Name: Brainstorm Design 2023: The Power Of Sketching
Uploaded: 2023-12-06
Duration: 18 min 34 s
Description: Keith Griffiths, Chairman and Founder, Global Principal Designer, Aedas In conversation with: Clay Chandler, FORTUNE

Keith Griffiths, Chairman and Founder, Global Principal Designer, Aedas   In conversation with: Clay Chandler, FORTUNE

Brainstorm Design 2023: The Power Of Sketching

Conferences

We're going to kick off this next part of the program with something really special alive demonstration of the design process by one of the world's leading architects, Keith Griffiths is the chairman and founder and global principal designer of Aido, one of the world's largest architectural practices. It has a staff of 100 people and 12 offices worldwide. Keith leads a diverse portfolio of high rise residential, commercial retail and mixed use buildings across Asia Europe, the Americas and the Middle East. He's one of the creator minds behind the famous city skyline of Hong Kong. Uh despite the creative possibilities, A I Fords designers and they are always vast. Keith still believes that each project must begin with a hand drawn sketch. And so we're going to ask him a little bit about that and show us, ask him to show us his design process. Please welcome Keith Griffiths, chairman and founder, Global principal from Ida Keith. Yes. Yes, welcome, welcome. Thank you so much for making time to, to join us. Can we uh roll uh just a quick video clip behind us to give us a little bit of idea of uh some of the the projects that Keith's been involved with. Uh we got, we get that video going. There we go. If you turn the volume down a tad so that we can just kind of talk here, talk us through some of what we're we're seeing here. These are amazing images. First of all, what you're seeing here is how many people we've got about 1100 people, 12 studios throughout the world. And that balances our business out through periods of up and down. So we can maintain a steady size of business flow of business, which we've been for the last 20 years, maintaining that very steady size. We've been in the top 12 largest architectural companies for the last 20 odd years. This is the West Kowloon train terminus. You just saw there. If you're coming into Hong Kong or leaving Hong Kong, you've got to go through an Idas building, which I think is rather smart. But here's one of the airport buildings, the midfield concourse. Uh that's uh Doha airport, that's one we just opened and then all the sand stuff here. That's Venetian Paris, Londoner sands itself. They're all, they're all done. Uh with Ida's lead architects as the London. It's just, you know, like a whole world under itself there. It's pretty huge. That's just across the water that was Hench and IFC, a little box we did for Hong Kong land. And these are two towers and a dancer in Shenzhen. This is a New Ban uh Tower ST Regis uh Center for Sinovac and uh Sa Plaza. We do our headquarter buildings. These are up in Beijing. These are five headquarter buildings in Beijing and Quint, this is the world's most twisty tower. So it looks a bit dodgy actually. But that's 10 cents headquarters in Chongqing. That's China's headquarters. That's an enormous building and that's DAO Alibaba and uh that one is Shanghai Lovi Center. We just opened Lovi Center which is an experiential online offline mall. So that's looking to the future of malls. And on the right, we've got commercial bank which is exposing its innards, exposing all the meeting rooms. This is an amazing array of different shapes and forms here. Yeah, they really get to be crazy, isn't it? According to the clients? I mean, we we follow where the client wants to go. What are the needs of the building? Unilever Indonesia World Trade Center. That's Hong Kong land. This is the G 20 summit. I was so pleased to see all the world's dignitaries outside that building recently. And over in Athens, we started work on the Lenon already. That's moving very well. A lot of railway stations, that's the Dubai Light rail system and some further projects in the Middle East as well. And this is the one that I want folks to see this. Tell us about the, this is wonderful. This is Neon Trana Ski village top of a mountain in Saudi Arabia. And this is manufactured snow, of course, it's a hot climate there and this is enormous and very, very exciting. One of the projects that Saudis creating alongside the line and the cube and so on. Yeah, part of this whole big neon project in Saudi Arabia, not the sort of images we're accustomed to associating with Saudi Arabia, but uh just remarkable. So, you know, particularly for those of us who travel a lot in, in uh in China, in Asia, in, in the Middle East. I, I think we've all seen your building that you've, you've worked on in one kind of way or another, perhaps without realizing it. So fascinating to be able to chat with you a little bit about how your, how your process works, how and how it's how it's changing. Um We've spoken a couple of times about how you use uh uh A I but tell us when you started using A I and kind of how that, how that shifted the way that you work. OK. Well, we've only been using A I for a matter of months. It's, it's a phenomenon that's burst upon the architectural world in the last few months. And I'm a drawing architect I sketch. So we see as architects, we three, we see three dimensional buildings in our heads and we use sketching as instant two dimensional views as we move around the building or the building looks like this. What does that part look like I look down and I sketch it out. Now, that sketch is not normally enough for a client to make a decision. He needs something more. He needs, he needs a rendering, he needs color, he needs light and shade and that's where the torture part comes in. Because when you start working with a client with a developer, he wants to see all the options, he wants them all. And that process of going from a sketch that in your head to a rendering good enough for a client. We have to put it into computer. We have to create a computer model. We have to render the computer model and then show it to the client. Now, that process takes a week for each option. So even if you're running, it simultaneously could be 23 weeks before you can get to the client with a bunch of options. And then he says we don't like any of them. Torture, go away and do it again. Uh It's totally torture. Now. A I has made us the torturers. We're now of the sadist because we can create and you'll see in a minute, we can create an image every 30 seconds. No problem sketch. Yeah, that's a good one. I will. OK. Next one. No, don't like that one, next one. And then A I takes over and it creates reasonably accurate to the renders of what we've created as a sketch. Good enough to show a client in the first instance. So should we show this example here to give people a kind of visual idea of what this looks like? This is a sketch that took you how long to, to this sketch took about 20 minutes. So this is I'm showing this one because what we're gonna do later will be very quick 32nd sketches. But this is a more detailed sketch. So this, yeah, this is quite a detailed sketch. Remember you're talking two D to two da I is two D sketch is two D. It represents 3d. You can see a 3D space, but it is a representation of the two D and now you'll see what happens with it there in an atrium with some planted trees. OK? Here we go with the first shot, put it in and what do we get? We get some structural trees. That's interesting. We get some strange spider webs going on. That's interesting. So we pick the best one of those and we run it again with a few more prompts. So this gives you four different choices and you say, OK, you know, I like that one and then let's tinker with that one and fix it a little bit. So a few more than four choices and we keep moving it forward, keep moving forward. Now that good enough, right? We can get that in front of the client. That's OK. Yeah, I mean, the fact that you, that you've got a hand rail on top of a tree. It doesn't matter. It's good enough, you know. So you can give him the feel of what that space is gonna be like. And he might say, yeah, I don't want a shopping center for us, you know, forget it. This is great. Yeah. So um I want to step over to the table but just give us an idea quickly about uh manpower the old way of doing it versus this way of doing it to, to before you started using how many people would have you have to put on that kind of rendering process. OK. So traditionally, um you know, let's say uh you got two designers, say myself and another designer working through a problem, we've done the planning, whatever we're trying to get some visualizations, probably take us a couple of days to get some sketch vis visualizations that will go through simultaneously to the, the the computer guys who will be modeling it. They will take at least it's a model like this at least three or four days to get that model at least. And then it's got to go to renderers and they're gonna take four or five days at least. So just for that, you're looking at 10 days. All right, let's step over and see what the process can now look like. Uh Just step here and this is fun. I'm gonna try to, I hope, I hope we entertain you. Stay out of Keith. Keith left handed here. So just kind of perch on his left hand. This is Gianna Giana is the guy who does everything. So, so, so I, I just sketch, we're gonna bring a camera so you can kind of look over Keith's shoulder while sketch it. And if I draw the thing which is making it completely impossible for me, I'll get a slap wrist and I'm not wearing my glasses. So it's gonna be very. Ok. Ok. And um we're gonna start off with some high rise. Yeah. Ok. Smooth high rise. See my instruction. He's a client. Ok. So we'll start with some smooth high rise. Can we see this on the? There we go. All right. Great. Nice. OK. Good enough. OK. It's gonna start feeding that and we're gonna do another one. What's, what's the next one? Ok. We can do some. I just want to point out here how small this scanner is. It's OK. Crystal fee. So we have, we have all the power of A I platform that's harnessed on a laptop here with a scanner that's like it's like the size of a can of tennis balls. Good enough. OK. Can we see this next one? Yeah. Yeah. So there's two. Ok. The program is running right now. Ok. What are we gonna try now? Let's see, just like that. Wow. So we're gonna go a bit more crazy. Ok. So I want to mention that uh Keith is using a platform called uh X Cool, which is specifically designed for architects and urban planners. And we're going to have the founder of the company that created that platform uh on our program a little later today. So look forward to that. There you go. Get that. Huh. Not bad. Huh. So this is what the technology can do and you can compare it just OK. Wow, back and forth. So you see sketch versus rendering maybe. Uh OK. I a curry. Yeah. OK. So we wanna be horizontal, horizontal. So he's the boss. OK. We're gonna do a curvy swerve. Uh Let me see. Let me see. OK. There you go. Oh Nice. Uh Good enough. Uh Yeah. Yeah. Sure. Try that. And then we'll do a boxy afterwards. Yeah. OK. We can do a box because those would be exciting, aren't they? So we're gonna do a um post pandemic box. Yeah, we Oh Nice. That looks pretty complicated. Let's see. Let's just put a few more bits to make it more difficult for the machine. Wow. Coming to a skyline near you. Um So tell me um can this process work without really that much attention to the sketch? Could you give it visual or, or verbal prompts and generate what the kind of structure and image and feeling that you have in your head? Yeah, you could. But all things are working together here. So I've got some verbal prompts already. We've, we've given it, we've given it a uh imagery prompt as well. Um So we're guiding it down a certain path. Uh In fact, Jenna is just changing the prompt now for this box, because we're gonna use a prompt that's gonna make it much more realistic. And you'll see that it's quite interesting what happens there, curtain wall will start to appear. Um So those prompts work with the sketch. Do you need everything together? But essentially the reason for the sketch is it gets it out of my head the best way of transmitting the image. That's the fastest way. That's why it's powerful because yes, you can use the words, but words are a very long winded way of describing a three dimensional building. Oh It is a long tower with square sides and a bump in the middle. No, so you need this and this is exactly what's my head? I mean, there's no thought process going on. I'm completely dumb when I sketch, I'm thinking, I'll put a tower over here. Well, that's something that I'll do something here. I'll put a box, I put a box on top and I'm not actually thinking I'm just drawing, drawing, drawing and that's going straight into that machine. So that's the, the magic of A I is. And I can't do auto C ad, I can't do anything on a computer and I haven't been able to do it for the whole of my career 40 years. But now. Hey, I've got, I've got a computer working for me for the first time. But, but what's remarkable about this is that you don't need to be writing code. You don't need to be a programmer to be able to get the results, the vision and, and translate it into something that you can actually take to a client. Exactly. It's very straightforward, very easy. And it puts us in the driving seat for the first time as a technology. Wow. Amazing. Can we see some of these other renderings? Can we show the, the images that are on the, on the computer? There we go, we go. What can you show us on? There? It is. Look at that, look at that from that sketch and that's the one and there's the box one complete with clouds. It's amazing. Oh, yeah. Look at that. Oh, yeah, that's why I need. So, do you see any flaws in these drawings or they look pretty good? Oh, there's loads of flaws. But it's, it's good enough, you know, I mean, it's good enough to show the intent. OK. I'm going to do a curvy building. That's waves on waves. Yeah, that's good. So how do you know that these are structurally gonna hold up? You don't? OK. So then what has to happen? Well, now, you know, so we go through that process of what could the outside of the building be? Now, we could have come at it planned the building from the inside. And now we're thinking about the outside skin. So that's one way that we've got basically a form and a shape, we just modeling it. So the sketch would be done over. So I'd be basically taking, let's say, a form and shape like that and then taking over that and saying, OK, fine, I'm gonna change something. So I've got the total sketch and I'll just change something in that sketch. Right. So that's one way, the other way is that we don't know exactly the planning, but we've got roughly the mass of the building and we're really playing with what the urban designer, the client, the government would like to see on this site. Uh It's amazing. I, I wish we had more time for this. This is so fascinating, but Keith, you've given us a great uh sort of quick snapshot of just how revolutionary uh this technology can be and how, how creative really uh it can be uh when it's uh when it's in the right hands. Keith Griffiths. Thanks so much for joining us.