AI in Architectural Practice

Randall Stevens: 00:00:00

Welcome to another Confluence podcast.

2

: 00:00:01

I'm Randall Stephens joined

as usual, uh, by Evan Troxel.

3

: 00:00:05

And our guest today is Charlie Portelli.

4

: 00:00:08

welcome Charlie.

5

: 00:00:09

Uh, I'll give a little brief intro

and then let you, know, tell more

6

: 00:00:12

about the kind of work that you're

doing, but, uh, met Charlie, uh, he

7

: 00:00:16

participated in our Confluence event

that we held in Brooklyn, uh, New York.

8

: 00:00:21

Uh, back in April and

gave a great presentation.

9

: 00:00:24

Charlie works with Perkins&Will, his

title is a digital innovation strategist.

10

: 00:00:29

Uh, our theme this year was all

around machine learning and AI.

11

: 00:00:32

So Charlie's working on a team of

people at Perkins&Will, uh, around, you

12

: 00:00:38

know, strategies, uh, and, uh, that's

what we're going to talk about today.

13

: 00:00:41

So welcome Charlie.

14

: 00:00:43

Charles Portelli: Thank

you folks for having

15

: 00:00:44

Randall Stevens: you're, uh, the title

of your talk, at the Confluence event

16

: 00:00:48

was, uh, AI in practice, you ended

up, uh, kind of walking us through.

17

: 00:00:54

know, some examples of how you

are thinking about, um, about

18

: 00:00:58

using machine learning and AI

techniques at Perkins&Will.

19

: 00:01:03

Um, I know, you know, one of the

things that you kind of kicked that

20

: 00:01:06

off was, was just the challenges

with, uh, so much unstructured data.

21

: 00:01:11

So maybe we can just kind

of kick it off there.

22

: 00:01:13

You can tell us a little bit

more about what your all's

23

: 00:01:15

team is thinking about and, uh,

24

: 00:01:17

let's

25

: 00:01:18

Charles Portelli: Yeah, no, totally.

26

: 00:01:19

I mean, when this whole sort

of, um, Data topic came.

27

: 00:01:25

Um, you know, sort of relevant to our

industry, you know, you give or take like

28

: 00:01:30

10 or so years ago, you know, buildings

equals data and so on, like the folks

29

: 00:01:33

at Case kind of coined that phrase.

30

: 00:01:36

Everyone's like, yeah, we have all

these models and we have, you know,

31

: 00:01:39

Revit files and spreadsheets and

yada, yada, yada, whatever, right?

32

: 00:01:43

We can do something with it, right?

33

: 00:01:45

Um, no, it's, it's highly

unstructured, uh, project teams change.

34

: 00:01:53

Um, There's not, sort of, like a really

rigorous, as rigorous as you'd like it to

35

: 00:01:59

be, there's not a really rigorous, sort

of, kind of, follow through on projects.

36

: 00:02:03

People on projects turn over and so on.

37

: 00:02:06

So, we have, in our industry,

a lot of unstructured data.

38

: 00:02:11

And that's that, right?

39

: 00:02:12

So, obviously, the more structure, in

my, this is now my personal opinion,

40

: 00:02:17

obviously, the more structured it is,

I feel like, um, it's easier to, to,

41

: 00:02:22

kind of, implement and use right away.

42

: 00:02:25

The more unstructured it is, the

harder it becomes to use, right?

43

: 00:02:29

I just kind of think of my own personal

life, like the more structured things are,

44

: 00:02:32

it's easier to kind of understand where

things are and how to find things, right?

45

: 00:02:35

Um, and with data, my mind

kind of thinks very similarly.

46

: 00:02:40

Um, with that being said, you know, so

how do we clean all this stuff up, right?

47

: 00:02:45

How do we actually make it usable?

48

: 00:02:48

So we've been looking at, um,

for example, naming conventions.

49

: 00:02:52

Of spaces and rooms, right?

50

: 00:02:55

The one example I brought up,

um, during the presentation

51

: 00:02:58

was like a restroom, right?

52

: 00:03:00

So, I mean, I say restroom.

53

: 00:03:02

We already all know what we,

what we're thinking about, right?

54

: 00:03:05

There's, there, there's a,

there's a toilet, there's a sink.

55

: 00:03:07

Um, you know, if it's a, uh,

person, if it's a residential, there

56

: 00:03:12

might be a shower, bath or bathtub

or something like that, right?

57

: 00:03:15

And then you can start to think about,

you know, other names like toilet.

58

: 00:03:20

and Water Closet and so

on and so forth, right?

59

: 00:03:22

And if you have kids, potty, right?

60

: 00:03:24

Like, you know, that's sort of the

thing that like peaks my ears when

61

: 00:03:28

my kid says, I gotta go potty and

you're like in a public space, right?

62

: 00:03:31

Um, so, you know, how do we

kind of clean all this stuff up?

63

: 00:03:36

So what we've done is that we can query

all this stuff via the API easily, right?

64

: 00:03:42

And what we can do is now take

this, run it through OpenAI's API.

65

: 00:03:48

and essentially start to

kind of unify it, right?

66

: 00:03:50

So from like a space naming convention, we

don't have to, with kind of force rigorous

67

: 00:03:58

standards, um, on design teams, right?

68

: 00:04:02

Um, and a reason not to do that also

is kind of clients, uh, have their own

69

: 00:04:06

sort of naming conventions for, you

know, conference rooms and workspaces,

70

: 00:04:10

breakout spaces, you know, communal

spaces, that sort of thing, right?

71

: 00:04:14

So we can still work

within the client's domain.

72

: 00:04:18

uh, and naming conventions and

so on, sort of their brand and

73

: 00:04:21

their strategy and all that.

74

: 00:04:22

Um, and then we can scrub all

that data, kind of unify it, clean

75

: 00:04:26

it up, and then we can run some

analytics on top of it, right?

76

: 00:04:29

So that's sort of one way we're looking

at sort of cleaning, cleaning data.

77

: 00:04:36

Randall Stevens: Kind of, know, it

brings up the point that, you know,

78

: 00:04:38

we talk about when you're working

on these things as far as data and

79

: 00:04:42

structure versus unstructured, but

really behind all of that is, language.

80

: 00:04:48

How do, how do, how are people thinking

81

: 00:04:51

about these things?

82

: 00:04:52

Uh, you know, the, the world of

design, architecture and design and

83

: 00:04:56

engineering, there aren't strict rules.

84

: 00:04:59

There aren't strict standards

on the way that we do that.

85

: 00:05:02

So I think that's, you know, it, it makes

sense that it's really, uh, things are

86

: 00:05:07

named the way that we talk about them

and that the way you talk about them

87

: 00:05:11

may be different with whoever the client

is, or even the different, you know,

88

: 00:05:15

especially a firm, the size of Perkins

will, you've got a lot of different people

89

: 00:05:18

from a lot of different backgrounds.

90

: 00:05:19

So even their language, you

know, so think it's, uh, it's

91

: 00:05:24

interesting to think about that.

92

: 00:05:27

Um, you know, it's interesting

to think about that.

93

: 00:05:28

I think maybe 10 years

ago is your example.

94

: 00:05:32

When everybody was talking

about data, everybody would

95

: 00:05:34

start talking about standards

96

: 00:05:36

and the reality is you're

97

: 00:05:38

to get everybody on that.

98

: 00:05:40

So the idea that we can now leverage

this technology to say, no, I can, I can

99

: 00:05:44

let you talk in whatever language you

100

: 00:05:46

to talk in, and I'm going to

translate that, which I think

101

: 00:05:49

is a very powerful way to

102

: 00:05:51

Charles Portelli: even comes

down to building codes.

103

: 00:05:53

which is something that came to my

attention, uh, on a personal project.

104

: 00:05:58

So, in New York City, if you have a

kitchen that's less than 80 square feet,

105

: 00:06:03

it's not a kitchen, it's a kitchenette.

106

: 00:06:06

And that's

107

: 00:06:06

like a building code thing.

108

: 00:06:07

And if you label your plan as kitchen,

and it's less than 80 square feet,

109

: 00:06:11

it will actually get rejected.

110

: 00:06:13

Evan Troxel: red flag.

111

: 00:06:14

Charles Portelli: Right?

112

: 00:06:14

It'll get flagged, right?

113

: 00:06:15

So, it's like, some, some of these things

are outside of our, our control, right?

114

: 00:06:19

So we just sort of, kind

115

: 00:06:20

of, work with it.

116

: 00:06:21

Evan Troxel: There's that side of it.

117

: 00:06:22

And then there's the architect side of

it, which is like, I will name a room

118

: 00:06:25

or a space, a certain label, just so it.

119

: 00:06:28

doesn't bring up a red flag.

120

: 00:06:29

Right.

121

: 00:06:29

So to, to your it's or, or because it's

political in nature in that organization.

122

: 00:06:36

Right.

123

: 00:06:36

So we, we won't call that space a, a.

124

: 00:06:38

Storage space, we'll call it a flex space

125

: 00:06:41

or something like that, where it's like,

that might be a terrible example, but,

126

: 00:06:44

but the idea is like, architecture is

all about the exceptions when it comes

127

: 00:06:48

to these kinds of things, because, I

mean, that's, that's basically what,

128

: 00:06:52

what we're doing with the building

code is finding the exceptions so

129

: 00:06:55

that we can achieve some certain goal.

130

: 00:06:58

And that goes against this whole

idea of quote unquote standards.

131

: 00:07:02

And, and every time you bring up

standards or naming conventions, the

132

: 00:07:06

hairs on the back of the architect's

neck raise up because it's like,

133

: 00:07:10

well, yeah, they're all great.

134

: 00:07:11

Mine is the best.

135

: 00:07:12

Right?

136

: 00:07:13

And, and, and so you.

137

: 00:07:15

You can't have a discussion even

in an office about standards

138

: 00:07:19

because every team has their own.

139

: 00:07:21

And I think it is particularly

interesting to think of this set of

140

: 00:07:26

tools as a way around that, right?

141

: 00:07:29

Number one, architects are

all about workarounds, right?

142

: 00:07:31

So this is a fantastic workaround to

the whole standardization of things.

143

: 00:07:35

You're, you're using a tool

in a large language model.

144

: 00:07:39

To take all of the possible things you

can call something and funnel it into

145

: 00:07:44

a single thing, which then allows you

to do that proactive analysis to it.

146

: 00:07:48

I think, like, I always bring up

like drywall as an example, right?

147

: 00:07:52

There's one name for it, then there's

gypsum board and then there's plaster

148

: 00:07:56

board and there's gyp board and

everybody in, in several details.

149

: 00:08:01

I mean, Randall, you probably see this

all the time at avail details on a sheet.

150

: 00:08:06

We'll call the same thing,

different things in every detail

151

: 00:08:08

because they're pulling from.

152

: 00:08:10

Some standards library where there's

no standards on the terminology for

153

: 00:08:14

those things because it's such a manual

labor, kind of a thing to go through

154

: 00:08:18

all those details and standardize that.

155

: 00:08:20

And it doesn't always

even fit on the line.

156

: 00:08:22

Right.

157

: 00:08:22

So we, we start to abbreviate things

and I mean, it is kind of a mess.

158

: 00:08:26

So this is a really interesting way

around this problem in our industry.

159

: 00:08:31

I'm pretty fascinated by this as

like who saw that coming, right?

160

: 00:08:35

Because everybody was attacking

it from the other way around,

161

: 00:08:37

which was the ground up.

162

: 00:08:38

It's like, now we have to manually go

through everything and change everything.

163

: 00:08:42

And basically what you're saying is, Nope.

164

: 00:08:44

We don't have to do that

165

: 00:08:45

anymore.

166

: 00:08:46

Charles Portelli: Yeah,

no, don't, don't bother.

167

: 00:08:48

And it's funny you bring up the

topic of standardization and so on.

168

: 00:08:51

I remember, um, in a different lifetime,

I was working at a different firm and

169

: 00:08:55

I was talking to one of the partners.

170

: 00:08:58

And he made a comment, he's like,

the idea of standardization and

171

: 00:09:00

innovation are sort of on the two

opposite ends of the spectrum, right?

172

: 00:09:03

Like, if you're constantly working

within standards, how can you, you

173

: 00:09:08

know, put yourself in a place where

you're constantly innovating, right?

174

: 00:09:11

Um, so they're, they're sort

of opposed to, to one another.

175

: 00:09:15

So we've kind of found a way to keep

working the way you work, and then

176

: 00:09:20

we will just kind of clean this up.

177

: 00:09:22

And we've done it in a way where, uh,

it's actually not that much effort.

178

: 00:09:25

Which is nice, right?

179

: 00:09:26

We're not going in renaming things.

180

: 00:09:28

We're not going in breaking projects.

181

: 00:09:29

We're not going in copying projects

Into you know, a template file or some

182

: 00:09:35

nonsense or any of that that stuff, right?

183

: 00:09:38

Randall Stevens: So maybe you can talk,

Charlie, how, you know, as you're, as

184

: 00:09:41

you all started out just by looking at

the data and doing the analysis of the

185

: 00:09:44

data, the, how are, how are y'all now?

186

: 00:09:47

How is that manifesting itself back into

either productivity enhancements or, you

187

: 00:09:52

know, what are the manifestations of that

now back to the designer, the design teams

188

: 00:09:58

Charles Portelli: Yeah, so it's still

sort of a work in progress, right?

189

: 00:10:02

I think When I when I did the talk back

here in New York, I called it building

190

: 00:10:06

intelligence um, or Intelligence

in Practice or something like that,

191

: 00:10:11

and I think it's a play on the word

practice, right, because we're, we

192

: 00:10:14

have a firm architecture practice, but

we're also practicing how we can kind

193

: 00:10:18

of put this to use, right, so it's

all sort of, um, work in progress.

194

: 00:10:23

So, we have spaces, right, we now

have a way of kind of grouping them

195

: 00:10:27

by naming convention, right, so cafes

in various projects, now we can start

196

: 00:10:33

to run queries across all of these.

197

: 00:10:36

Um, what we're looking at is, uh, by sort

of standardizing the naming convention

198

: 00:10:43

and running these queries, we can

start to understand, okay, so you have

199

: 00:10:46

a hospital of X thousand square feet,

how many cafes do you have, what's the

200

: 00:10:52

typical, you know, cafe size, what's the

typical equipment in a cafe size of that

201

: 00:11:00

sort, um, and we can start to flush out.

202

: 00:11:04

Very quickly, sort of, um, a

jumpstart on a project, in a way.

203

: 00:11:12

So, I need to start a project, I need

to do a science and technology lab.

204

: 00:11:17

The performer says it needs, you

know, this amount of wet labs,

205

: 00:11:21

this amount of dry labs, this

amount of clean spaces, and so on.

206

: 00:11:24

Help me flush this out very

quickly and help me, you know,

207

: 00:11:28

mass it out quickly, right?

208

: 00:11:30

Um, just to kind of put something on the

paper and then we can start to massage

209

: 00:11:36

it, you know, more and more, you know,

as the, as the, the, the project evolves.

210

: 00:11:40

That's sort of the direction

that we're going, right?

211

: 00:11:42

So a little

212

: 00:11:43

Randall Stevens: you have a.

213

: 00:11:44

I know you have an architecture

background yourself and a lot of what

214

: 00:11:47

we try to, Evan and I try to, you know,

as we get people like yourself onto

215

: 00:11:51

this podcast, we're, we always try to

dig into like, well, what was the, what

216

: 00:11:55

is the process of going through this?

217

: 00:11:57

So, can you maybe talk to us a

little bit about, you know, if

218

: 00:12:01

you're like us, you kind of start

with putting a stake in the ground.

219

: 00:12:04

You'll kind of dream up what to use

the tech for, but talk to me about

220

: 00:12:08

how do you all actually engage design

teams in this process and then say,

221

: 00:12:13

here's what we're able to do and then

get their feedback in that, like what

222

: 00:12:17

would help you, can you kind of walk us

through what that process looks like at

223

: 00:12:21

Charles Portelli: Yeah, it's very dynamic,

um, and that's sort of kind of my way

224

: 00:12:25

of saying it's not standard, right?

225

: 00:12:27

So there's some folks within

226

: 00:12:30

any format.

227

: 00:12:30

Evan Troxel: when you said that, I was

like, Oh, that's a nice way to say that.

228

: 00:12:35

Randall Stevens: Dynamic.

229

: 00:12:36

Charles Portelli: It's a,

it's a dynamic thing, right?

230

: 00:12:38

So, at any firm that I've ever

worked at, there's sort of, you

231

: 00:12:41

can kind of group folks into almost

like three categories, right?

232

: 00:12:45

There's the people that know

what you do, and know how to

233

: 00:12:47

leverage what you do, right?

234

: 00:12:49

And you work really well together, and

sometimes you work so well together that

235

: 00:12:52

you actually spend little time working

together, because you just churn out

236

: 00:12:55

results, and everyone sort of moves on.

237

: 00:12:58

There's the opposite end of the

spectrum, the people that have no clue

238

: 00:13:02

what you do, and have no clue how.

239

: 00:13:04

They can engage you, um, and then the

in-between group that know what you do,

240

: 00:13:08

but don't understand how to engage you.

241

: 00:13:10

Right?

242

: 00:13:11

Um, so depending on those

three, that's why it's a little

243

: 00:13:14

bit of a dynamic situation.

244

: 00:13:16

Um, so I go to the office frequently,

um, and even though my team is remote,

245

: 00:13:22

when I say remote, you know, my team

is in Boston, Poland, Copenhagen,

246

: 00:13:28

Chicago, you know, so on and so forth.

247

: 00:13:30

They're all in different offices, right?

248

: 00:13:32

Um, even though I have no.

249

: 00:13:33

physical reason to go to the office.

250

: 00:13:35

I go to the office.

251

: 00:13:36

I interact with designers.

252

: 00:13:37

I talk with them.

253

: 00:13:38

I give presentations to

the office and so on.

254

: 00:13:40

So that's one method of engaging folks.

255

: 00:13:43

Randall Stevens: I guess there, I guess

there, Charlie, just a question about is

256

: 00:13:48

your team, is it viewed as a, a services

group within Perkins Will that can be

257

: 00:13:54

engaged by any of these teams and you

have to go out and kind of sell your

258

: 00:13:58

services of what you all can do how

much are you all proactively kind of

259

: 00:14:04

things on your own and then trying to

260

: 00:14:06

into the

261

: 00:14:06

Charles Portelli: Yeah, it's both.

262

: 00:14:08

It's both.

263

: 00:14:08

We, we tackle it from, from

a multi perspective view.

264

: 00:14:12

Um, so there is my, my co lead.

265

: 00:14:14

So I co lead the IO group with,

uh, my, my colleague Thomas Kearns.

266

: 00:14:18

He's based out of Chicago.

267

: 00:14:20

Um, Yeah.

268

: 00:14:21

Uh, he focuses more on the, on the

services side of things, uh, so

269

: 00:14:25

much so that some of those services

actually become externally facing,

270

: 00:14:28

client facing services and so on.

271

: 00:14:30

And he's kind of leading that, that

effort, um, sort of more leading

272

: 00:14:33

the internal side of things.

273

: 00:14:35

And now whether you want to call it still

services or not, you know, I'll leave

274

: 00:14:39

that open for everyone to interpret.

275

: 00:14:41

Um, but we do, we do

talk to project teams.

276

: 00:14:44

Uh, we do talk to folks within the

office, you know, leadership folks,

277

: 00:14:48

but also people on projects kind of.

278

: 00:14:51

in the trenches and we kind of express to

them what can be done, what are you doing,

279

: 00:14:56

how can we help out and sort of thing.

280

: 00:14:58

We keep it very open.

281

: 00:14:59

Um, it's also a way of kind of

understanding more and more how

282

: 00:15:04

projects are evolving over time.

283

: 00:15:07

Um, I've slowly, I've gotten pulled

out of projects over the years.

284

: 00:15:12

Um, and that's necessarily, it's

not, I think being on projects

285

: 00:15:16

once in a while helps out, but

it's hard to kind of do everything.

286

: 00:15:19

Um, so we do engage in a lot

of different, different ways.

287

: 00:15:24

Um, I try to do it at a grassroots

level, because like, those are the

288

: 00:15:27

folks executing the work, those are

the, uh, the folks that are going to

289

: 00:15:30

see immediate benefit, and if they see

immediate benefit, then the project

290

: 00:15:33

sees immediate benefit as well.

291

: 00:15:36

But then there's also sort of larger

initiatives, um, that we're working on

292

: 00:15:40

that kind of the whole firm will see.

293

: 00:15:43

Randall Stevens: Themes.

294

: 00:15:44

Yeah.

295

: 00:15:44

Yeah.

296

: 00:15:45

you talk maybe about, like, even around,

as you all began to look at how you could

297

: 00:15:50

structure this data, was that presented

back to, to some of the practitioners

298

: 00:15:55

in the group and what kind of response

did you get from them as you started to

299

: 00:15:59

show them the kinds of things that you

300

: 00:16:01

Charles Portelli: Oh yeah,

no, a hundred percent.

301

: 00:16:02

When, when we gave that presentation,

we actually, I forgot if it was the day

302

: 00:16:05

before or after, uh, my colleague was

from Boston, came to give the presentation

303

: 00:16:09

with me, the avail presentation with me.

304

: 00:16:11

So while he was here in New York,

we also gave that same presentation

305

: 00:16:14

to the New York office and we

gave it to a couple other offices,

306

: 00:16:17

and it was, it was well received.

307

: 00:16:19

It's still one of these things where

they, they, it, it's, it's sort of, um, a

308

: 00:16:25

light at the end of the tunnel in a way.

309

: 00:16:27

So we're not just there yet.

310

: 00:16:29

Right, we are thinking about being able

to jumpstart a project, but because we

311

: 00:16:33

can't do it and we can't, you know, put

it on someone's desk and say here, next

312

: 00:16:37

time you have a performer, you know, click

these buttons, you know, it's hard for

313

: 00:16:41

folks to, you know, grasp it really well.

314

: 00:16:46

Like you can, you look at like a new

car that comes out on the market and

315

: 00:16:50

until you drive it, it's hard for any

one of us to say like, oh, it really

316

: 00:16:54

drives really smooth or, you know.

317

: 00:16:56

You know, it has really bad throttle

response or something like that, right?

318

: 00:17:00

So until someone actually gets

to test drive it, there's always

319

: 00:17:03

sort of like, okay, that's nice.

320

: 00:17:05

What's next?

321

: 00:17:06

Okay, that's nice.

322

: 00:17:07

When can we use it?

323

: 00:17:07

Sort of.

324

: 00:17:08

So we're working towards that.

325

: 00:17:09

Yeah.

326

: 00:17:12

Yes.

327

: 00:17:14

Yeah.

328

: 00:17:17

Yeah.

329

: 00:17:20

Randall Stevens: you know, we f we have

the same challenges, which is, I, I

330

: 00:17:23

don't do what you do every day, so I

need your input, I need your feedback.

331

: 00:17:27

And it's almost like you have to

find the, somebody with almost

332

: 00:17:30

a personal interest in it.

333

: 00:17:32

We, that, the way you described it

as light at the end of the tunnel

334

: 00:17:35

is a really good way to say it.

335

: 00:17:36

When you, you know, you show 'em

something and then their wheels,

336

: 00:17:39

you know, start turning, which is

usually what the people kind of.

337

: 00:17:43

To do what we do.

338

: 00:17:44

Uh, you know, we're always imagining

and then you see what sticks.

339

: 00:17:48

It's like, okay, I've got to

go out now and have a bunch

340

: 00:17:50

of conversations around this.

341

: 00:17:51

I've done enough to hopefully get

your wheels turning, but, uh, it's

342

: 00:17:56

really hard to find people that, you

know, that don't want to just say,

343

: 00:18:01

let me know when you're done, you

know, and, and engage in that process.

344

: 00:18:05

So I don't know if you've got any,

any tips or tricks that you've

345

: 00:18:09

learned in doing that within the firm.

346

: 00:18:11

Uh, but, uh, Yeah,

347

: 00:18:18

Charles Portelli: We're part

of the firm but external.

348

: 00:18:20

Um, and I'll, I'll, I'll explain.

349

: 00:18:23

So in a, in a different lifetime, uh,

you know, I started my career kind of

350

: 00:18:28

as kind of like a computational designer

within a firm and I kind of worked more

351

: 00:18:32

on projects, um, and I touched them more.

352

: 00:18:36

But I would go to, I would play soccer

with them on Thursdays, I the happy hours,

353

: 00:18:41

I would go to, uh, the softball games, and

then play softball that I would just go.

354

: 00:18:46

Um, so I did a lot of that sort

of socializing, and you give any

355

: 00:18:52

architect a beer or two and they

just start ranting on like all the

356

: 00:18:56

issues they're running into, right?

357

: 00:18:58

And I'm one of the few people

that'll say, you know, I want to hear

358

: 00:19:01

your complaints because that is an

opportunity for me to make a change.

359

: 00:19:06

Um, so I can tell you like countless

times how on a Thursday after soccer

360

: 00:19:11

you'd be at the bar and someone

would complain about something.

361

: 00:19:15

And then Friday morning, you know,

we stagger into the office and I

362

: 00:19:18

start, you know, churning away.

363

: 00:19:21

And within like a day or so, I

was like, here, install this,

364

: 00:19:24

and let me know what you think.

365

: 00:19:26

And they'd be like, oh, this is

great, but can it do X, Y, Z?

366

: 00:19:28

Or, no, this isn't what I was

talking about, but can it do this?

367

: 00:19:30

So, the iteration that you do as sort of

like a, you know, a software provider,

368

: 00:19:37

you know, it happens only so fast, because

you're kind of at arm's length away.

369

: 00:19:42

But within a firm.

370

: 00:19:44

We're literally desks away, so

that churn can happen pretty fast.

371

: 00:19:47

And within a week, you can have

something that's on someone's desk,

372

: 00:19:52

you know, lightweight, nothing crazy,

that, you know, made them happy.

373

: 00:19:56

Right?

374

: 00:19:56

And you, you want that person.

375

: 00:20:02

Yeah.

376

: 00:20:03

We, we

377

: 00:20:06

Randall Stevens: uh, you know, anytime I'm

in, in their offices, I'm always asking

378

: 00:20:10

what else would you want this to do?

379

: 00:20:11

What, uh, what, what else

problem could it solve?

380

: 00:20:13

And it's like taking notes.

381

: 00:20:15

And, uh, you know, I just had a call the

other day, Kate brought one of those back.

382

: 00:20:19

Engineer coded it up real quick.

383

: 00:20:21

And, you know, we were on the call

with them and tell them about this.

384

: 00:20:25

And all of a sudden we

were sharing the screen.

385

: 00:20:27

They're like, Oh my gosh,

you've already coded it.

386

: 00:20:29

It's like, yeah, if you'll, if you'll

tell us we'll go, you know, but it's,

387

: 00:20:34

it's like pulling teeth sometimes to get

388

: 00:20:36

Evan Troxel: Well, they're all

389

: 00:20:36

conditioned to deal with the

people who don't listen to the

390

: 00:20:41

companies who, who build software, who

actively ignore them as the audience.

391

: 00:20:46

Right?

392

: 00:20:46

So

393

: 00:20:46

Charles Portelli: want to pay

394

: 00:20:47

Evan Troxel: nobody

395

: 00:20:48

Randall Stevens: yeah, you, you have

to be a little bit careful that,

396

: 00:20:52

Charles Portelli: Yeah.

397

: 00:20:52

Yeah.

398

: 00:20:55

Randall Stevens: and you know, all that

kind of stuff, but when it's small things,

399

: 00:20:58

especially small efficiency, things that

can be like, um, you know, added, know,

400

: 00:21:03

quickly to kind of tools that you've

already built, it's like, okay, thank you

401

: 00:21:07

for spending the time to just explain it

402

: 00:21:09

to me.

403

: 00:21:09

And that's all it took for us to

go like, okay, we can actually

404

: 00:21:12

go add value in this way, but

405

: 00:21:15

Charles Portelli: No, totally.

406

: 00:21:15

And

407

: 00:21:15

Evan Troxel: for you about the,

about your, your latest AI endeavors.

408

: 00:21:19

And you've talked about light

at the end of the tunnel.

409

: 00:21:21

So you provide kind of this

vision in your presentation.

410

: 00:21:24

And I'm sure there's some.

411

: 00:21:26

Something somebody could start

doing right now, but, but maybe

412

: 00:21:29

you could just talk about how a

tool like this would, would work.

413

: 00:21:32

So you talked about giving somebody kind

of a great kickoff for their project.

414

: 00:21:38

How would they interact

with, with a tool like that?

415

: 00:21:41

How do you see that working?

416

: 00:21:42

Is it, is it still very large

language model kind of chat based?

417

: 00:21:46

Or is it more of a, an app that

they run and they check boxes And

418

: 00:21:49

something comes, comes back to them?

419

: 00:21:52

Charles Portelli: it's, it's,

it's sort of evolving, right?

420

: 00:21:55

It's kind of, it's kind of

my other word versus dynamic.

421

: 00:21:57

It's evolving.

422

: 00:21:58

Um, so we, we, we have looked at

that kind of like chat based kind

423

: 00:22:02

of efforts and we've showed some

of that during our presentation in

424

: 00:22:06

New York, uh, being able to query a

model and say, you know, so in this

425

: 00:22:11

hospital, how many restrooms are there?

426

: 00:22:13

So you can get a sense of, okay, so

a hospital this size has this many

427

: 00:22:16

restrooms or something like that, or.

428

: 00:22:18

How do you get from, you know,

uh, from the pre op room to the

429

: 00:22:22

op room, uh, in a hospital, right?

430

: 00:22:25

What are, what are the stages,

you know, the cleaning and

431

: 00:22:27

all that other stuff, right?

432

: 00:22:28

Because in some places, there's,

in some projects, there's,

433

: 00:22:32

there's a sequence of a flow.

434

: 00:22:34

Airport is an obvious one, right?

435

: 00:22:36

Like, you, you check in,

436

: 00:22:37

you drop off your bags.

437

: 00:22:39

You go through security and you go to

your gate, right, like that, that is

438

: 00:22:42

a one way sequence and anytime anyone

breaks that one way sequence You're

439

: 00:22:47

either escorted out the building

or you just, you know, people just

440

: 00:22:50

don't like you, all right, having

to go back through the security line

441

: 00:22:56

So from a chat based perspective, you

know, we've tested that and it can work,

442

: 00:23:00

right Um, with regards to sort of like

the app based, you know, sort of options

443

: 00:23:04

and clicking and so on, uh, that's, that's

something we're looking into as well.

444

: 00:23:09

Uh, we have the ability of doing a

quick search and navigating kind of

445

: 00:23:12

the spaces and understanding elements

within spaces and that sort of thing.

446

: 00:23:16

Um, there's no sort of final,

nothing's kind of finalized yet.

447

: 00:23:21

Uh, but as sort of as a grand

scheme, I can envision this

448

: 00:23:25

moment where a client comes, is

awards a project to the architect.

449

: 00:23:30

and says, okay, this is my Performa, this

is my budget, you know, as architects

450

: 00:23:34

we have to design within that Performa

and within, within that budget, right?

451

: 00:23:38

The Performa sometimes can be detailed,

sometimes it can be a little vague, um, if

452

: 00:23:43

you're working with like, for example, a

hotel, you need X amount of hotel rooms,

453

: 00:23:48

this is, you know, the specifics of each

hotel room, you know, the, the king suites

454

: 00:23:53

and, you know, so on and so forth, right?

455

: 00:23:55

Uh, you need so many numbers

of, of these rooms, right?

456

: 00:23:59

The goal would be to kind of feed,

use this Performa as sort of like a,

457

: 00:24:04

uh, the, the initial point of feeding

this application that will then kind

458

: 00:24:08

of take that and say, Okay, based

on past projects that are similar

459

: 00:24:12

size, this is sort of kind of the

breakdown and the setup that you have.

460

: 00:24:16

And then also we would kind

of feed in the building block.

461

: 00:24:20

I've worked on projects where before

where they gave us a Performa and

462

: 00:24:24

we tried to put not perform on

the building lot and said, Hey, it

463

: 00:24:27

doesn't fit, it doesn't work, right.

464

: 00:24:30

It just can't.

465

: 00:24:31

We got within like 98% and like

we, we just can't get the other 2%.

466

: 00:24:36

So there's moments where the

performance isn't a hundred

467

: 00:24:39

percent, you know, up to date.

468

: 00:24:41

Evan Troxel: Mm hmm.

469

: 00:24:42

Charles Portelli: so the idea is

to kind of take those two things,

470

: 00:24:44

you know, the property and the

perform and feed them in and, and

471

: 00:24:46

basically get a rough block model.

472

: 00:24:49

Rough.

473

: 00:24:50

I want, I don't wanna call it massing.

474

: 00:24:52

because I don't want it, I want,

I don't want folks to think it'll

475

: 00:24:55

take away from their creativity, but

a bubble diagram of sorts, right?

476

: 00:24:59

Say these are the spaces, these spaces

need to be adjacent to one another,

477

: 00:25:02

this is ballpark, size of the space

is based on, you know, um, usage

478

: 00:25:08

counts and that sort of thing, right?

479

: 00:25:10

So very quickly bubble diagram

ish of, um, of the building.

480

: 00:25:16

Evan Troxel: It seems too that like when

you upload that proforma or program or

481

: 00:25:20

whatever you want to call it, right?

482

: 00:25:21

It's like there's some kind of analysis

that could happen that says, We've

483

: 00:25:25

typically done these kinds of spaces

when this is the majority of the program.

484

: 00:25:30

Here's what's missing, kind of a thing.

485

: 00:25:31

Cause that happens too, right?

486

: 00:25:33

We get, we get programs

and it's like, there's just

487

: 00:25:35

stuff that's been omitted, not on

488

: 00:25:37

purpose, but it's just,

wasn't, wasn't there.

489

: 00:25:39

And if you, and so then it just

becomes more of a conversation

490

: 00:25:43

back to the client, right?

491

: 00:25:44

So it's like, here's some

prompts for you to take back to

492

: 00:25:46

the client and talk about these

493

: 00:25:48

things.

494

: 00:25:48

It just seems like kind of

a natural progression of

495

: 00:25:50

a, something like

496

: 00:25:51

Charles Portelli: A

497

: 00:25:51

hundred percent.

498

: 00:25:52

Yeah, the less sexy side of kind of

this whole endeavor that we're looking

499

: 00:25:56

at is essentially a QAQC process,

500

: 00:25:59

right?

501

: 00:25:59

Saying, okay, so we have this plot of land

and these are all the spaces and so on.

502

: 00:26:04

And typically for a pre op room,

you have these pieces of equipment

503

: 00:26:09

because you need to, you know, the

surgeons need to prep themselves to to

504

: 00:26:12

go into, um, into the surgery space.

505

: 00:26:16

But your space does not allow for that.

506

: 00:26:19

All these pieces of equipment, so

you need to kind of rethink this.

507

: 00:26:23

And it's less about having the

application kind of try to automate

508

: 00:26:28

rethinking it, and more about

flagging it to the designer, right?

509

: 00:26:32

Um, anything, anytime a per, I feel

like a person's pulled out of the

510

: 00:26:36

process, uh, it, it, it opens it

more for distrust and errors, right?

511

: 00:26:43

So

512

: 00:26:44

keep the designer engaged as it moves on.

513

: 00:26:48

Evan Troxel: Yeah, I think you brought

up building codes earlier too, and

514

: 00:26:50

this is where things I think could

get really, really interesting, right?

515

: 00:26:53

It's because maybe a couple more steps

down the line from the program stage

516

: 00:26:57

that we were just talking about is

when you actually start to come up

517

: 00:27:00

with the stacking diagrams and the

adjacencies and all these things.

518

: 00:27:04

And that's where the building code

plays a huge part in this, because you

519

: 00:27:07

could choose a certain construction

type, and you could choose to sprinkle

520

: 00:27:10

your buildings, and you could have

certain size side yards on projects,

521

: 00:27:13

which, Provide exceptions, right?

522

: 00:27:16

And so depending on occupancy

types, types of spaces, you can

523

: 00:27:20

see where this just leads to.

524

: 00:27:23

And I talked with Shane at, it was

Scott at Upcodes a long time ago, right?

525

: 00:27:27

And it was like, well, what if.

526

: 00:27:29

You could, if it could give you questions

that you need to ask yourselves, your

527

: 00:27:33

design teams, your clients, to say,

okay, if I choose this construction

528

: 00:27:38

type, it allows me to do this.

529

: 00:27:39

But if I choose this construction type,

it allows me to do this and this and this.

530

: 00:27:43

And then that becomes a decision

point that you have to make, like it's

531

: 00:27:46

not going to make it for you, right?

532

: 00:27:48

And so it is inclusive, but it's

also prompting you to things

533

: 00:27:52

you may not have been aware of.

534

: 00:27:54

Um, just by making a decision like that

early on, it has huge impacts on the

535

: 00:27:58

outcome of the project, what you can

do, how many floors you can go, how big

536

: 00:28:02

your areas can be, et cetera, et cetera.

537

: 00:28:04

And this is architecture, right?

538

: 00:28:05

Every municipality out there has some

weird adoption of the building code.

539

: 00:28:09

You know, there's, there's something

like 30, 000 municipalities and they

540

: 00:28:13

all have a different recipe of the

codes that they are using right now.

541

: 00:28:18

And so that becomes part of what you're

feeding potentially into the system

542

: 00:28:22

to say, Okay, this is the project.

543

: 00:28:24

This is the location.

544

: 00:28:25

This is the code cycles that

we're going to be using.

545

: 00:28:28

These are the codes and here's the

program and it, it just seems to me like

546

: 00:28:34

there's a, there's a lot of opportunity

here to, to be that co creator along

547

: 00:28:39

the path of, of the design phases.

548

: 00:28:43

Charles Portelli: Yeah, a hundred

percent, a hundred percent, I think,

549

: 00:28:45

and it's, you know, the designer's

still having that kind of control

550

: 00:28:49

and sort of being, sort of, uh,

made aware of some of these things.

551

: 00:28:54

The other, kind of, sort of, reason

behind some of this is not, not to just

552

: 00:28:59

automate and speed things up, right, and

allow designers to do more, is that there,

553

: 00:29:04

there's, there's a lot of macro decisions

that are made on projects and there's a

554

: 00:29:07

lot of micro decisions made on projects.

555

: 00:29:09

Um, I argue that the macro decisions

are actually quite easy, because they're

556

: 00:29:14

usually done by committee, or there's

some really, um, hard and fast rules,

557

: 00:29:19

like, this is what we're doing, and

this is why we're doing it, right?

558

: 00:29:22

Whereas, uh, the micro decisions are

things that designers do at their

559

: 00:29:26

desk on a day to day basis, right?

560

: 00:29:29

And if these errors of the micro

decisions start to kind of, once they

561

: 00:29:34

start to add up, you know, hear They

can lead to rework and things like that.

562

: 00:29:38

Um, and a lot of times to solve these

market decisions, you have to go to the

563

: 00:29:42

senior person in the office and say,

Hey, you know, this is the condition I'm

564

: 00:29:45

dealing with, how would you address it?

565

: 00:29:47

Or, this is what I'm facing, do you have

any ideas or any, you know, precedence

566

: 00:29:51

of projects that we've done in the

firm that can do this and so on, right?

567

: 00:29:55

But if anyone's worked in an office before

and from the junior level all the way

568

: 00:29:58

up, right, you come to realize that that

senior person isn't always available.

569

: 00:30:02

It's because they're

working on many things.

570

: 00:30:05

They're traveling, you know, we all

have lives, we all have priorities

571

: 00:30:09

and, you know, everything needs

to be addressed right from, you

572

: 00:30:11

know, addressing client needs to

addressing project needs and so on.

573

: 00:30:16

So if we can have this setup that

can allow for, you know, dealing with

574

: 00:30:21

some of these micro decisions, right?

575

: 00:30:24

We don't have to worry about

kind of veering off too

576

: 00:30:26

much off on the mark, right?

577

: 00:30:28

Oh, someone missed, you

know, these rooms or.

578

: 00:30:31

This is undersized, or something like

that, or oversized, you know, and then

579

: 00:30:34

you have to go through a VE exercise,

which is sort of like blasphemy, right?

580

: 00:30:39

Evan Troxel: Right.

581

: 00:30:39

Yep.

582

: 00:30:40

And

583

: 00:30:42

Charles Portelli: Yeah,

584

: 00:30:45

Randall Stevens: because it seems

to be a natural kind of way to

585

: 00:30:48

think about these kinds of things.

586

: 00:30:50

You want some, little co pilot

doesn't mean there's still not

587

: 00:30:53

another part of the process, but you'd

love to have right along, another

588

: 00:30:58

personality right along with you.

589

: 00:31:00

Maybe that knows a little bit more about

certain things and can be kind of push

590

: 00:31:04

you in certain kinds of directions.

591

: 00:31:06

And that's what a lot of these

discussions end up being.

592

: 00:31:09

It's a very natural thing to think about.

593

: 00:31:10

It's not going to, it's not the end

all, it's not going to make all this

594

: 00:31:14

stuff, but it's going to help kind of

guide me and propel me, me, challenge

595

: 00:31:18

me, help me do all those kinds

of, uh, those kinds of adjectives.

596

: 00:31:23

Charles Portelli: it's all about

me as the architect, right, as

597

: 00:31:25

the designer, and not necessarily

taking me out of the picture, right?

598

: 00:31:30

And sometimes I even pull back from

the idea of co pilot, because the idea

599

: 00:31:34

of co pilot means that there's someone

else there that if the pilot can't

600

: 00:31:37

navigate, The co pilot can take on, right?

601

: 00:31:40

So

602

: 00:31:40

I, I, I, I kind of even step it down a bit

further and say, It's an assistant, right?

603

: 00:31:46

Like, so this is gonna,

it's gonna help me, right?

604

: 00:31:50

Randall Stevens: Yeah.

605

: 00:31:50

And you Yeah, you're right.

606

: 00:31:52

And, and that, that's an example

where language matters to somebody.

607

: 00:31:55

It's like, are you gonna

get pushback from this?

608

: 00:31:57

Or is it your assistant

609

: 00:31:59

this, you know, equal?

610

: 00:32:00

And it's like, well, they're not

equals it's uh, uh, well, one of

611

: 00:32:04

the things, uh, uh, I know Charlie,

when you were presenting, uh,

612

: 00:32:09

you know, at the New York event.

613

: 00:32:11

We've talked a lot about the data

side of it and where you all started

614

: 00:32:14

around trying to, you know, put

some structure to all of that.

615

: 00:32:17

But you also talked some about,

um, how you're using, looking

616

: 00:32:21

at the image generation.

617

: 00:32:23

I know you all were experimenting

with putting some new front ends,

618

: 00:32:26

uh, What's that tool called?

619

: 00:32:28

Comfy UI or something that

you're putting in front of

620

: 00:32:31

Charles Portelli: Yeah.

621

: 00:32:32

Randall Stevens: And maybe you can

talk a little bit about that side

622

: 00:32:34

of it, um, you know, uh, geometry

generation and or image generation

623

: 00:32:39

and, and where, where you'll see

624

: 00:32:41

Charles Portelli: Yeah, the

image generation is sort of,

625

: 00:32:43

uh, like low hanging fruit.

626

: 00:32:44

It's a no brainer at this point.

627

: 00:32:46

Almost everyone's doing it.

628

: 00:32:48

Um, We generate so many images in our,

in our industry that it's, it's, again,

629

: 00:32:53

it's a no brainer, right, to kind of

allow you to kind of iterate very quickly.

630

: 00:32:58

You do have to kind of read through the

fine print of who owns it and what's your

631

: 00:33:04

sort of contractual agreement with your

client and, you know, all that other fun

632

: 00:33:08

stuff, right, so do, do talk to kind of

counsel about all that stuff, um, and user

633

: 00:33:14

license agreements and all that, all that.

634

: 00:33:16

Um, But yeah, so we are using

Comp UI, um, for image generation.

635

: 00:33:21

We can take screenshots out of

Rhino, for example, of a mass ink,

636

: 00:33:25

and we can kind of very quickly

iterate ideas on top of it, right?

637

: 00:33:28

Um, so that's kind of nice and handy.

638

: 00:33:31

We can apply different styles to it

and so on, and we can use that for,

639

: 00:33:35

um, for presentations and so on.

640

: 00:33:38

We've gone through the fine print, right?

641

: 00:33:40

Um, it's all run locally, so none of

it is, um, It's cloud based, right,

642

: 00:33:46

so we're not breaching any security

requirements that clients ask us to

643

: 00:33:52

abide by, right, so that's totally fine.

644

: 00:33:55

It's not feeding back into, you know,

global models and so on, right, so it

645

: 00:34:03

kind of, which is still kind of like an

open question about like IP discussions

646

: 00:34:08

and all that stuff, that's beyond me.

647

: 00:34:11

Um, so it's, it's a safe.

648

: 00:34:13

Playground, it's a safe environment

that, yeah, kind of iterate,

649

: 00:34:15

play around with, and so on.

650

: 00:34:20

Randall Stevens: since you brought that

up, I think there's probably still a

651

: 00:34:23

lot of, um, I know we talked about it

at the event, but, I You know, it takes

652

: 00:34:29

a lot, a lot of data to train a model.

653

: 00:34:33

And a lot of the kinds of stuff

that we're talking about doing

654

: 00:34:35

is not really training models.

655

: 00:34:37

We're using existing models to

then exercise on training models.

656

: 00:34:43

local data, uh, in the example.

657

: 00:34:45

Maybe you can, uh, kind of

talk a little bit about that.

658

: 00:34:48

Even, you know, even at the scale of a

Perkins will, one of the largest firms

659

: 00:34:51

in the world, do you all ever perceive

that you'd be able to have enough data

660

: 00:34:55

to even train your own model, or is it

always going to be to leverage these

661

: 00:34:59

large models, you know, public models

and, and execute something with it?

662

: 00:35:04

Charles Portelli: I think our

current, um, our current sort

663

: 00:35:08

of consensus is more the latter.

664

: 00:35:10

There are a lot of models, they're

trained on, you know, probably

665

: 00:35:15

trillions of images, who knows, right?

666

: 00:35:17

Um, and we're gonna sort of just

use those right now, because, you

667

: 00:35:22

know, they're definitely more robust.

668

: 00:35:25

we feel like anything that we

could probably put together.

669

: 00:35:28

Now there's other firms that

argue differently, right?

670

: 00:35:30

I've seen presentations by other firms

that say, um, Hey, we can, you know,

671

: 00:35:37

we've trained our own model and yada,

yada, yada, and so on and so forth.

672

: 00:35:40

And, you know, Godspeed.

673

: 00:35:41

I think, uh, certain firms

have very specific styles that

674

: 00:35:47

can lend themselves to that.

675

: 00:35:49

So that it, it almost as if

it's building a heavy bias.

676

: 00:35:53

into, um, into that, which is

good for them because, you know,

677

: 00:35:58

they want to generate images

based on their style and that's,

678

: 00:36:01

Randall Stevens: the same.

679

: 00:36:02

the Frank Gehry's of the world,

the Zaha Hadid's have a very,

680

: 00:36:05

you know, kind of iconic style.

681

: 00:36:07

Charles Portelli: exactly, exactly, yeah.

682

: 00:36:09

Um, so yeah, so things like that.

683

: 00:36:11

So that's sort of where, where we

are and how we're looking at it.

684

: 00:36:14

Um, and, uh, it's kind

of where we are today.

685

: 00:36:19

Nothing saying we may not train

our own models on our own images

686

: 00:36:23

and so on because we do generate.

687

: 00:36:25

A bunch of images regularly for

projects, but to be determined, right?

688

: 00:36:30

Um, as it gets easier to do some of

these things, it might be easier to

689

: 00:36:38

train our own models and so on, right?

690

: 00:36:40

Yeah.

691

: 00:36:44

Randall Stevens: but I think, The

current state is it's, it takes

692

: 00:36:48

a lot of data to train a model.

693

: 00:36:50

And even at the scale

of, yeah, it's expensive.

694

: 00:36:52

And that's why you see these

companies, the main, the large

695

: 00:36:56

organization spending billions of

dollars for, you know, that's usually,

696

: 00:37:02

uh,

697

: 00:37:02

time, it's like it takes, and

that's just because they're

698

: 00:37:05

ingesting so much information.

699

: 00:37:07

Um, but, uh, yeah, I think it's, uh,

I think a lot of what we're going

700

: 00:37:11

to be talking about over the next.

701

: 00:37:13

Next couple of years probably is just

educating, helping everybody understand

702

: 00:37:18

that when we're talking about, you know,

putting an image in or asking it to do

703

: 00:37:23

something, we're not training a model.

704

: 00:37:25

We're just ask, we're

using the model to do some

705

: 00:37:29

on top of the data that we're giving it.

706

: 00:37:31

And, and that, you know, you have to,

you have to dig into what's really going

707

: 00:37:35

on to start to understand and to build

a mental map about what's going on.

708

: 00:37:38

And, and then that gets into

the overcoming the fear.

709

: 00:37:41

We're not giving our intellectual

properties, not going into, you know,

710

: 00:37:45

the, engine that's, you know, another

firm's going to now be able to, you know,

711

: 00:37:49

Charles Portelli: Yeah.

712

: 00:37:50

And even that, that sort of intellectual

property is kind of, you could probably

713

: 00:37:54

have a conference on just that alone.

714

: 00:37:56

Randall Stevens: Right.

715

: 00:37:57

Charles Portelli: So it's kind of

open and, you know, once an image is

716

: 00:38:02

put out there on the internet, it's,

you know, it's accessible, right?

717

: 00:38:05

So

718

: 00:38:06

Randall Stevens: Sure.

719

: 00:38:07

Charles Portelli: whether they

train it beforehand or after

720

: 00:38:09

the fact, you know, whatever.

721

: 00:38:10

That being said, an

image is an image, right?

722

: 00:38:13

You still have to build

the building, right?

723

: 00:38:15

Like, so you still have to

detail it, you still have to get

724

: 00:38:17

fabricators involved, you still have

to get all these folks involved.

725

: 00:38:21

So just because you can, you know,

generate a new Gaudi building.

726

: 00:38:24

It doesn't necessarily mean that

you're actually going to be able to,

727

: 00:38:28

you know, build it, um, and so on.

728

: 00:38:30

So it takes, it takes more than just that.

729

: 00:38:33

Evan Troxel: The interesting thing

about this whole movement with

730

: 00:38:35

AI that we've been seeing is this

idea of regurgitation, right?

731

: 00:38:39

Like the, the Things tend to get

more and more watered down as the,

732

: 00:38:44

the thing, the, the information it

spits out then goes back in to train

733

: 00:38:48

it even further, right, depend, with

a little bit of feedback to maybe push

734

: 00:38:53

it in one direction versus another,

but it's the same with images, right?

735

: 00:38:56

It's like, If you train it all on

your own images, you are regurgitating

736

: 00:39:01

very similar ideas, and sometimes

projects call for that, right?

737

: 00:39:05

It depends, is an answer that you'll hear

a lot when it comes to this stuff, like,

738

: 00:39:09

should I use, should we train it on ours?

739

: 00:39:12

Well, it depends.

740

: 00:39:13

Like, you want options, right?

741

: 00:39:16

When you're a designer.

742

: 00:39:18

And you want inspiration sometimes,

sometimes you do want the fastest path

743

: 00:39:23

down the road, sometimes you, you want

to take the circuitous, you know, scenic

744

: 00:39:27

byway to, to start to inform your ideas.

745

: 00:39:30

So, I think for, for a lot of

people, they're looking for novel

746

: 00:39:35

ideas, and we saw this very early on

with, uh, What seemed to be pretty

747

: 00:39:40

exceptional prompt engineering, right?

748

: 00:39:43

Which was a facade made of feathers and

like people just exploring weird mashups

749

: 00:39:48

that you wouldn't normally have done.

750

: 00:39:50

You wouldn't have never taken the time

to design something like that before

751

: 00:39:54

because of how tedious it would have

752

: 00:39:57

been to model and texture and

light all of those things.

753

: 00:40:01

And so in the pursuit of novelty, I think

it was, it was pretty phenomenal, right?

754

: 00:40:05

Because it's not just trained on

buildings, it's trained on images, right?

755

: 00:40:10

Across the board.

756

: 00:40:12

and and then it's taking that and

using your prompt to kind of guide it

757

: 00:40:16

and come up with, with novel things.

758

: 00:40:18

I think that that is what

caught everybody's eye.

759

: 00:40:21

It's like, whoa, what?

760

: 00:40:22

Because nobody would have taken

the time to figure that out.

761

: 00:40:25

or to do the manual labor to

actually do that earlier on.

762

: 00:40:29

But, but this whole idea of training

your own model on your own things,

763

: 00:40:33

it might just be too far down

the regurgitation path as well.

764

: 00:40:37

Right?

765

: 00:40:37

And, and if architecture is usually,

I mean, I, I guess I really can't,

766

: 00:40:42

can't say usually, but let's just

go with it for a moment, right?

767

: 00:40:45

It's like, it's usually look in the

pursuit of, of new, fresh, I mean,

768

: 00:40:49

that's what we do when we practice,

we're looking for fresh ideas.

769

: 00:40:53

And it is a practice like we are

continually evolving and getting, you

770

: 00:40:57

know, going, going in different ways.

771

: 00:40:58

And so, uh, it, it is something that I

think firms need to think pretty deeply

772

: 00:41:02

about because it does cost a lot of money.

773

: 00:41:04

And do you want to spend all the time

to do something like that so that

774

: 00:41:07

you can just come out with the same

architecture over and over and over again?

775

: 00:41:11

Like that's something

you kind of have to ask.

776

: 00:41:13

Charles Portelli: Yeah,

no, a hundred percent.

777

: 00:41:14

And I think, um, it, it, you know,

it helps take the stick back and say,

778

: 00:41:20

you know, this was all trained up

and dated on, on the internet, right?

779

: 00:41:23

Now, I'm not going to get into

like truthfulness or not, or

780

: 00:41:27

any of that nonsense, but.

781

: 00:41:28

There was a point in time where we were

all taught not to believe everything

782

: 00:41:32

that's on the internet, right?

783

: 00:41:33

So, come on, like, we should also

be, you know, vetting a lot of

784

: 00:41:38

this that comes out our way, right?

785

: 00:41:41

Not every image on the internet

is true, not all the text on the

786

: 00:41:44

internet is factual, or it might have

been factual for that state in time,

787

: 00:41:49

but then research evolved

and led to believe that.

788

: 00:41:53

Actually, with a modification in medicine

and so on, that like now this is like the

789

: 00:41:58

more modern fact and modern treatment for

an ailment or something like that, right?

790

: 00:42:02

So, something may be factual for

its own period of time and evolve.

791

: 00:42:04

So like we, we have to be

a little more, we have to

792

: 00:42:07

remember it's artificial, right?

793

: 00:42:09

So, keep that in mind.

794

: 00:42:12

Randall Stevens: So from, uh, you know,

with your head down working on these kinds

795

: 00:42:16

of things every day, Charlie, what's,

uh, what's, what's got you excited?

796

: 00:42:20

What do you, what do you kind of foresee

is, You know, something over the next

797

: 00:42:23

6 12 months that you think you're

going to be working on that keeps you

798

: 00:42:27

excited because, you know, a lot of

times you're probably, you're probably a

799

: 00:42:31

year or two ahead of of what people are

able to, uh, you know, even put their

800

: 00:42:37

Charles Portelli: I don't know,

we'll see, hopefully, right?

801

: 00:42:39

Um, I don't know, it's hard to

get excited these days, there's

802

: 00:42:42

like not much, I don't know.

803

: 00:42:45

And I don't mean that in like

any like sort of condescending

804

: 00:42:49

way or anything like that.

805

: 00:42:51

It's just sort of like

806

: 00:42:54

Anyone can easily get excited and

I think that's probably one of the

807

: 00:42:58

big problems is that there's so much

flashiness and all this stuff that like

808

: 00:43:02

Anyone can easily get excited and I

think that's probably like if you think

809

: 00:43:05

about the the Gartner hype cycle I was

explaining it to our intern earlier today.

810

: 00:43:10

I think that that excitement is like

Anyone can easily kind of jump on that

811

: 00:43:14

bandwagon, and I think it's cutting

through all the fluff and seeing

812

: 00:43:18

sort of the real potential to it.

813

: 00:43:19

I think that kind of excites me.

814

: 00:43:22

Randall Stevens: The

pragmatic side, right?

815

: 00:43:24

Charles Portelli: like, I mean,

anyone can make a pretty picture,

816

: 00:43:27

and it's actually gotten way easier

to make a pretty picture now, right?

817

: 00:43:30

But the problem is, can you

build the building, right?

818

: 00:43:33

Because I've worked on a lot of projects

prior to, you know, this, this wave of

819

: 00:43:38

AI prevalence in our, in our industry,

and some of them didn't get built.

820

: 00:43:43

But, some of them did get built.

821

: 00:43:44

Actually, quite a few of them got built.

822

: 00:43:46

Um, and those are sort

of the exciting ones.

823

: 00:43:48

Like, I go for a run in the morning, and

I run past one of my towers that I built.

824

: 00:43:52

And I don't, it's not like

I'm waving a flag saying, hey,

825

: 00:43:54

I built this, or anything.

826

: 00:43:55

But it's just something

for me, like, oh, wow.

827

: 00:43:57

I remember those facade panels.

828

: 00:43:58

I remember those stack joints.

829

: 00:44:00

I remember the, the BMU tracks,

and all that other stuff, right?

830

: 00:44:03

Like, so, we do have to

build, build buildings, right?

831

: 00:44:06

We're building them

for, for humans, right?

832

: 00:44:08

Um, so if we kind of can't build

these buildings, right, and sort

833

: 00:44:13

of ask, like, what's the point?

834

: 00:44:15

So I look at that pragmatic side of it.

835

: 00:44:17

In terms of, like, what gets me excited,

I think, you know, this, this effort

836

: 00:44:20

of being able to quickly jumpstart

a project and have sort of a design

837

: 00:44:24

assistant running with designers,

right, in conjunction, so that we can

838

: 00:44:31

be a little more intentional with our

efforts, a little more informed with

839

: 00:44:35

our efforts, Part of this is not just,

you know, space requirements and so

840

: 00:44:40

on, but also material requirements.

841

: 00:44:42

You know, Evan, you talked about what

if we use, you know, concrete versus

842

: 00:44:46

steel versus timber or whatever, right?

843

: 00:44:48

Those are not just quantity, um, you

know, uh, significances, but also

844

: 00:44:54

spatial, you know, considerations

have to be taken into account.

845

: 00:44:57

Concrete, you can only span so much.

846

: 00:44:59

Steel, you can span so much.

847

: 00:45:00

Timber, you can span so

much and so on, right?

848

: 00:45:02

Like, so being able to

understand material quantity.

849

: 00:45:05

And now if you understand material

quantities, you can start to take

850

: 00:45:07

into account carbon, embodied carbon,

and so on and so forth, right?

851

: 00:45:11

So, that kind of excites me, being

able to stitch together a lot of these

852

: 00:45:18

things, um, or these topics that we

take into consideration, but right

853

: 00:45:24

now it takes a lot of mental effort

to kind of stitch them together.

854

: 00:45:28

Like, okay, we have a concrete

building, how much concrete do I have?

855

: 00:45:30

Well, I gotta go figure it out.

856

: 00:45:32

Okay, which concrete

supplier are we using?

857

: 00:45:35

These are the ones that the contractor are

recommending, and the engineer recommends

858

: 00:45:40

a specific PSI, so we have to kind of

do a lot of legwork to really get to

859

: 00:45:46

where we are, um, or where we want to be.

860

: 00:45:49

So I think if we can cut down on a

lot of that legwork, that's, that kind

861

: 00:45:52

of excites me a little bit, right?

862

: 00:45:54

We can be informed about our decisions,

and if the client says, change the

863

: 00:45:57

building from concrete to steel,

we could say, okay, that could be

864

: 00:46:00

done, the implications are this much

CO2, spatial requirements change,

865

: 00:46:07

you know, so on and so forth, right?

866

: 00:46:09

Which has happened, I've worked

on a project where we went

867

: 00:46:12

Evan Troxel: Oh, for sure.

868

: 00:46:13

Charles Portelli: one

material to another, yes.

869

: 00:46:15

Evan Troxel: Right.

870

: 00:46:16

Randall Stevens: If I, if I remember

right, you kind of posed the question,

871

: 00:46:20

uh, when we were all together, uh,

at this event in April of, you know,

872

: 00:46:25

for the group, how, how is AI gonna,

you know, affect both, you know,

873

: 00:46:29

the profession, the, the person, you

know, working in this profession, uh,

874

: 00:46:35

uh, you know, the individual versus

the, the career of that individual

875

: 00:46:39

versus the profession as a whole.

876

: 00:46:41

So I think a lot about

what you just said is.

877

: 00:46:45

Like those are positive

directions for the profession.

878

: 00:46:48

You're going to be better stewards

of resources, energy, all those kinds

879

: 00:46:52

of things around the problem solving.

880

: 00:46:55

Um, maybe you can, um, what have you

seen with the work that you're already

881

: 00:47:00

doing, uh, or are you far enough yet to

kind of understand how at an individual

882

: 00:47:04

level that these tools are going to help.

883

: 00:47:07

know, an individual in their career,

what, what, what I maybe a question

884

: 00:47:11

would be, what advice would you give to

a architecture student today about what

885

: 00:47:15

they should be thinking about and doing

because of what, what's going to, what

886

: 00:47:19

the profession is going to look like in

887

: 00:47:21

Charles Portelli: Oof,

that's a hard one, um,

888

: 00:47:23

and it

889

: 00:47:23

Evan Troxel: Stay nimble.

890

: 00:47:24

Stay nimble,

891

: 00:47:25

kids.

892

: 00:47:27

Charles Portelli: Um, and it

actually should be an easy question

893

: 00:47:28

because I, I teach, so I should be

like, you know, teaching folks in

894

: 00:47:32

the future, but, um, It's tricky.

895

: 00:47:35

I think, uh, it, and it also

really depends on, on, on sort

896

: 00:47:38

of their career path, right?

897

: 00:47:39

Whether it's a technology career

path, a design career path,

898

: 00:47:42

or a technical career path.

899

: 00:47:43

So to kind of, I think the staying

nimble is definitely sort of key.

900

: 00:47:49

Sort of being able to ask the

right questions, um, is important.

901

: 00:47:54

Um, kind of like what is your end goal?

902

: 00:47:58

Um, keep that clear and be

able to kind of, you know, Find

903

: 00:48:04

a path to get to that goal.

904

: 00:48:06

Why?

905

: 00:48:07

Um, because today, you know, we,

we do what we do as architects.

906

: 00:48:12

Twenty years ago, I doubt that

architects were taking into account

907

: 00:48:17

the amount of operational carbon and

embodied carbon that goes into projects.

908

: 00:48:22

I doubt architects were taking

into account, you know, the

909

: 00:48:26

reusability of the building, uh,

repositioning of projects and so on.

910

: 00:48:31

There was always presentations

where they said, Oh yeah, and you

911

: 00:48:33

can scale your building up this

way, or you can, you know, expand

912

: 00:48:37

this way, and that sort of thing.

913

: 00:48:38

But it was always within

the same typology, right?

914

: 00:48:40

Here's an office building, you make a

bigger office building down the road,

915

: 00:48:44

here's a residential, you build a bigger

residence down the road, sort of thing.

916

: 00:48:48

And now we're in a position where

we're realizing that, um, things

917

: 00:48:52

can change quite significantly.

918

: 00:48:54

So, being able to kind of wrap

our heads around that more,

919

: 00:48:59

It's going to be incredible.

920

: 00:49:00

The architect of the future is going to

be able to do more than we can do today.

921

: 00:49:06

Evan Troxel: Something you said earlier,

Charlie, I think would also be great

922

: 00:49:08

advice, which is how you went to the

soccer games and went to the bar and hung

923

: 00:49:13

out and asked a lot of questions, right?

924

: 00:49:15

And like that, the idea of just

being curious and, and Randall,

925

: 00:49:19

you spoke about this too, right?

926

: 00:49:20

Visiting clients in LA.

927

: 00:49:22

When you're, when you're brainstorming

ideas, in Randall's case, or when

928

: 00:49:27

you're just trying to get at, like, what

people hate about their job, Charlie,

929

: 00:49:32

like, for an example, right, it's

like, like, what pisses you off about

930

: 00:49:35

the practice or the technology or this

part of the design process or whatever?

931

: 00:49:40

I mean, that, that, to me,

shows that, like, you always

932

: 00:49:45

want to be learning about it.

933

: 00:49:47

something, right?

934

: 00:49:47

And, and just the ability to

ask questions and be curious.

935

: 00:49:52

I think a lot of times architects are

trained in a vacuum to solve the problems

936

: 00:49:56

and figure things out themselves.

937

: 00:49:58

And it's amazing what you can

learn when you listen, right?

938

: 00:50:02

And, and ask good questions.

939

: 00:50:03

If, if you're just trying to advance

your career, then you're doing the same

940

: 00:50:08

thing to a more senior person, right?

941

: 00:50:10

You're asking a lot of questions.

942

: 00:50:12

You're trying to figure out why

they would make a decision in this.

943

: 00:50:16

Particular instance versus,

you know, a different decision

944

: 00:50:19

in a different instance.

945

: 00:50:20

Right?

946

: 00:50:20

So I think when it, because,

because it goes against the way

947

: 00:50:24

that we're trained, which is heads

down, figure it out, get it done.

948

: 00:50:29

And, and you make a lot of decisions

without any input in those circumstances.

949

: 00:50:34

You get out into the world, right?

950

: 00:50:36

Every project is a team sport.

951

: 00:50:37

Every project involves

multiple stakeholders.

952

: 00:50:41

And, and there are a lot of people doing

it who know a lot more than you do, right?

953

: 00:50:46

And so it's, it's always going into

these situations with an open mind,

954

: 00:50:50

with a lot of curiosity, asking

great questions to real people, and

955

: 00:50:55

then listening to their answers.

956

: 00:50:57

And, and again, I think architects

are often in the position of

957

: 00:51:01

having to have the answers.

958

: 00:51:03

And, and that, I think, is something

you want to be very careful about

959

: 00:51:06

is, is just, Again, being open

minded and curious to find out why

960

: 00:51:11

you may not know the best thing.

961

: 00:51:12

And even if you do know, you

don't need to spread it out.

962

: 00:51:15

You don't need to just say

it to stop the conversation.

963

: 00:51:18

I think a lot of times keeping the

conversation going is going to be a lot

964

: 00:51:22

more enlightening and get you somewhere

great than if you just simply accept

965

: 00:51:28

what somebody says and say, that's the

answer, let's move on to the next thing.

966

: 00:51:31

So maybe there's a time and place for

both of those, but I would just, I would

967

: 00:51:34

err on the side of curious and openness.

968

: 00:51:36

Charles Portelli: Oh, totally, 100%.

969

: 00:51:38

And I was sort of forced into that sort

of method of thinking more, deeper,

970

: 00:51:45

into that method of thinking when I was

a product manager at a tech startup.

971

: 00:51:49

Right, so I was doing this sort of

internally at a, at an architecture firm,

972

: 00:51:53

but, cause it made, it made sense to me,

it was sort of common sense, but then

973

: 00:51:56

when I worked at a tech startup, you

know, I realized I was doing that already,

974

: 00:52:00

but I was really kind of, like I said,

forced to go deeper down that path of

975

: 00:52:05

not making assumptions, but really kind

of, um, fleshing out a lot of these, uh,

976

: 00:52:12

hurdles that people are running into and

sort of, kind of, they're, they're sort

977

: 00:52:15

of, you know, What they, what their kind

of goals are and how they want to work.

978

: 00:52:21

And then it's also worth kind

of fleshing out even more so,

979

: 00:52:23

like, how often does this happen?

980

: 00:52:25

We get, we, I used to get

a lot of one off requests.

981

: 00:52:28

It's like, oh, I'm working on a project

that's like XYZ, can you help me?

982

: 00:52:32

It's like, alright, so how,

how often do you run into this?

983

: 00:52:35

What's, what's the situation?

984

: 00:52:36

Sort of, because it's also

understanding how did you get to this

985

: 00:52:43

state where you're in a kind of jam?

986

: 00:52:46

Right?

987

: 00:52:46

For lack of a better word, you know,

and is the solution a Band Aid just to

988

: 00:52:53

get you to the next hurdle, you know,

taking the same baggage with you, or

989

: 00:52:58

should you rewind and find that point

where you got on the off ramp, you

990

: 00:53:03

know, accidentally and kind of get

yourself, you know, course correct

991

: 00:53:07

and kind of progress forward, right?

992

: 00:53:09

And then, and also depends on time of.

993

: 00:53:11

Uh, for developing the solution,

you know, so on and so forth.

994

: 00:53:14

If you give me two days, I'm going to

give you a two day solution, right?

995

: 00:53:17

If you give me two months, I'm going to

give you a two month solution, right?

996

: 00:53:20

And you can course correct and everything.

997

: 00:53:23

Randall Stevens: That brings

up a good, a good Charlie.

998

: 00:53:26

Like, the assumption with the, you know,

of applying technology, you know, to

999

: 00:53:33

the industry is it's at least started,

a lot of it is talked about as, as

: 1000

00:53:39,243 --> 00:53:43,963

efficiency gains, automate, automating

mundane tasks or things that, you know,

: 1001

00:53:43,963 --> 00:53:46,703

you would have had to have manual,

a person would have manually done.

: 1002

00:53:47,333 --> 00:53:51,323

guess, I think it's largely

true that that's happening.

: 1003

00:53:52,043 --> 00:53:55,893

I guess a question that we have to

ask though is, The assumption is,

: 1004

00:53:55,893 --> 00:53:59,043

is that's going to free me up to

do higher order kinds of things.

: 1005

00:53:59,133 --> 00:54:01,973

Is, is that Are you?

: 1006

00:54:02,863 --> 00:54:05,483

Charles Portelli: Uh, short answer, no.

: 1007

00:54:05,793 --> 00:54:08,063

Long answer, possibly, right?

: 1008

00:54:08,133 --> 00:54:08,683

Like, so,

: 1009

00:54:09,073 --> 00:54:09,443

Randall Stevens: Yeah.

: 1010

00:54:09,893 --> 00:54:13,243

Charles Portelli: I can, if I can automate

a specific task that now takes an eight

: 1011

00:54:13,243 --> 00:54:18,203

hour task and brings it down to six

hours or seven hours, it doesn't mean

: 1012

00:54:18,203 --> 00:54:20,773

you get to go home an hour early, right?

: 1013

00:54:20,773 --> 00:54:20,793

Like,

: 1014

00:54:21,353 --> 00:54:24,383

It's not gonna, it's not gonna

really flesh itself out that way.

: 1015

00:54:25,063 --> 00:54:27,123

Randall Stevens: But maybe you

get to think about the project for

: 1016

00:54:27,123 --> 00:54:28,763

those two hours in a different way,

: 1017

00:54:29,043 --> 00:54:30,533

Charles Portelli: Correct,

a hundred percent right.

: 1018

00:54:30,543 --> 00:54:33,593

So you can, you can now look at

things slightly differently, right?

: 1019

00:54:33,593 --> 00:54:38,313

And that's where sort of analytics kind

of helps kind of take the coin and be

: 1020

00:54:38,313 --> 00:54:39,803

able to see it from the other side, right?

: 1021

00:54:39,813 --> 00:54:47,453

You can now build projects with more

agile teams, more informed teams.

: 1022

00:54:47,563 --> 00:54:51,363

I was talking to my intern this morning

and I said, one of the projects I worked

: 1023

00:54:51,363 --> 00:54:52,793

on, which is, it falls in the top five.

: 1024

00:54:53,083 --> 00:54:54,793

Ten tallest towers in the world.

: 1025

00:54:55,173 --> 00:54:58,203

At one point in time, there was only

five of us working on it, right?

: 1026

00:54:58,213 --> 00:55:02,933

So like five people built or helped design

one of the tallest skyscrapers, right?

: 1027

00:55:02,983 --> 00:55:06,333

And it's because we had like a

really fleshed out system, right?

: 1028

00:55:07,393 --> 00:55:10,643

We were able to do more

with less, so architects are

: 1029

00:55:10,643 --> 00:55:11,583

going to be able to do more.

: 1030

00:55:12,303 --> 00:55:15,223

When I say do more, they are taking

into account considerations like

: 1031

00:55:15,603 --> 00:55:19,773

material quantities, carbon, Where

are the materials coming from, right?

: 1032

00:55:19,773 --> 00:55:20,913

How sustainable are they?

: 1033

00:55:21,013 --> 00:55:21,653

And so on.

: 1034

00:55:22,023 --> 00:55:24,483

How do we reposition the

building in the future?

: 1035

00:55:24,853 --> 00:55:26,303

Design for disassembly, right?

: 1036

00:55:26,313 --> 00:55:29,943

These are all topics that are

conversational right now, and

: 1037

00:55:29,943 --> 00:55:36,013

we're trying to take into our

practice that weren't before, right?

: 1038

00:55:36,013 --> 00:55:41,323

So as we, as the practice evolves,

we're going to do more and we're

: 1039

00:55:41,323 --> 00:55:43,193

probably going to have to do

it in the same amount of time.

: 1040

00:55:45,083 --> 00:55:48,273

Randall Stevens: I think, uh, uh, you

know, one of the things that you kind of

: 1041

00:55:48,363 --> 00:55:53,923

wrapped up with when we did the, uh, this

session in April was, know, kind of posing

: 1042

00:55:53,923 --> 00:55:58,113

the group, uh, around the implications

of this technology of whether you want

: 1043

00:55:58,113 --> 00:56:00,913

to be a driver or a passenger, right?

: 1044

00:56:00,923 --> 00:56:05,823

And, uh, I think it does come down to kind

of attitude about this positive attitude.

: 1045

00:56:05,823 --> 00:56:10,333

Like, I do want to automate away the

mundane things so I can work on, you

: 1046

00:56:10,333 --> 00:56:15,148

know, you perceive as the higher order

things or the more important things.

: 1047

00:56:15,158 --> 00:56:21,398

Um, so anyway, I think, I think that's,

uh, probably a good, a good analogy

: 1048

00:56:21,398 --> 00:56:24,458

or a good way to think about it and

why, you know, if you just want to go

: 1049

00:56:24,458 --> 00:56:28,028

to work every day and keep, you know,

spending eight hours filling those

: 1050

00:56:28,028 --> 00:56:32,938

eight hours with the way they used to

do it, or do you want to pose and ask

: 1051

00:56:32,968 --> 00:56:35,718

that question of Could I make that

: 1052

00:56:35,763 --> 00:56:36,063

Charles Portelli: Yeah.

: 1053

00:56:37,618 --> 00:56:39,348

Randall Stevens: Do I need

to be the one doing that?

: 1054

00:56:39,358 --> 00:56:41,558

Because I'd rather be spending

my time doing something else.

: 1055

00:56:41,738 --> 00:56:44,098

I think it's a good way to think about it.

: 1056

00:56:45,158 --> 00:56:45,368

Charles Portelli: Hmm.

: 1057

00:56:45,793 --> 00:56:49,163

Evan Troxel: I'll put a plug in here for

a recent episode I had on Troxel podcast

: 1058

00:56:49,173 --> 00:56:50,903

with Shane Berger from Woods Bagot.

: 1059

00:56:50,953 --> 00:56:56,393

When we addressed the topic specifically

about what happens to those newly

: 1060

00:56:56,393 --> 00:57:00,003

found hours that the, uh, the

automation or the efficiencies gained.

: 1061

00:57:00,383 --> 00:57:05,353

created and I won't spoil the outcome

of that conversation because I would

: 1062

00:57:05,353 --> 00:57:08,613

love it if people would listen to it and

then, and then provide feedback on it.

: 1063

00:57:08,623 --> 00:57:12,113

But I do think that that is something

firms need to think about is how

: 1064

00:57:12,113 --> 00:57:18,053

they are going to use those hours and

what, what is the best use of those?

: 1065

00:57:18,433 --> 00:57:22,433

And, and there's a whole gamut

of possibility there, right?

: 1066

00:57:22,463 --> 00:57:28,023

Uh, so architecture has, has often

been, uh, you know, in the column of.

: 1067

00:57:28,083 --> 00:57:31,183

Uh, we just do more stuff,

right, in those hours.

: 1068

00:57:31,243 --> 00:57:35,403

Like, we, we've, we've always

fallen into do more stuff, but

: 1069

00:57:35,403 --> 00:57:36,823

that's not the only option, right?

: 1070

00:57:36,863 --> 00:57:40,073

And lots of other, uh, verticals out

there have shown that, that there

: 1071

00:57:40,073 --> 00:57:43,353

are other things you can do with,

uh, with the efficiencies gained

: 1072

00:57:43,363 --> 00:57:44,803

that, that the tools have provided.

: 1073

00:57:44,803 --> 00:57:47,743

So, I'll put a link to that in the

show notes for this episode, so

: 1074

00:57:47,743 --> 00:57:49,183

people can, can listen to that.

: 1075

00:57:49,563 --> 00:57:50,443

Um, but,

: 1076

00:57:50,523 --> 00:57:51,343

but that, I

: 1077

00:57:51,343 --> 00:57:53,763

do think that's an interesting

topic to, to ponder.

: 1078

00:57:54,923 --> 00:57:57,553

Randall Stevens: just want to say,

you know, thanks again, Charlie, for,

: 1079

00:57:57,593 --> 00:58:01,653

uh, participating, you know, at the

live event in, uh, back in April.

: 1080

00:58:01,693 --> 00:58:05,933

And you're going to be coming to

Lexington here this fall and participating

: 1081

00:58:05,933 --> 00:58:09,243

with us, uh, for the three day

confluence event that we have here.

: 1082

00:58:09,243 --> 00:58:10,253

So looking forward to that.

: 1083

00:58:10,263 --> 00:58:15,728

It's always, uh, you know, uh, it's, uh,

I don't want to, uh, diminish that, you

: 1084

00:58:15,728 --> 00:58:19,468

know, it takes time to do these kinds

of things, but I think for the industry

: 1085

00:58:19,468 --> 00:58:22,488

sharing and being able to come on and talk

about these kinds of things in the way

: 1086

00:58:22,778 --> 00:58:24,908

that we do is a, uh, it's why I do it.

: 1087

00:58:24,948 --> 00:58:28,818

I feel like it's an important part of the

service to the industry to propel things.

: 1088

00:58:28,928 --> 00:58:34,078

I learn every time I have, uh, you know,

that I'm around, you know, uh, people

: 1089

00:58:34,078 --> 00:58:37,568

like you, Charlie, and the members of

your team, uh, about what you are doing.

: 1090

00:58:37,568 --> 00:58:39,428

So much appreciation for that.

: 1091

00:58:39,513 --> 00:58:40,273

Charles Portelli: The feeling is mutual.

: 1092

00:58:40,273 --> 00:58:42,713

I mean, having events like

the ones that you hosted and

: 1093

00:58:43,233 --> 00:58:44,563

the one coming up in the fall.

: 1094

00:58:45,343 --> 00:58:48,493

You know, we're all, we're all

in this together in a way, right?

: 1095

00:58:48,493 --> 00:58:51,773

Yeah, we're competing for, you

know, projects and so on, right?

: 1096

00:58:51,803 --> 00:58:52,333

I get it.

: 1097

00:58:52,923 --> 00:58:56,533

But you always hear a lot of

folks talking like, oh, I want

: 1098

00:58:56,533 --> 00:58:57,543

to move the industry forward.

: 1099

00:58:57,543 --> 00:58:58,993

I want to move the needle forward, right?

: 1100

00:58:59,463 --> 00:59:03,273

But if you keep it all internal, right?

: 1101

00:59:03,433 --> 00:59:07,313

You're, you're moving your firm

forward, but you can only move as

: 1102

00:59:07,953 --> 00:59:12,833

It's as fast or as much forward as,

you know, some of the people other

: 1103

00:59:12,843 --> 00:59:14,073

in the industry as well, right?

: 1104

00:59:14,073 --> 00:59:18,633

Like, to really collectively move,

everyone has to have, like, you know,

: 1105

00:59:18,633 --> 00:59:23,553

these large initiatives and kind of share

ideas and share thoughts and so on, right?

: 1106

00:59:23,883 --> 00:59:27,683

It's still up to designers and so on

to go out and win projects and design

: 1107

00:59:27,683 --> 00:59:32,693

beautiful buildings and all this other

stuff, but we have to collectively do it.

: 1108

00:59:33,163 --> 00:59:34,383

Do it a little differently, right?

: 1109

00:59:35,308 --> 00:59:35,638

Randall Stevens: Great.

: 1110

00:59:36,138 --> 00:59:39,138

Well, I think that's probably

a good, good way to end this.

: 1111

00:59:39,148 --> 00:59:42,288

Uh, again, thanks for coming

on and sharing your thoughts.

: 1112

00:59:42,288 --> 00:59:46,388

Looking forward to, uh, an update

in September when you present

: 1113

00:59:46,388 --> 00:59:47,948

at the next Confluence event.

: 1114

00:59:48,068 --> 00:59:54,078

And, uh, Evan, uh, as we were talking

about, or before, before we went

: 1115

00:59:54,098 --> 00:59:59,238

live, uh, probably be seeing several

people out in October in San Diego

: 1116

00:59:59,388 --> 01:00:01,288

area around the Autodesk University.

: 1117

01:00:01,288 --> 01:00:03,488

So, a bunch of us will be

out there for that as well.

: 1118

01:00:03,823 --> 01:00:05,133

Evan Troxel: A lot of

good stuff coming up.

: 1119

01:00:05,263 --> 01:00:05,543

Yeah.

: 1120

01:00:05,603 --> 01:00:06,283

Thanks Charlie.

: 1121

01:00:06,323 --> 01:00:08,043

Great to have this

conversation with you today.

: 1122

01:00:08,893 --> 01:00:09,333

Charles Portelli: Likewise.

: 1123

01:00:09,333 --> 01:00:09,653

Thanks, folks.

Share Episode

Shownotes

Transcripts

Follow

Links

Chapters

Video

More from YouTube