is an
experienced architect and founder of Archaworks, an AI
(00:20):
consultancy for the built environment.
He spent the last 15 years designing key projects in London
for a number of prominent architecture practices,
including Morris and Company. Keir was an early adopter and
advocate for artificial intelligence and generative
design within the profession andwith Archaworks he is supporting
architects and implementing AI tools, evolving their processes
for these new and growing opportunities.
(00:41):
What use cases are they finding for this and how are they
adapting their practices? I talk about the work I do in
two big buckets. There's images and video, and
then there's text and data. And those two realms have
different implications. You know that meme of like, I
want AI to do the dishes. I don't want it to write poetry
or whatever. Yeah, I think of a lot of the
text work as the dishes. Maybe the image models are more
(01:02):
like creative stuff that we'd rather do ourselves.
If I were going into practice right now, I would be trying to
find ways to become essential. But then how do you do that now?
Well, it's probably different towhat it was five years ago.
So that combination of working out what's essential and also
being flexible, it's the idea that probably the role of
architect is going to change a lot and that we need to be ready
for that. So Keir, thanks for joining me.
(01:24):
To start off a question that I ask most of the guests.
Why architecture? What is it about architecture
that keeps you going, keeps you coming back for more?
You can't switch it off, can you?
Once you have the architecture brain?
So like walking into this room, I was like registering all of
(01:46):
the finishes, all of the furniture and the various
things, and the same with the building.
And then while we were talking earlier, I was noticing the
atrium detailing and the ring beam and all of that good stuff.
So I think as soon as I got an eye for it, it's like it became
a bit of an obsession. So yeah, I just, I like all
(02:07):
kinds of visual things, all kinds of aesthetic things, all
kinds of design. But there's something about
buildings and being in them thatmakes you feel different.
I like the idea of making thingsthat make people feel different.
That's a really nice summary. Is that an unusual take?
No, no, I, I think it's for, forme, it's an interesting one
because I always had these conversations with my wife about
how other people experience, youknow, space and, and, and what
(02:29):
people see sort of thing. Because she is in no way
architectural and has, you know,no interest in most of the
things that I say. So OK, it's, it's quite
interesting. I always find to understand what
we as architects see and how we perceive things versus the
general public. And yeah, I, I think that's
always kind of an interesting conversation and, and how much
(02:49):
we as architects put into details and things like that and
how much that's appreciated by people sort of thing, what they
feel in the space. Yeah, I've, I so my wife is an
architect too. And so we can just, we just talk
about it all the time. And unfortunately our children,
I fear, are afflicted with a similar problem.
So like Theo is extremely good at Minecraft and his like 3
(03:11):
dimensional reasoning is like slightly worryingly.
I mean I fear for him a little bit, but we'll see.
Don't know if I'll be encouraging it or dissuading it.
We'll see. So just with that in mind, what
was your kind of early, you know, idea into architecture?
Why, why did you get into it? Are there any sort of early
memories or early experiences that kind of pushed you that
(03:32):
way? I do remember being taken to
certain places by my dad in particular, because he's taking
me to kind of cultural places like galleries or theaters and
things. And I remember feeling different
in those sorts of buildings. I didn't really think it was.
I didn't consider it was for me.I was sort of into music and art
and then I did quite well at science at school.
(03:53):
And then my dad, who's a doctor,sort of said you should be a
doctor. And so that's was my plan for a
couple of years. And then I wasn't very good at
chemistry. And I also, I don't think my
heart was in the idea either. And then my, I remember my
sister saying in the car one day, oh, she wanted to be an
architect and I didn't really know what it was.
And then she said, you know, like Tom Selleck from Three Men
(04:16):
and a Baby. And I was like, Oh yeah, that
does seem like a cool job. And then, and then my career
advisor said something similar. He's looking at what I was good
at. I was very into graphic design
as well and, and, and things like that.
And then I got, because I did all science at A level, I got
into engineering and then and then I had a second thoughts
(04:37):
about that and I took a gap year.
And then in the gap year I spent, I basically decided I'm
just going for architecture. So I went around the schools, My
mom came to all the schools withme and I didn't look back.
To be honest. It was, I felt like once I got
into undergrad, it was like the first time I was doing a subject
(04:57):
that was designed for me. As in, I felt finally, like all
of the other subjects had done, there was somewhat a match.
But then with architecture at 5,that was fully a match.
I was really obsessed with it and it was, yeah, a total
pleasure to study and to do. And I know you went, you did
your Part 1 at Liverpool and then you were Part 2 at
Westminster. That's right.
(05:18):
Yeah. What's the, what was the
difference between those experiences?
I know you mentioned previously when we were speaking about the
the appreciation of the studio space at Liverpool, for example,
and I think it's kind of an interesting thing when you jump
between Part 1 and Part 2 in those sort of experiences.
Yeah. So Liverpool was, I think quite
traditional in the sense of we did a lot of like technical
theory with a lot of history of architecture.
(05:40):
I had an amazing professor, David Dunster and he he was very
inspirational. Like we studied the modernist.
We were really, there was a realsort of rigor to it.
When I went to Westminster it was way more experimental and
that was what I wanted. So I think I got a good balance
(06:01):
of quite good technical background at Liverpool.
It was, it very much looked and felt like I thought an
architecture education should, the studios, culture, drawing
boards, making models fantastic.It it, it and great friends
there as well. And then at Westminster, I'd
done my Part 1 in in London and I wanted to get to a London
(06:22):
School. I wanted that.
They're much more avant-garde actually in terms of design
culture. I think, I mean, at that time, I
think like the President's medals were being won in that in
at that school. And it was, it was wild.
Like we had 6 or 7 studios. The studios all had very
different remits and different styles and you could pick, you
(06:43):
could be postmodern for, you know, a year and then you could
go parametric, you know, and they had every kind of design
school in a way in that in that school microcosm.
Sort of thing. Yeah, exactly.
So you sort of picked your tribeand I was in Gene Sillet's unit
there and we did. It was, I would say it was very
(07:03):
place based history, history based, deep research,
phenomenological, very much about atmosphere and like the
tectonic feeling of spaces. And we did, we did one project,
you know, we did the whole year was 1 project and then the whole
year of fifth year was 1 projectpractically.
And it was, it was great. They were both brilliant for
(07:26):
different reasons. I'm, I'm quite glad I got a mix
and didn't do the, I actually went back to Liverpool and I
remember David Dunst was there and he, he, he saw me in
reception. He was like, what the F are you
doing here? As in he was like, you need to
go do something else. And it was such good advice
because I wanted to go back and work with him again.
And, and his advice was no challenges like as not that it
(07:49):
wouldn't have been a challenge. It was more like do something
different. You've got an opportunity here.
You know, he actually told me togo to Switzerland and I did get
a place and that I found out toolate.
And anyway, I'm very glad I wentto Westminster.
It was good. I think that is the fascinating
thing, isn't it? Like the older you get, the more
you kind of reminisce and think about the sort of early days of
when those experimentation things can happen and, you know,
(08:10):
all the different opportunities that, yeah, I mean, you know,
the amount of people that I talked to about pushing them
into different things when you are in that early career, you
know, point of just trying everything and, you know, no
experience is kind of negative. It's just about getting a
different type of experience in architecture with that sort of
in mind, your your early career was at HMM and field and Fowles,
(08:32):
obviously quite two well known and and very well regarded
practices, but two very different sizes and work, you
know, types of work. I wonder if we could just spend
a little time just kind of looking at the differences
between those two and your differences in their sort of
experiences. Yeah, I had a HMM, something of
a baptism of fire. I went in Part 1.
I was definitely the youngest inthe whole practice.
(08:54):
I think there were two-part onesthat year.
I'm still friends with James, who was the other Part 1.
It was, HMM, very much that timewas like, I think, I think there
was 75 when I joined and then they took quickly.
It was 100 and it was a time of great growth.
It was incredibly exciting. And they put me on this amazing
(09:15):
office building. It's called the yellow building
in Latimer Rd. It was Monsoon Accessories
headquarters and it was, they were topping out when I joined
and I basically got to work on it on site.
I did, you know, doing various in the packages, package
information, subcontractor information, like all that real
nitty gritty. It just went straight in for it
(09:37):
and it was just brilliant. And then I think my take away
from working with AHMAM is that they are such a well run
machine. Like it's such an effective
practice, design practice. And they have at that time, like
my big impression it's still true.
Is, is there was a sort of system for everything like they
(09:57):
had, they'd worked it out. They were and they didn't, they
were obsessed with like taking something that they did well and
then making sure they didn't forget that in the future.
So everything you did was exemplary, as in you would try.
I mean, and I had a very good project architect called Sarah
and she was obsessive about detail and really used to pick
(10:18):
me up on stuff and red line things.
And, you know, immediately you'll sort of, you have to
really be on it. So I think the thing I got from
H Man was like really good background and sort of thinking
about the best way to do things in practice and systems and make
sure things get better over time.
And I still obsessed. I'm still obsessed with that
now. And that's sort of a lot of what
(10:39):
I do now is systems thinking with practices working out.
OK. Your practices is an engine.
Like how do you make sure it's running well?
How do you take, yeah, change something and make it run
better? That's very much what I'm doing
now. And then we filled in files.
I wanted to do something very different from H&M and they were
great. I I worked at H&M through my
(10:59):
part too. They were very supportive and
I'm still good friends with themand my wife works there still
now field and files. I knew Fergus personal friend
and I was very excited by what they did.
They did T Prem and it was yeah,very early days.
Like we were in this like very cold warehouse, freezing cold
warehouse in, in Hackney. And he, he let me come in and do
(11:22):
a couple of things and then theythey hired me as a Part 2.
And at that time we didn't have a Part 3 in the whole practice.
We were all part twos. Yeah, incredibly young practice.
Wasn't it so young? Yeah, they.
And because they just had the attitude of like, just just do
it, you know, we'll work it out.It'd be fine.
And. And it was, yeah.
It wasn't without risk. And so without stress, it was
(11:43):
very, very much, it was very stressful and hard work.
But because I went in to a smallpractice, I, you become, you
know, it's just everything, every, you have to do everything
and you become quite essential quite fast.
And then I, I was able to get projects which were way really
beyond my years and thankfully managed to deliver them well.
(12:07):
And so I did my part three thereand then it was, it was just
something about the kind of justgot to get on with it and do it
all and and we'll work out and we'll make it work.
And they and the work was very high quality and it's continued
to be high quality. And naturally, I always joke
that like the best thing that happened to field and files was
me leaving because they've just gone from strength to strength
(12:29):
ever since. Fabulous practice.
Yeah, I mean, it's even in, in just two experiences.
It's it's just such a variety oflike you say, experiences and,
and, you know, responsibilities and, and, and things like that.
When obviously your next move is, is Doug and Morris or, you
know, now Morris and company. I I mean, obviously the kind of
(12:49):
interesting thing is, is Joe is,is XAHMM.
And at the time they were probably, you know, quite a
young practice as well. So it, it feels like, you know,
the field and powers and the H&Mtogether made, made a Doug and
Morris sort of thing. I wonder what that sort of
experience then going into that practice.
And I know on your CV you talk about being employee number 20
(13:11):
and, you know, leaving that practice 10 years later as a, as
a director. Yeah, so.
It was a great experience. Yes, it was very much so.
I did big, then I did small and then and then I guess medium, I
don't know, it's 2020 is sort ofsmall to medium, isn't it?
And then we by the time I left, we were sort of more, we were
practicing up at like 50 or 60 ish kind of kind of range.
(13:35):
So it very and yeah, so I I do think Joe and I worked well
immediately together because probably because we had that
shared history of HMM, like we Ithink immediately he kind of
like I recognized the filing system when I went in, you know,
and and so and like we in those days, we were micro station as
(13:55):
well. It was, you know, it was dreamy.
And so you just sort of, I just kind of it felt like, and we had
a lot of like shared staff members.
I don't know, I just felt like very familiar.
And he at that point at 20, theywere starting to go from lots
of, they'd done lots of like private houses, lovely small
jobs and got a lot of recognition.
(14:16):
And then now they were starting to be courted by bigger clients,
PLCS and bigger organizations. And you needed that sort of
understanding of what it meant to work at more of a commercial
scale and with large clients andwith more risk.
And so I think I came in and I was given sort of attended to
work on the bigger jobs I would say in the earlier day it like
(14:37):
in my early days there. And the first project I did is
it was delivered last year. I mean, it took 10 years to
build practically. That's the architecture.
That's the way it goes. Yeah, Yeah.
So it was a good blend. And you know, if you know, in in
terms of advice to sort of younger architects, I do think
it's good to have a sense of different, there are different
types of practice and they have very different cultures and it's
(14:59):
good to just understand what they're all about.
And I wanted to be part of something that was changing and
fast and like at field and Fowles.
I mean, it was, it was I left like, you know, with, with very,
very fondly, but I, I basically,I'd seen Doug and Morris's Otis
building in Denmark Hill and I was like, oh man, I just have to
(15:23):
go work with that practice. I really wanted to go work with
them. And because I had the feeling
that they were just going to go on and like suddenly sort of
kick on to the next level. And I wanted to be part of
something that where that feeling of like rapid sort of
growth, I suppose and and experience and that was that was
a good decision. Yeah, I mean, it does feel like
(15:45):
they're, they're one of those practices of that generation
that have been able to sort of navigate the commercial aspect
whilst retaining that sort of design quality that really, you
know, stays with the projects ofspace.
Honestly, they they went throughthat sort of difficult period
and and, you know, the breaking up of the band sort of thing.
And yeah, obviously that as a, as an individual and as a, as a
(16:05):
person in that company, that must have been a really kind of
strange thing to experience and navigate, I suppose.
Yeah, it was really stressful. It was really, really difficult
period for everyone involved. John marry and like, you know, I
was in me, David Miranda and Charlotte at the time, Co
directors and we were we'd sort of we were tasked with, I
(16:31):
suppose, helping the practice through that period and sort of
navigating it and making sure that it was quite a disruptive
time for obvious reasons. And it was a lot of difficult
meetings, frankly, And that's but you know, we came through it
and it was everyone came throughit and it's just you just stuff
(16:52):
like that happens in practice. It's not it's not a it's not
always a walk in a park, you know, and you've good score is
be like, we had that sort of like, well, OK, what's next,
what's next? And you just keep working
through it. And then we came out on the
other side. Mary came out with a great
practice and she's done wonderful work.
And you know, we're still we're in touch.
(17:13):
And I think probably it was, youknow, with the benefit of
hindsight, the practices went indifferent directions.
Like you can see what Mary's done and Joe and the stuff that
we did with Joe afterwards. It's sort of, yeah, I think it
was the the founding of MorrisonCompany, which I was sort of
pivotal in along with the others.
(17:36):
It was a repositioning and sort of we were trying to be be clear
about the sort of what we wantedto do and the way we wanted to
do it basically. And it was, yeah, it was an
interesting, an interesting timefor sure.
And was was the point of I mean,obviously with the naming of the
company, Morrison company sort of thing and and opening that up
(17:56):
to a sort of, you know, was the intention was that it it became
more about the OpenTable with the rest of the directors, I
suppose. Yeah, so the and company thing
was a big deal. We actually talked.
It was the last idea we had. So there was AI won't name the
other names that we had. There was some funny ones,
someones I'm really glad we didn't do and we were called
(18:17):
something else for a few weeks. I definitely won't say that, but
we were trying to not do a name name.
And then I think Miranda said, hang on a SEC, like, what are we
doing? Like there's a reason Doug and
Morris is, has has a value. I think she, I remember she used
the word cache or and, and, and there's a reputation there.
And people do come to us becauseof Joe.
And so the thing was like, OK, Joe is the sort of spiritual
(18:42):
leader of it. And then we're the company.
And it was very much about shared voices, trying to very
much share the culture of leadership across the practice.
What were the sort of values that, you know, you as a
director really tried to sort ofhone in your role and, and
amongst your sort of staff and the sort of, you know, the
values you try to push amongst your people to get the best out
(19:04):
of that? I get I have a thing, which is
what you do is more important than what you say.
So I think the way turning up and like leading by example is
important. So if things go wrong, like how
do you take responsibility for them yourself rather than, you
know, blaming other people? I think another thing, like I
definitely made mistakes earlieron when I got to the point where
(19:26):
I was maybe overseeing other people or starting to run
projects where I definitely micromanaged people and made
them. You know, I could see that it
was not an effective approach. And it was coming.
It was coming out of probably because, yeah, when you're
stressed, that's what you end updoing.
And if you've not understood howto do it properly, like that's
(19:46):
the first major error that people make when they go, when
they reach more management or senior level or leadership level
is they think they have to control every detail and they
try and and they try and, and they end up micromanaging.
I don't, I don't, I mean, you'd have to ask my teams, but I
don't think I did that. I was very much always about
like I just trusting. We had amazing talent, talented
(20:08):
people there. So just trusting people and
being available if they needed support.
I think so, yeah. Having a culture of like doing,
not saying I think is important and I, you know, I genuinely
just really liked all the teams that I worked with.
We, we, we were lucky to have amazingly talented people.
So it does make it a lot easier when does.
(20:30):
Yes, you know, so eventually youleave Morris and company.
Yeah, onto. Different things.
Different, different things, let's say.
Was it, well, I suppose we'll, we'll come on to the sort of
incredibly exciting AI stuff andconsultancy and, and software.
(20:52):
But the thing I'm kind of interested in what was it that
allowed you to have your kind offinger on the pulse with that
and, and take that kind of opportunity because it's, it's
something that, you know, so many of us are trying to grapple
with. Yeah.
And I suppose what is it that you saw as probably an early
(21:14):
adopter to, you know, those sorts of things, those sorts of
conversations, those sorts of rumblings that made you kind of,
you know, stop and, you know, stop practicing as a architect
and jump into this kind of new field?
My, in my family, I've got people that are in tech and
people that are in AI. And so we've been talking about
(21:35):
it for about over a decade aboutwhat what it would all mean.
And I kept thinking, when's thisgoing to come to architecture?
When's it going to start impacting us?
And I would say 2022 was the year when we saw that really
happen. And I'd spent probably the five
years before the, if I had sort of, if you could describe them
(21:56):
as side hustles, I'm not sure. But like hobbies or whatever it
was that all the hobbies were around tech and software.
So I was doing sort of evening classes in Python And product
management and listening to and reading about tech entrepreneurs
and just getting kind of just following my nose down that
(22:17):
rabbit hole. Like I was, I, I'm quite an
obsessive character and I was kind of a bit obsessed in that
realm. And I did sort of several things
beforehand. Like I had this, I did a thing
with a family member on like algorithmic trading, like high
frequency trading with, with using Python to do that.
(22:41):
And then I did another, which was like a statistical method to
predict the next Michelin star restaurants in London.
Like so it was like constantly coming up with these things.
Then we did a passive house retrofit passive house tool.
I did that with two, two other guys, one of whom's gone on to
found he's got a start up that'sthat's doing that.
So there basically there were rumblings, people talk about
(23:03):
finding your niche. And my brother-in-law had said
to me around the time we were tricking around with things
together, well, your niches mustbe architecture plus like
software and tech or whatever. And I was like, oh grace, that
sounds awful. And I just couldn't really see
what that would look like. And because then I was like prop
tech, like not really interestedin prop tech per SE.
(23:23):
And then in, yeah, in 2022, we, me and Mateo, who's the manager
at Morris and Company, we were looking at thing called Delve,
which is a Google, originally Google owned product.
And it's like a master planning generative design tool.
And we'd, we'd gotten a trial for it.
And I remember two things. So the, the, the first thing was
(23:43):
it, it didn't really work. It wasn't robust.
It didn't at all do what I wanted.
And I was annoyed that I felt like the people who designed it
hadn't really understood the anatomy of buildings and of
places and everything was numerical.
They'd been turned into a statistical assessment.
And I felt that was very utilitarian and not really what
the world needed. And the other thing was they
(24:04):
were pricing it very highly. So it was, it was a yeah, yeah,
it was, it was this, it was a sort of full time salary type of
price. And I just thought I kind of
came away from that experience thinking, OK, that feels like
quite important as a moment. And then I went home and I, it
was like, say it didn't really work, but then it also did also.
I was impressed. I was very impressed.
(24:24):
I was like, oh, wow. And then I just thought, we'll
give this five years. This is going to be a huge
thing. And maybe it won't be that
product. I'm sure it'll be other things.
And I suddenly start to think, Ithink we should be, I, I don't
understand where the architects are in this like, and I went on
their website and I started looking at the job section.
And so I'm, and then I realised I'm, I'm thinking I'd quite like
(24:45):
to work on, on this product. And that's a strange thing to
think if you're a director in a company and I was doing
literally my dream role. I had a direct ship in a great
design practice as delivering projects.
We probably finished 5 or 6 projects that I'd been, you
know, pivotal in, in my last year.
And I'm going online like seeingif I can get a job in like New
York at this tech company, whichis ridiculous.
(25:07):
But it was a bit of a signal. And then very, very soon after
that, the first release of ChatGPT happened.
And I, because we'd had this narrative and discussions for so
long, you know, I was the first,I was like on the first day
logging and using it. And Dali was happening around
that time as well. And I guess I sort of looked at
(25:29):
it and I've had a few things in life where you know, when you
get these nudges and you see things and then you ignore it
and then two years later you seeit again and you ignore it.
And then a couple of years later, it's like a huge thing.
And you just think, why didn't Ijust act on the 1st impulse?
And, and, and then I started playing in my head, well, what
would that look like if I just acted on this impulse that I
have? Like I'm spending all my time on
this, why don't I do this? And so I think I was also at a
(25:54):
moment in Morrison Company wheremy major projects were coming to
a close and I was looking not atthe next year or two years, I
was looking at the next 10 years.
And we were also appointed as a sort of leadership team where,
you know, we were looking at very longer term plans in terms
of, you know, sharing company and all the other good stuff.
So it was all in. It was all in it was like
(26:15):
you're, you know, basically see you at retirement or do
something else. And I, I just decided to, I
think I, I, I said something like, I'm not sure if this is
the fish I'm going to catch. I think I need to go catch a
different one, which is mental actually, in hindsight.
And it, it really was does feel still a bit mental and it, and
(26:35):
it, but it's been an amazing pleasure the whole thing since
then. And thankfully it's worked out.
And I, you know, if it hadn't worked out, I probably would
have just, you know, gone and tried to get a job again and
carried on being just a traditional.
No, no bad thing. So I've yeah.
And then it's interesting once you spend for when you work on
something full time, you, you get, you know, quite good and
(26:56):
quite obsessed with it quite fast.
And then the other thing I thinkabout this whole space is that
really anyone that tells you that they're an expert in
generative AI before 2022, before they'd even seen ChatGPT,
they weren't because, yeah, there's phenomenal theory around
it. But in terms of applying it to
real world domain expertise and and use cases that's happening
(27:18):
right now, you know, and it's we're two or three years in
basically. So with that in mind, obviously
your company ARC works, your consultancy, you work with other
architects, other practices navigating this whole thing.
Yeah. What is the industry?
You know, how are the people that you're working with, what
(27:41):
use cases are they finding for this?
And I suppose how are they adapting their practices to
these things? And I suppose at a time when so
many of us feel like, you know, we want to get on this train and
and we don't want to miss it sort of thing, I guess it's kind
of like understanding what thoseopportunities are.
So I talk about the work I do intwo big buckets.
(28:03):
There's images and video, that'sone bucket.
And then there's text and data. And for the images and video,
we're talking about diffusion models mostly things like mid
journey, stable diffusion, flux.In the text world, we're talking
about large language models essentially.
And those two realms have different implications for
practice. So the the image realm is I
(28:26):
guess there's that, you know, that meme of like, I want AI to
do the dishes. I don't want him to write poetry
or whatever. I suppose I think of a lot of
the text work as the dishes. It's the more the administrative
supportive tasks, project support tasks and maybe the
image models are more like the not dishes, whatever that is,
the creative stuff that we'd rather do ourselves.
(28:47):
What I find is every practice falls in a different place.
So some people want to do everything and they are all in
on let's change the way we process all written information
in the practice. And also at the same time, let's
like completely revolutionize our design process.
And it does go to the core of your design culture if you want
(29:08):
to do it. And then other practices, you
know, I'm very lucky to work with a lot of amazing practices
with, you know, international reputations where they are more
circumspect, particularly about the design side and which I
completely understand. And that's kind of one of the
reasons I set up Archer works, which is I could see this thing
was going to go straight, straight to the core of our
(29:29):
practice culture. And they are not really, some of
those are just saying we're justnot going to touch image models
because we don't really want, wedon't really want to bring AI
into that part of our world, butwe'd love to have help with.
You know, the stuff we don't enjoy.
The stuff we don't enjoy and thestuff we don't enjoy would
include things like regulatory audits, briefing trackers,
(29:52):
meeting minutes, condition trackers, all kinds of
administrative project management tools which are
hugely aided. Now if you if you can learn how
to do it in the right way by theuse of these phenomenal
particularly, yeah the latest reasoning models that are coming
out. If you use them in the right way
(30:12):
with the right setup and you andyou know how to, how to avoid
certain risks, they are astoundingly effective.
And I'm, I continue everyday I'mshocked by what I see is
possible. And one of the reasons I love
doing workshops and, and meetinglots of different practices is
(30:33):
in every single workshop, someone comes up with a
different thought and they'll say, oh, could I try this?
I've just been trying to do thison a project and suddenly you're
taking something that exists in the world, technology and you're
actually applying it to a domainwith, and you're applying your
expertise and you know your field.
So I know what I understand the risks of being on an
(30:54):
architecture project. I understand the sort of
imperatives and the stresses of it and also the opportunities.
So your, your job then is how doI take this thing that's out
there and fit it into that worldin a way that's actually not a
pain in the arse and, and actually something that people
are so grateful for. So when it's working well,
people genuine, you know, you get texts like people just
(31:15):
saying, Oh my God, this is unbelievable.
I'm so delighted with this thing.
Or they're really happy to show you something that they made.
And so that's what I'm saying. I think I talk about heatmap and
AI heatmap for every organization.
And the idea is that your particular sensitivity to this
technology depends on the culture you're practicing, how
you operate already. And, and so you can look at a
(31:37):
practice and you can say, where are the opportunities?
Where do we, and then where, where do we have like where do
we want to be really aggressive and where do we want to be in
more circumspect? And you'll, what will emerge is
a picture of opportunities. And I call it a heat map.
And essentially you just work. There's a way of looking at it,
analysing what you'd like to do 1st and you just go straight
into those things. And you, you, you have a quite a
(31:59):
sort of objective way of saying,well, there's 100 things we
could do, but what are the five things we should do in the next
6 months or so? And you just work at it like
that basically. So with the, with the pressures
of the industry and we all know this, this idea of racing to the
bottom in terms of our, our feesand our value.
Do you see the AI opportunity asa, as an option for efficiency
(32:21):
and, you know, the idea of tackling these problems or I
suppose with the rest of the world, you know, speeding along
on that same thing, Are we just going to be in the same position
that we were all along? I wrote something last year
which was we need to talk about fees and we do need to talk
about fees in general, like outside of AI there is, I don't
(32:44):
like talking about, I find the word efficiency slightly crude
when we talk about AI because it's so 1 dimensional.
I think I want practices to become more effective and I also
want to make sure that any benefits that they get that they
actually feel themselves and that they don't just pass them
back to the market because I think we've been giving value
(33:05):
away to the market for too long.I'm worried if we just are
particularly worried about the idea of time saving, if we are
moving our fee structures to a time standard, time based
standard. I in the 15 years or so since I
sort of did my part to the culture of fees has shifted from
percentage based value based fees to lump sum fees derived
(33:30):
from resource plans or and, and unfortunately those are usually
underestimated. And I see that as being a major
problem for undercutting as well.
And the problem with that is if you make a time saving and you
use it as a competitive measure against another practice, then
what you will do is you will drive our fees into the ground.
And that's how we lose at this whole thing.
(33:53):
But we, we are uniquely positioned to discharge the
responsibilities and duties of architects, particularly with
the Building Safety Act. We're, we are essential.
I think I still believe that we need to understand our value.
And with the, with the rise of AI, we need to be the best at
using it and not, and we need tonot become at risk from some
(34:14):
product that pops up that promises to do what we do.
We should be better. We should be way better at that
at, at any product than that. And we should be paid well for
it. And so I, I sort of push back.
I try not to talk about time saving.
Of course there is time saving involved.
(34:34):
If you're doing it the right way, you should find the quality
goes up and you should also findthat your enjoyment of the tasks
way goes up. So you, you get great pleasure
and satisfaction when from say sitting in a meet.
You know, the big, the big one that I'm enjoying at the moment
is transcriptions are just so phenomenal.
So it and there's ways of takingtranscriptions in person that
(34:56):
that recognises different voices.
So you know who's speaking and when and taking those and
basically producing your, your written records, your reports,
updating trackers, doing all kinds of things with that raw
material. So the meeting becomes
incredibly important and what you say in the meeting is
consequential. So you start, you're more
declarative when you're communicating.
(35:16):
You start making sure, OK, so what's the action?
And what have we decided or whatdidn't we agree?
And you start saying these things out loud because you know
it's going to get picked up and it's going to end up in the
tracker. It's going to end up in the
minutes or it's going to end up in the planning statement or
whatever it is that this thing'sgetting put into.
And that means that you work better because you, you, you
work. You have better meeting hygiene.
(35:37):
You have better communication. What fascinating feedback loop
to say that we're changing our behaviour because of that.
Oh yeah, we definitely will do. Yeah.
I mean, I'm seeing it already and and I actually in that case
I think it's a positive. Change.
Yeah, completely. There is an interesting sort of
counterbalance to that, which isif if an AI is listening, then
we're we could be on the hook for all kinds of things in terms
of indemnity. So, and if you're a cynical
(36:03):
project manager or, or contractor, then you know, you,
you might want to use transcriptions.
And I know that they're already thinking about doing this and
essentially to pursue architectsand make claims more
effectively. So maybe what happens is the
culture of meetings, technical meetings become extremely,
they're just procedural and no one says anything of any
consequence. And then what happens is you'll,
(36:25):
you know, I'll take that away iswhat people say.
I'll just have to confirm that. And of course, the real
conversation happens somewhere else, so we'll see.
With with with the whole, you know, changing attitudes around
AI and how practice is using it.I wonder if we talk about your
sort of new start up Omni chat, if you can give me a kind of a
(36:45):
brief intro into what the origins of that are and and why
you decided to go down that route.
Yeah, I mean I'm incredibly excited about it.
So I met Stephen, who's my Co founder on a on an AI panel at
Arab an event and we, we, we sort of clicked.
We immediately went looking for innovate UK funding for for a
(37:08):
project. We didn't actually get the
funding, but it's interesting the idea we had, or that's
really Steven had, I shouldn't take credit for it, is what
we're going on. We're going to probably be
implementing in the next year orso.
And that's a more agentic approach to an API based agentic
approach to building different automated workflows for
practices. But he sort of said to me,
(37:30):
we're, we're working in a similar space.
So he's, he's software development background AI
specialist and he's working withsome, some practices.
I know. And he's, he said to me, well,
it's like your biggest challengeat the moment.
And I said, you know, I, I thinkthere's huge potential in large
language models, but I'm worriedabout privacy and I'm worried
(37:50):
about access because what I was,what I see is the private tools
you have to pay for. When I say privacy, I mean
protection from training. You still have to trust third
parties. These are not private like
airlocked LLMS. But you're at the moment more
than if you do a survey in a practice, you will find more
than 50% of people using free chat BT and they are just giving
(38:11):
all the data away for training. So they're making the models
better. All of the stuff you put into
ChatGPT, all the attachments, whether you own them or not.
And you probably don't own a lotof them.
Probably they belong to other people.
You're giving them away. They might be confidential, they
might be GDPR sensitive. And you know, there's, so
there's a, there's a data issue there.
There's also, there's a, there'sa quality issue.
(38:34):
I I did a test last week, just the way you prompt will
massively change accuracy. So if you drop an attachment
into freech HBT, if you drop theEU act into freech HBT and you
ask it to retrieve all of the articles it does about it gets
about 4% accuracy. Which is shocking.
If you do the same in GBT team using attachment methods, you
(38:57):
get about 16% it, you know, getsa few of the articles doesn't do
a good enough job. But if you do this method we
endorse could unabridge prompting you get it.
It retrieves every single article 100% accurate.
And that's using the same model.It's just interacting with it in
a different way. So I was talking to Stephen
about this and he said, well, I've got this prototype and
basically as soon as I saw it, Isaid like, this is what everyone
needs. So we implemented it first for a
(39:19):
big practice in London and now it's the main tool I use for all
use for all of my training. And the main idea is it's we've
got all of the language models that we think are any good
connected up inside it. OK, So it's one interface that
you learn and you can flip the models at any point.
So you can do work in Claude andthen you can have GBT come in
(39:41):
and check the check the homeworkof Claude and see how it did, is
it accurate? Did it make any mistakes?
And then you can maybe use a different model for data
analytics or whatever. So you've, you sort of have
configuration set up for different tasks and you, you get
to know when you get very into it, you get to know which ones
are good at what and they have different personalities.
They're sort of like your team and you have to decide who in
(40:01):
your team you'd like to pick forcertain tasks.
So I think of GBT 4 O as like a very, almost like an engineer.
It's like very sort of analytical, very quite
utilitarian. Whereas clawed, the clawed
personality is much more able tohandle nuance.
And to, to, to deal with uncertainty and to use and to
(40:24):
adopt style of language. So you would, you would use
those two types of characters for different things.
So that's why it's called Omni chat because we have these,
these Omni models basically. So it's multi model.
And then the other thing that it's allowed us to do is to
standardize workflow. So when we're working with
practices, I'll build them a solution for, I don't know,
getting basically taking something from one position into
(40:45):
another position and and outputting material that's
essentially usable. And we we set those up as pre
made templates. So you just click a button and
they run. And yeah, we're starting now to
implement agents. I don't know if you want to talk
about agents, but agents can do sort of end to end tasks now and
link different tools up with AP is so that's also going to sit
(41:08):
in Omni. So that's becoming extremely
busy. We've got clients in Denmark,
Ireland, Germany and the UK and.So that's almost like having a
staff member that will go and doa task for you and put the
pieces together. It's exactly, it's well, you
still have to say. What I like to encourage is that
(41:28):
people learn to prompt really well.
And then they essentially have asuite of workflows that they
have, they've built themselves using this method.
And it'll be like, it might be quite a big job that the thing
does, or it might just be a small section of a job that you
do regularly, but it just reallyhelps you with that part of it.
And rather than doing it in thisad hoc way, I describe it like
(41:49):
what most people do in GPT is they kind of open the fridge and
decide what to cook for dinner. And it's like, oh, I've got this
and this and I'll just throw them together.
With this method, it's much morelike you've got a recipe and it
executes the same and you're just changing the ingredients
each time a little bit. So there's a couple of things
that change, but most of it stays the same.
And so you're looking around thepractice constantly for like,
what are these things we do all the time that actually, yeah,
(42:12):
they're different. But there, there is a structure,
there is a template, there is a method that we follow.
Those things are perfect for this.
So there's, I think we worked out there's something like 80
typical, there's, there's about 80 things you can do across a
whole practice. And that's just as a starting
point that we, that we've started to do and that.
(42:33):
So a lot of what we do is that, and that's becoming more and
more of my efforts and, and focus.
And it's also made me become waymore specialised on large
language models because I'm working with Stephen and Stephen
reads every single research paper.
He's every new model release, he's trying it on the day and
we're able to connect new modelsvery quickly.
And that whole, the process of working with Stephen has been
(42:57):
just the most amazing education for me as well.
And I'm, you know, a lot of my clients benefit from that
directly. So yeah, I, I probably, in terms
of my, where I see opportunity, I love the image realm.
I'd, I find it, it this sort of,it kind of interests a different
part of my brain. It activates the sort of the
part of the brain that you enjoyusing for sketching and that
(43:18):
sort of part. But the thing I think is really
transformative is probably the language models work, yeah.
Yeah. I think just sitting here
listening to you talking about the processes and understanding
the different processes, you know, an architecture practice
goes into, you know, I am thinking about those early days
at HMM that you're talking aboutwith the systems and stuff.
And that's right. I think it's a really
(43:39):
fascinating thing to you. Know Go go.
Full circle with that and and you to be looking at new
technologies but solving the same problems that you were
thinking about. I like that you closed that loop
because I, I, I think that is about right.
Actually, I think there's a, there's a method of thinking
about businesses as systems and,and the within a system, if you
(44:00):
can observe, can you observe it objectively and say, can you
describe it to someone else? Like how do we do bids?
Can you describe it or is it totally made-up every time?
If you can't describe it, it's not a system and you need to be
able to draw it. And then you need to say, OK,
well, there's 10 parts to that process or 20 parts to it,
whatever it is that you're doing.
Like maybe you're bidding for awards or whatever it is, though
(44:21):
you could probably describe it and draw that process.
And then you should, then you can become obsessed about
improving the system. And then it's almost like
you're, you're constantly refined, refining this overall
vehicle or engine or whatever, whatever the right metaphor is.
And that's how I think about practice.
And I think maybe it did come from HMM.
Or maybe it's just the way I tend to think about problems.
(44:43):
I don't know. With with that in mind, just to
kind of close up the conversation, what do you think
people can kind of take away from your career journey?
Considering, considering that like we've just kind of alluded
to you, you've kind of come, youknow, you've gone full circle
and then, but now are, you know,incredibly sort of future
(45:03):
looking and, and trying to solveage-old problems.
But in a, in a, you know, clearly sort of contemporary
manner. What is, you know, if, if you're
a young architect starting in practice or, you know, even, you
know, you're a young person starting into architecture, you
know, what is that? What is that journey going to be
looking like in the future? And you know, what are those
(45:26):
sort of takeaways that we can get from your journey, I
suppose? I find this a bit hard actually.
I run it really hard to think about the future right now in
terms of career, career advice for other people.
But I think maybe what we can besure of is there's going to be
lots more change. And I think there's maybe a way
of thinking about your career ina more flexible way.
(45:50):
I've probably had a problem withthe title architect for some
time, as in I'm very proud of the fact I'm an architect.
But it's also, it's something ofa straight jacket because there
is a certain way that we think about what we must do each day
and how we must do it. And that is, it's not, I don't
(46:14):
know that that's, I think that that's sort of, I'm an architect
and that's my ego and it's, and I define myself in that way.
I find that to be limiting and not ready for the what needs to
be much more flexible approach. So I, I think the skill set for
architecture is phenomenal. What we're really good is like
(46:35):
quite abstract problem solving, like we are taking very complex
things and weaving an answer andalso communicating that story
from beginning to end to clients.
And those sorts of skills are universal and those are the
things they're not going to be alienated from us.
Like if we can communicate with clients, connect, understand
their problems, find solutions to them.
(46:58):
Building's incredibly stressful.Also, it happens in the real
world, like with like in with like muddy holes and people
digging stuff on site. Like it's a, it's a messy
business and we need to lean into that sort of those parts if
we want to make in terms of AI and disruption.
We have that's where we bring value.
I think is, is our, the, the ability to come in and, and see
(47:21):
the problem, identify and, and be able to weave a solution and
then communicate the solution. And then you're, we're almost
like producers rather than like artists.
Like we, we sort of we part produce a part director or
something and we're building a vision and we need to bring
people in at the right time. But I also think like, I don't
that we've got this thing where maybe it's our training, but we,
(47:42):
we sort of feel like we need to be like the best.
We have to be totally like all rounders at everything.
And I also think that's a bit ofa problem.
So it just in terms of if you, if I were going into practice
right now, I would be trying to find ways to become essential.
And I think that's probably, that's probably true of any time
in, in, in human history. If you want to do well, you
(48:05):
basically have to work out in any situation, how do I become
essential and indispensable? And that's that's true now.
But then how do you do that now?Well, it's probably different to
what it was five years ago. And I think so that combination
of working out what's essential and also being flexible to the
idea that probably the role of architect is going to change a
(48:26):
lot. And that we need to be ready for
that in order to have a flourishing and happy career.
And not be too rigid about like,I don't know if you have the
same thing. I have this sort of monograph.
It's my joke is like in my head,I have a monograph and it's like
the, it's the like career architect name on a book.
(48:48):
Here's My Portfolio. I've delivered these projects.
Aren't they amazing plan sections, elevations and who am
I trying to impress? Like mostly other architects.
Like that's all of the things I just said.
Most architects that I know, I think are vulnerable to that
fantasy and I definitely am, butit doesn't serve us well and
it's very limiting because actually there's all kinds of
(49:10):
other opportunities. If you just say, well, how can I
be essential and where can I provide value?
So like my, it may be my thing is, you know, when you're
looking at the next, your next business opportunity or your
next move, it's like you could think, well, what do I want for
myself? Or you can think?
And in my situation, I probably would have come away from Morris
(49:30):
company thinking I should set upa practice and, you know, set up
my own practice, architecture practice.
Maybe it's AI driven. And then I thought to myself,
I'm not sure we need another practice right now.
It feels like there's a lot of really good practices.
So what maybe the world needs isfor me, or what I could do that
would actually be more valuable is go help focus my time on
(49:50):
becoming really good at something and then go help lots
of other practices with that. And that's a different way of
looking at a problem is like, rather than what am I trying to
impose on it? It's like, well, what does, what
do they, where could I be valuable?
So that's I would say that to young, young up and coming
architects. But I do think it's a funny time
and I I'm, I'm, I'm slightly, I don't.
(50:13):
What do you think about the future of the what would you be
saying to architects coming up now?
I mean, you're seeing all this stuff too.
Yeah, I mean, like you say, it'sit's something that we're all
trying to navigate and I think it's, it's an incredibly
difficult time and, and particularly, you know, even
just choosing to study architecture at the moment is a,
you know, such a, a big, you know, decision to make and, and
(50:33):
things like that. So, yeah, it's, it's just
complicated, isn't it? It's.
A huge undertaking financially. Crazy undertaking.
It doesn't make any financial sense anymore.
And that's, that's something that's changed in my career.
You know, I, I, I was, I met, I went, I went to the Arsenal last
night and I met with a friend from Part 1 and he went and did
(50:54):
contracting. And lots of my friends actually
have ended up going and doing different things later in their
careers. But we were reflecting on our
starting salaries at Part 1 and how and how proud we were.
And you know, we look at the numbers now and they're not that
different. And yet we know everything costs
about 10 times more. So I, I find that to be
incredibly unjust and upsetting for, for up and coming
(51:15):
architects. They need to be angry about it
as well. Like it's not acceptable.
And I think the, the slight problem, I mean, it's a, it's
fee driven. I mean, practices are not making
our fees, they're not making ourprofit.
I think there's maybe a sense of, I don't know that those
things are felt by incumbent owners and leaders and practices
(51:36):
when they've been running them for many years, maybe maybe many
decades and they are fairly welloff and they're probably doing
it for the pleasure of doing it and it's less essential for
them. And I think that it got worse
during COVID in just in terms ofthe disparity of perspective.
And it's a difficult time. We need to sort it out.
I don't know how we sort it out.It would, it would really help
(51:59):
if, you know, our professional institutions were more direct on
this and took a lead. I think they've been a bit
agnostic about it for too long. And I think I don't know what
happens. I mean, basically the only logic
at the moment, you need to be pretty privileged, I think to,
to look at the economic equationfor architecture and think that
(52:20):
it makes any sense to and I probably am, I'm probably lucky
to have come. I had like my parents were
always like, you know, do what you love and, and which was
lovely. But that's the sort of that's
like generational advice. I'm not sure that I'm not sure
that's so appropriate. Not anymore.
(52:40):
In a way, it's like it's, it's, that might be quite an
irrational thing to do now, you know?
Completely and and you know, particularly given now the
government is talking about moving, you know, the
apprenticeship schemes away, youknow, the funding away from
apprenticeship schemes, which again, it was quite a hopeful
thing for. Obviously it was never going to
solve the problem, but it was a hopeful kind of thing.
So yeah, it's a difficult time and I don't think anyone's quite
(53:02):
got the answers. But like you say, at least our
professional bodies pushing and and having those conversations
might help. They really need to, yeah.
Brilliant, absolutely fascinating conversation.
Thank you so much for joining me.
Thanks for having me, it was very fun.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
The Clay Travis and Buck Sexton Show
The Clay Travis and Buck Sexton Show. Clay Travis and Buck Sexton tackle the biggest stories in news, politics and current events with intelligence and humor. From the border crisis, to the madness of cancel culture and far-left missteps, Clay and Buck guide listeners through the latest headlines and hot topics with fun and entertaining conversations and opinions.