WEBVTT

00:00:00.020 --> 00:00:03.850
<v Tim Fu>I see UAE as a key player here. They really are geopolitically,

00:00:03.850 --> 00:00:07.770
<v Tim Fu>economically, and the AI progress as well. Honestly, this is one of the reasons

00:00:07.770 --> 00:00:12.040
<v Tim Fu>I'm here. And one of the reasons why a lot of our clients, if not most of our

00:00:12.040 --> 00:00:13.870
<v Tim Fu>clients, are actually here in UAE.

00:00:19.070 --> 00:00:24.620
Tim Fu, an absolute pleasure having you here in the Ajman International Urban Planning Conference.

00:00:25.080 --> 00:00:26.080
<v Tim Fu>Thank you. Pleasure to be here.

00:00:26.080 --> 00:00:28.880
You told me that you started your studio about two years ago,

00:00:28.880 --> 00:00:34.100
two and a half years ago, and you're exploring AI mostly within your designs.

00:00:34.100 --> 00:00:37.920
Lately, you had Lake Bled close to Slovenia?

00:00:39.220 --> 00:00:42.340
I don't know whether this is the project that you want to pinpoint today,

00:00:42.720 --> 00:00:48.090
but you did mention that this was an AI-created project from start to finish.

00:00:48.090 --> 00:00:53.580
<v Tim Fu>Yeah, maybe just a bit of a clarification is that it's not driven completely

00:00:53.580 --> 00:00:59.960
<v Tim Fu>through AI or pure AI project. It's more of a.. the first time that we've been utilizing

00:00:59.960 --> 00:01:03.010
<v Tim Fu>AI at every stage, every decision.

00:01:03.830 --> 00:01:10.240
<v Tim Fu>Meaning that, although, it is not driven by AI in these decisions, we incorporate

00:01:10.240 --> 00:01:16.520
<v Tim Fu>it into our team so it asks, for example as consultants, we are communicating

00:01:16.520 --> 00:01:20.020
<v Tim Fu>between clients large language models such as ChatGPT comes in

00:01:20.020 --> 00:01:23.000
<v Tim Fu>and gives us summaries, informs us what's a

00:01:23.540 --> 00:01:28.570
<v Tim Fu>good decision to potentially make and at the end of the day we have a panel

00:01:28.570 --> 00:01:34.180
<v Tim Fu>of humans that drives the design. So AI comes as an additional support to that.

00:01:34.640 --> 00:01:39.230
<v Tim Fu>So I think that's something that we should definitely clarify because I think

00:01:39.230 --> 00:01:44.410
<v Tim Fu>it's important for us to maintain in the driver's seat of any project,

00:01:44.940 --> 00:01:49.380
<v Tim Fu>the values of a project such as Lake Bled is the values of human,

00:01:49.380 --> 00:01:51.240
<v Tim Fu>of heritage, of culture.

00:01:51.240 --> 00:01:55.670
<v Tim Fu>And these things are fundamentally human. For us, it's about really...

00:01:56.410 --> 00:02:03.330
<v Tim Fu>finding out what is the best decision to make at every stage and adding AI helps

00:02:03.330 --> 00:02:06.410
<v Tim Fu>us inform ourselves in the decision-making process.

00:02:06.670 --> 00:02:10.800
What do you mean by every stage, especially if this was during the ideation

00:02:10.800 --> 00:02:12.390
creation or the creation of the...

00:02:12.710 --> 00:02:17.360
<v Tim Fu>Sure, sure. So for example, we are designing window details.

00:02:18.050 --> 00:02:23.340
<v Tim Fu>And all of a sudden, we are looking at question of materiality,

00:02:23.340 --> 00:02:25.460
<v Tim Fu>the type of metal, what is your ship.

00:02:25.740 --> 00:02:33.110
<v Tim Fu>So we upload the image to ChatGPT and we ask, what is the most applicable supplier

00:02:33.110 --> 00:02:36.430
<v Tim Fu>for this specific type of window type? What would you suggest?

00:02:36.770 --> 00:02:40.340
<v Tim Fu>And what is the carbon footprint of, for example, shipping them from there to here?

00:02:40.800 --> 00:02:45.370
<v Tim Fu>So I'm just giving you a very accurate, specific example of one of the things that we do.

00:02:45.810 --> 00:02:52.190
<v Tim Fu>So you can always try to ask AI very specific questions and see how useful it is.

00:02:52.400 --> 00:02:57.430
So it's really about utilizing as much data that you have in terms of decision

00:02:57.430 --> 00:03:02.350
making when it comes to creating these concepts and designs and really making

00:03:02.350 --> 00:03:06.820
it more realistic with the optimum solutions that AI gives you.

00:03:07.120 --> 00:03:13.460
<v Tim Fu>Yes, yes. It's all about data. And us humans, we also want to make informed

00:03:13.460 --> 00:03:20.180
<v Tim Fu>decisions. Our concept is using AI to gather more data pre-digested

00:03:20.180 --> 00:03:23.040
<v Tim Fu>for us in some ways. That's what AI does.

00:03:23.640 --> 00:03:29.240
<v Tim Fu>It's a human, large-scale language model, and there's a lot of implicit knowledge

00:03:29.240 --> 00:03:31.260
<v Tim Fu>understanding already in there.

00:03:31.680 --> 00:03:34.830
<v Tim Fu>And it's creating links of connections that you may not see,

00:03:34.960 --> 00:03:40.430
<v Tim Fu>but the AI does because it's like a super powerful network. It's operating in

00:03:40.430 --> 00:03:44.190
<v Tim Fu>the same way as our brain is by taking different memories and creating synapses.

00:03:44.520 --> 00:03:48.520
<v Tim Fu>But the AI is so much more powerful because it can comprehend so much larger

00:03:48.520 --> 00:03:51.350
<v Tim Fu>amount of data and then bring that to us.

00:03:51.610 --> 00:03:55.000
<v Tim Fu>So why not use that to our advantage? More information is power.

00:03:55.210 --> 00:03:58.480
How can you make sure that AI does not hallucinate

00:03:59.240 --> 00:04:00.500
during the design process?

00:04:00.500 --> 00:04:06.860
<v Tim Fu>I can't. That's the answer. I cannot be sure that this is right or not. I have to confirm.

00:04:07.330 --> 00:04:12.790
<v Tim Fu>But that being said, same with humans. I cannot ensure that human error is not

00:04:12.970 --> 00:04:17.630
<v Tim Fu>inherent in our memory, in our biases, in our way of thinking.

00:04:17.630 --> 00:04:23.370
<v Tim Fu>It's the same critical thinking that you will require for AI as you do dealing with humans.

00:04:23.880 --> 00:04:30.420
<v Tim Fu>And I think the beauty of that is AI is gradually decreasing because it's improving so fast.

00:04:30.590 --> 00:04:32.400
<v Tim Fu>And humans, we're slow, right?

00:04:32.670 --> 00:04:40.210
<v Tim Fu>It takes us millennias to adapt and evolve. Whereas for AI, it's just going off

00:04:40.210 --> 00:04:43.270
<v Tim Fu>the rails. We're going to see it develop even further in the next decade.

00:04:43.270 --> 00:04:45.450
<v Tim Fu>And we have no idea what it's going to become by then.

00:04:45.820 --> 00:04:49.920
Let's talk about AI as Agentic AI. How is that implemented?

00:04:50.230 --> 00:04:55.890
And can you also, Tim, take us through the design stages that you have where AI really comes in?

00:04:55.890 --> 00:05:00.280
<v Tim Fu>I think we're still at an early infancy of this technology of Agentic AI.

00:05:00.280 --> 00:05:05.610
<v Tim Fu>It's not as mature as the intelligence that is showcased by ChatGPT.

00:05:06.360 --> 00:05:09.450
<v Tim Fu>So I would say I would still wait for a bit.

00:05:10.110 --> 00:05:15.810
<v Tim Fu>We've been developing websites as well, HTML code using Manus,

00:05:15.810 --> 00:05:17.480
<v Tim Fu>which is Agentic as well.

00:05:18.010 --> 00:05:20.990
<v Tim Fu>And I've compared it with what we've done with Gemini. And for example,

00:05:21.390 --> 00:05:25.960
<v Tim Fu>this language model has done a way better job at coding HTML and producing the results you want.

00:05:26.820 --> 00:05:30.310
<v Tim Fu>So it's a matter of the data set, it's a matter of the training.

00:05:31.210 --> 00:05:34.970
<v Tim Fu>Agentic AI has not impressed me to the extent as large language model has,

00:05:35.530 --> 00:05:38.270
<v Tim Fu>but we foresee that it will be.

00:05:38.400 --> 00:05:40.630
<v Tim Fu>It will be very, very impressive.

00:05:41.160 --> 00:05:47.510
<v Tim Fu>I foresee for sure that we will ask it to do secretary task easily,

00:05:47.930 --> 00:05:53.980
<v Tim Fu>ask it to arrange meetings, talk to clients even, talk to people for us.

00:05:54.100 --> 00:05:58.020
<v Tim Fu>The inefficiencies in humans is how we communicate with each other and how slow

00:05:58.020 --> 00:06:01.440
<v Tim Fu>we are and how disorganized we are, to be honest.

00:06:01.440 --> 00:06:04.390
So to use it more for managerial work.

00:06:04.730 --> 00:06:11.540
<v Tim Fu>Yeah, yeah, yeah. I think honestly, having AI as a psychic and starting a company,

00:06:12.730 --> 00:06:15.390
<v Tim Fu>is a Godsend. You know, alhamdulillah, as you say.

00:06:16.210 --> 00:06:19.540
<v Tim Fu>It's really, really good. Really good. If you want to start a company today,

00:06:20.010 --> 00:06:22.230
<v Tim Fu>it's so much easier than five years ago.

00:06:22.430 --> 00:06:27.910
How do you think the future would look like with superintelligence AI in terms of architecture?

00:06:28.410 --> 00:06:34.560
<v Tim Fu>This is the fun question. So, you know, there's a famous saying that it's hard

00:06:34.560 --> 00:06:38.130
<v Tim Fu>to speculate in three years and it's impossible to speculate in five years.

00:06:38.380 --> 00:06:45.160
<v Tim Fu>We are living a technological era that its progress is at an exponential rate.

00:06:45.160 --> 00:06:49.370
<v Tim Fu>For architecture, I think we will see a

00:06:49.370 --> 00:06:57.230
<v Tim Fu>point where we can deploy prototypers to Mars and to the moon and let them

00:06:57.610 --> 00:07:00.870
<v Tim Fu>harvest resource and self-develop and create cities,

00:07:01.790 --> 00:07:06.930
<v Tim Fu>create first factories, factory pump out, you know, different robots and they

00:07:07.290 --> 00:07:10.770
<v Tim Fu>autonomously keep creating and building until there's a moon base or a Mars base.

00:07:10.770 --> 00:07:14.920
<v Tim Fu>Like that's completely within the realm of possibility. It's just a matter of when.

00:07:14.920 --> 00:07:18.430
It's quite interesting. Are you more optimistic when it comes to AI?

00:07:20.430 --> 00:07:22.210
<v Tim Fu>Outwardly to the public, I try

00:07:22.210 --> 00:07:30.730
<v Tim Fu>to be but there's things that keep me up at night. I absolutely fear

00:07:32.170 --> 00:07:37.630
<v Tim Fu>where AI could go wrong, because if we're talking about this on a more fundamental,

00:07:38.150 --> 00:07:41.780
<v Tim Fu>human level, it's like, what are we doing here?

00:07:41.780 --> 00:07:46.240
<v Tim Fu>You know, what is our goal? And in creating something that we eventually can't

00:07:46.240 --> 00:07:47.590
<v Tim Fu>understand how it thinks.

00:07:47.730 --> 00:07:52.300
<v Tim Fu>Yeah, that's what's driving me with our research always, is in the realm of art

00:07:52.300 --> 00:07:55.260
<v Tim Fu>and architecture, we want to maintain authorship.

00:07:55.390 --> 00:07:59.400
<v Tim Fu>And so that to me is a critical aspect of what we do.

00:07:59.950 --> 00:08:05.130
<v Tim Fu>And that's what drove us this entire process. I think the reason why we pump

00:08:05.130 --> 00:08:10.000
<v Tim Fu>out the things with AI that speaks to people is because you can see the human

00:08:10.000 --> 00:08:12.080
<v Tim Fu>elements that we embed in all of our work.

00:08:12.410 --> 00:08:14.310
<v Tim Fu>And I think that speaks for itself.

00:08:14.710 --> 00:08:19.340
<v Tim Fu>Now, the moment we lose that, it's going to be what they call slop.

00:08:19.660 --> 00:08:21.540
<v Tim Fu>You know, AI slop is happening everywhere.

00:08:21.540 --> 00:08:25.920
A lot of the AI experts talk about that stage, but then they also talk about

00:08:25.920 --> 00:08:31.430
the later stage that is more optimistic. And, you know, so hopefully,

00:08:31.650 --> 00:08:32.690
hopefully we reach to that.

00:08:33.860 --> 00:08:39.960
My last question is, how do you feel like the MENA region, especially the UAE,

00:08:41.140 --> 00:08:43.900
is ready for adaptation of AI?

00:08:44.310 --> 00:08:51.520
<v Tim Fu>I see a lot of investment to data centers, to resources and collaborations with,

00:08:51.520 --> 00:08:53.880
<v Tim Fu>for example, U.S. companies and with China.

00:08:54.470 --> 00:08:58.580
<v Tim Fu>And I see UAE as a key player here. They really are geopolitically,

00:08:58.580 --> 00:09:02.760
<v Tim Fu>economically, and the AI progress as well. I think...

00:09:03.520 --> 00:09:07.620
<v Tim Fu>Honestly, this is one of the reasons I'm here and one of the reasons why a lot

00:09:07.620 --> 00:09:10.800
<v Tim Fu>of our clients, if not most of our clients, are actually here in UAE,

00:09:11.290 --> 00:09:13.070
<v Tim Fu>Abu Dhabi clients, Dubai clients.

00:09:13.730 --> 00:09:19.010
<v Tim Fu>I see, first of all, they're very open to technology and open to this advancement.

00:09:19.320 --> 00:09:23.800
<v Tim Fu>And they are very strategically intelligent in understanding that this is something

00:09:23.800 --> 00:09:27.310
<v Tim Fu>that we have to invest in intellectually and financially.

00:09:27.810 --> 00:09:33.860
<v Tim Fu>And so it's through this process that I'm engaged with this part of the world quite a lot.

00:09:34.340 --> 00:09:40.900
<v Tim Fu>And they have to make a lot of important decisions as a key player in this technology

00:09:40.900 --> 00:09:45.820
<v Tim Fu>as well, how to be careful, to be protective of how we develop.

00:09:46.510 --> 00:09:51.780
<v Tim Fu>But I would say that the general outlook I have is quite positive of UAE.

00:09:52.200 --> 00:09:56.270
<v Tim Fu>I've seen a lot of the ways urbanism has developed here in UAE to be more responsible.

00:09:56.810 --> 00:10:01.040
<v Tim Fu>And I see the society here experiencing it.

00:10:01.210 --> 00:10:06.480
<v Tim Fu>I can compare that it's really looking to the future and very,

00:10:06.900 --> 00:10:08.880
<v Tim Fu>I would say, multicultural.

00:10:08.880 --> 00:10:17.290
<v Tim Fu>It embraces difference of opinions and it welcomes new information and knowledge and collaborations.

00:10:17.560 --> 00:10:24.360
<v Tim Fu>So I think this fosters a positive outlook on where I see this country can take AI.

00:10:24.560 --> 00:10:29.720
<v Tim Fu>So hopefully we're going to be more integral part of this process, Inshallah.

00:10:29.720 --> 00:10:33.360
And Tim, coming into Ajman, it's always interesting, you know,

00:10:33.360 --> 00:10:36.880
from an architecture perspective, when you go into a city and you look into

00:10:36.880 --> 00:10:39.270
the city, you start observing the city in.

00:10:39.850 --> 00:10:44.900
And then you start going like, yes, this is a potential opportunity for me.

00:10:45.360 --> 00:10:50.080
I want to do one, two, three. What are the one, two, threes looking at Ajman

00:10:50.080 --> 00:10:53.240
that you feel like you want to do and, you know, work on?

00:10:53.560 --> 00:10:58.590
<v Tim Fu>This is my first time here, right? My first impression is that it's actually,

00:10:58.990 --> 00:11:01.090
<v Tim Fu>I would say on a more human scale, more...

00:11:03.150 --> 00:11:10.800
<v Tim Fu>I feel it's more approachable. And also, I understand the ecological, the climate here,

00:11:11.520 --> 00:11:18.490
<v Tim Fu>It doesn't really promote that. However, maybe there is ways of creating walkability,

00:11:18.490 --> 00:11:22.880
<v Tim Fu>shaded areas, passive solar strategies, and underground tunnels.

00:11:22.880 --> 00:11:26.090
<v Tim Fu>I don't know. There's so many things that we can do to...

00:11:26.810 --> 00:11:32.920
<v Tim Fu>re-approach this issue. I think, you know, there is health issues related to non-walkable

00:11:32.920 --> 00:11:37.480
<v Tim Fu>cities and we will see that in cities that are very reliant on cars,

00:11:38.160 --> 00:11:43.000
<v Tim Fu>but again bringing it back to Ajman I think it's kind

00:11:43.000 --> 00:11:47.450
<v Tim Fu>of on the more edge of the center of where it's happening so there's more

00:11:47.450 --> 00:11:53.610
<v Tim Fu>space to invent and develop and improve, so I see it, especially these cities that are,

00:11:54.290 --> 00:11:59.820
<v Tim Fu>on the more edge of like, Dubai, we can see that they have opportunities to develop differently.

00:12:00.130 --> 00:12:05.480
<v Tim Fu>And hopefully to see it develop in a way that is more approachable as a walkable

00:12:05.480 --> 00:12:10.360
<v Tim Fu>city because I'm not going to compare with Netherlands because it's a very different climate.

00:12:10.680 --> 00:12:12.390
<v Tim Fu>But there's a lot of things we can learn from it.

00:12:12.390 --> 00:12:15.870
Did you know that Ajman is the 15-minute city?

00:12:16.020 --> 00:12:17.760
<v Tim Fu>Ah, yes. I heard about that.

00:12:18.240 --> 00:12:21.640
And especially in the older areas, people are always walking.

00:12:21.640 --> 00:12:27.700
<v Tim Fu>I'm one of the weird people that when I come to UAE, I try to walk to see it. To observe it. Yeah.

00:12:28.060 --> 00:12:33.700
<v Tim Fu>And then I feel, obviously, I'm the only one walking. Yeah. And it's a very strange experience.

00:12:34.280 --> 00:12:39.020
I think it was John Gale when he said, you're looking at city from different perspectives.

00:12:39.700 --> 00:12:43.860
And he started, you know, creating that if you're in the street,

00:12:43.860 --> 00:12:47.940
if you're in a car versus you're on the bike versus you're walking,

00:12:47.940 --> 00:12:52.550
it's really how you observe it and you take it in. Yeah. So it's a different experience.

00:12:52.550 --> 00:12:58.750
<v Tim Fu>I mean, I would even put myself in the eyes of a bird or animals that roam around

00:12:58.750 --> 00:13:04.110
<v Tim Fu>the streets just to understand the entire ecology of how the perception of a city is.

00:13:04.110 --> 00:13:08.960
Tim Fu, thank you so much. Thank you for your time. And we look forward to welcoming you back again.

00:13:09.220 --> 00:13:09.910
<v Tim Fu>Thank you. It's a pleasure.