Can a robot do a site inspectors job today or no? - Short answer is yes. Question is whether you want it to be as accurate as you want it to be. So everything becomes like a glass house. - Should construction be worried about the next five years? - They shouldn't panic yet, but they definitely have to pay attention because the issue of getting labor, the cost pressures are gonna be there. There is going to be robots that will be significantly faster and significantly able to outperform and in fact are super human in its capability to engage the physical world. So if we are not on top of this, the impact on humans is that they will be felt this and franchise. - Chin Lim built a tech company in Singapore, sold it to Autodesk, spent five years on the inside, watching how the world's most used construction software actually works. Then the Singapore government recruited him to lead their smart nation program, deploying robots, AI and sensor networks across an entire country. Then he walked away from all of it, start again from scratch. His company, Deconstruct Robotics, now builds robots as scan construction sites, build digital twins and catch errors before your team even gets to site the next morning. His clients include some of the biggest companies on the planet. But here's the thing, this conversation goes way beyond robots. Chin has a view on AI, on power, on what happens when a handful of companies control the infrastructure, the whole world runs on. He taught me something about Autodesk that genuinely change how I think about this industry. And he's got an analogy about elevators and autonomous vehicles that might change how you think about the future of building designs and society. This, my friends, is one of those episodes. I'm Owen Durery, and this is Bricks and Bites. Enjoy. (whooshing) - Okay, Chin. Thank you so much for joining me all the way from Singapore and a quick stopover on your world tour of bringing Deconstruct to the world, eh? So we're gonna dive into some quickfire questions, then we're gonna find out about you. So quickfire number one, is the world more or less stable than it was five years ago? - Definitely less stable, but it's an opportunity to lay the groundwork for new forms of resilience. - Nice. Can a robot do a site inspectors job today, yes or no? Or you wanna have the opportunity to look at all the different fuzzy areas and negotiate a better way to do stuff. - Hmm, interesting. Has AI lived up to the heart? Barry Minus is at 15th of April, 2026. So just to put the timestamp there. - Well, AI has certainly met the hype in terms of capability and breakthroughs, but I think he hasn't quite caught up in a real world for adoption across all industries. - Mm-hmm, nice. So you're based in Singapore, from Singapore. It's just done of what is, I don't know, it's Singapore, the smartest country in the world. - I'll say this way diplomatically, it's strong in executing planning and the system level thinking is there, but it's certainly not the best in the world, right? Because smartness really depends on value in innovation, the scale, the governance, different countries have different priorities. - Very cool. Should the construction be, bosses be worried about the next five years? So moving early from a structural perspective is gonna be a problem. We have to find out for themselves if they are too far behind to catch up. - Mm-hmm. Where are your kids, and my kids living a better world than you did? - Well, technologically, certainly be more advanced, yes. And you will be more complex and the goal in many ways, because the opportunities that come with access to innovation and now the AI, it means there's a lot that the new generation have to navigate to balance their perspective in life. - Yeah, great. Okay, last one then. What does, okay, so yeah, let me give you some context here. You coin this term when we first met a few weeks back, quote, and the phrase was tech for good. Can you explain what that means? - Sure, I didn't quite coin it, but it was from organization, the government I used to work with called Draftec. And tech for good essentially means building system that amplifies our human potential, ensuring that this potential is accessible to as many people as possible. - Autodesk Singapore government startup founder, you actually were a startup founder, sold your startup to Autodesk work there for a while. Can you give us a brief overview of your story? - Yeah, start that my company, building solution that brings clients or type application into the cloud. And when I was building this, I got an opportunity to work with Autodesk because they were transitioning from selling just their software from disk to cloud enable application. So we worked with them for a little while, they made an offer after we worked for them and say, well, don't you, we'd be interested to join Autodesk family. So I did that. I was with them for five years. And after five years, the Singapore government came knocking on my door and say, look, we got a project for you, if you're interested, would you be interested to be a senior director to build a smart nation platform solution? So I thought it was a great opportunity to look from another perspective that is the government. So I joined that. I was there on the contract and we went from three years and ended up doing a half. So I got ants in my pants and wanted to get back to my startup roots and decided that AI robotics was the way to go. - Why did the Autodesk buy your company? - I think there was a couple of reasons. One, we demonstrated that we could be able to integrate into the existing Autodesk platform. That's one. And second, we had the talent that they needed in order to think beyond their current business of distributing the software through disks. So CD disk. And we were already plugging into the cloud. And so we knew and talked the language of cloud first and they were transitioning from a company that was doing something that was very physical to something that was more digital. So that transformation that they were going through gave us an opportunity to have that conversation with them and our IP that we have on board and the talents that we have, we were able to then make a deal for us to get into the Autodesk family. - One poorly managed incident. That's all it takes to demolish the entire profit on your project. And right now, the only thing standing between you and that disaster is your superintendent showing up at 3 a.m. buried in paperwork before the days even started. That's why breadcrumb exists. Not to add more technology to the job site, but to give the people running it tools that actually work the way they do. Safety plans signed off. Orientations completed on any phone in any language, perm its process without the back and forth. No paper, no chasing signatures, or synced directly into ProCore. With OSHA penalties reaching into the tens of thousands, that kind of oversight isn't just a nice to have, it's everything. Over 1 million workers across 50,000 projects globally have breadcrumb in their pocket. Book a demo at breadcrumb.co. That's br-e-a-d-c-r-u-m-b.co. Links in the show notes. Major construction projects are complex, and construction managers know that every day things don't go as planned. Unfortunately, most don't know what and why until it's too late. The best construction leaders know that the secret to delivering on schedule is for everyone to be planning together, committing to targets, and tracking their performance every day. By combining look-aheads, weekly work plans, visual staging diagrams, and daily activity briefings into one multiplayer platform, AFX is where the world's leading projects are planned and coordinated. Join the largest projects and contractors across Europe, the UK, Asia, Australia, and New Zealand and discover why at AFX.co. In construction, precision matters. Every budget line, every schedule, every document. But too often, project data is scattered across tools and departments creating risky silos. That's where ARKDes comes in. ARKDesk is the only one construction management platform designed to connect your entire project operation from opportunity through to tender, all the way to hand over. With real-time dashboards, cost-control, scheduling, procurement, and document management in one place, ARKDesk gives you complete visibility and control. Trusted by contractors, subcontractors, and asset owners worldwide, ARKDesk has powered landmark projects from shopping centers to Europe's tallest buildings. If you're ready to reduce risk, protect your margins, and deliver projects with confidence, discover ARKDesk today. And visit ARKDesk.com, that's arcdisk.com, and see you.
how leading firms are building smarter with Art Desk. What did you learn from the process of building this company and then selling it? Yeah, first of all, you learn a lot about yourself, right? And you want to figure out what you prioritize is important. But one of the things that when you're running a company is certainly unless it was a philanthropic company, we have to figure out that we have top line and we've got bills to pay and we have talents that we've got to feed. So one of those things is making the decision how much can we push the company to the level that we are all happy with. That's one thing we learn. The second thing is we have to figure out how can we contribute beyond the horizon that we see. The companies that are big and are already established gives us a lot of opportunity to look beyond horizon and certainly gives us a lot of opportunity to learn. The bit that actually is exciting is to see our technology skilled up. So as a company that started from Singapore and we have a smaller market for us to be able to participate in a global company was very, very exciting. Very cool. Okay, so tell us about what was it like to work for Singapore government? Singapore has an ambition that was made clear that they wanted to become a smart nation and not just a smart city. So the idea of a nation includes not just infrastructure but the people. So I was brought in specifically to look at what kind of infrastructure can we build in Singapore that allows us to use this technology for increasing and improving the quality of life with the citizens. So this was an exciting mission for us to take not only because it is at the national level but is also looking at discovering what technology we can apply that achieve the goal that our cabinet in Singapore has set itself on achieving. What was like, can you give an example of some of the initiatives that you wrote out? Sure. So one of the major considerations is what I talk about is can we have conversations with a city or infrastructure or a town. So I was actively working about team to build out the open digital platform which essentially is a facility management tool that allows for messaging to the different IoT devices to the endpoints of a building of cluster of buildings and together into a small town. And then we are able to take this information and analyze that information and visualize the information for not only policy making but also for situational awareness. So the system allows us to understand how much power we use. If even the door or room that is occupied, how are elevators being used? What was the crowd looking like? And it gives us information about the equipment and whether it is in good shape. Because we built this, we are allowing that conversation with robots as well. So robots doesn't necessarily have the hands or arms to open doors. So we were able to build the self-opening door with connecting to a robot demand to open that door or how robots needs to get around horizontally by going through turnouts as well as vertically through going through different floors. Of course the robots that we can integrate with back then were robots that can deliver things or in some cases also clean the floor. So that's how we started. And that conversation allows us to understand what kind of security system we need to put in because we don't want the security system and protection that is needed for physical infrastructure, especially now that you have robots talking to it. So this kind of information, we are able to collect the data and we are able to understand from that data how buildings are used, how rooms are used, and it gives us an idea how to better manage the assets that we have and maintain. The other projects that we've done was around smart land posts. So one of the things that we were discovering is how does smart land posts improve the quality of life for our citizens? Whether does it bring more safety, whether does it bring more convenience, were they able for us to be able to provide safety because the land posts normally provide light but also allow for collecting of information like if there's a traffic or if it is a note where we can use it to disperse wireless communication and so on and so forth. So we learned quite a bit about it. So one of the lessons I learned while working on a smart city for the land posts was that it may not be the most efficient solution because Singapore is very urban already. Whatever that we want to hang on the land posts, there are enough buildings nearby that we could have made it easier to install and learn to use the surrounding infrastructure doesn't need to roads to do the same job. The reason being there is specific use cases for land posts, first of all providing lights and sometimes land posts are not powered when it's not needed. So for example, between 7 a.m. and 7 p.m. there is not enough power for the land posts to put on other stuff, other sensors and other equipment and also for safety measures should be an excellent as well. So all these were taken into consideration with the proper authorities and managers this and conclusion that we have is Singapore is urban enough, dense enough where the other buildings can provide that infrastructure together data much more effectively and cheaper. I think the most critical thing is we are in a government to serve the people. So everything that we do has to have a legal basis for it and the outcome has to be crystal clear. Not only are we measured against how we spend public money but is also how can we use that money to provide services that the citizens, the citizens want and the citizens feel that when they have it, it improves their life. So this is one of the key factors and Singapore government is very diligent about this and they are very careful about this. So sometimes we think about bureaucracy, there's a reason behind it because we have to cover not only how we deploy a solution but we have also to look at the legal implication. You have to think about there is politics involved and so anything that we do we have to be extremely careful and we must be able to defend our position. So as a result we need to document what we think and do and then have those in a sense scrubbed by different perspectives so that when the outcomes come out it doesn't remove any part of the society that cannot benefit for it. So we try and work very hard to think about to be inclusive at the end of the day. The other lessons that we learn from it is that when we are looking at deploying any solution and using the massive resources that are available, we want to ensure that it's aligned with the overall direction of where the cabinet has promised the citizen. So in many regards this helps us align what are the priorities that we focus on because we have so many resources and we have so many opportunities to do a lot of things we could get carried away. So having very focused outcome is extremely important and because of our research that we do and we do very deep research in what we want to deploy, we will scour the world to figure out what makes sense, we learn from what other people do well, we also learn from the mistakes they make and we ensure that we don't make the mistake and then we enhance on all the good points so that the solution that
that we have in mind, there's a lot of sense to everybody. - Very cool. - Tell us about construction. Tell us about technology within the construction industry. - Wow. That's construction technology, I think, as we know, is a very physical one, right? Because you don't make a digital burger, you go to construction, you really gotta get the thing built, whatever that thing happens to be, is a house, is a building, is a road, and so on. So it's a very physical activity. So how does this, in a sense, align with what's happening in AI and so forth? So I will say this thing, while the construction industry is very physical, there's a lot you can do with the data. So I believe the future of construction needs to be also in data collection. Because if you wanna use AI in the world of construction, without data, you can do enough. You wouldn't be able to link between what is digital and what is physical. So one of the amazing things that I think the people in the construction will, may not realize it is, you know, people live in it. So it has an amazing influence in the happiness of people. So in many ways, while construction is a physical thing, there is that other touch that it will inspire people, you'll be able to give us memories, you'll be able to give us a safe place to live in. So construction on its own, has a, in my opinion, is important in that context. Now, how do we figure out in the construction space that it can continue to deliver in an effective way and cause effective at that? So, and also the problem construction, I believe is facing is also, there may not be enough people to be going into the construction industry. So the conversions of trying to manage costs as raw materials go up, and then to be able to find the talents, to be able to achieve the outcome that construction projects need to do is not an easy one. So there is going to be a high need for automation, for AI, and in some cases, robotics. I say some cases because not all robots can do the work of a human, and it will take some many years before robots can even achieve some of the things that we as humans do naturally. One distinctive feature that construction companies, I think, always face is that there is always the design that people have in mind, and then there's a process to build it, right? There is that gap that is you sometimes discover too late. So the question now is, can construction be able to understand what was designed against what is being built at a consistent milestone or real time so that if mistakes are made, they can be captured. If we can do this, I think we can reduce the cost of construction from errors, and then have that savings to be distributed, not only to the ones that finance the project, but also to the talents that is working on it, then outcome would be excellent if I say if we can reduce more errors. - Yeah, interesting. Tell us about that. When did you decide that and see the opportunity and think like, yeah, this is where I'm now going to dedicate my life to, I'm guessing you had a few options to choose from as well, so I would be good to hear that decision. - Yeah, so what stood out right away is actually we're not enough talents around. So companies struggle from finding the talents to keep their business going. So that was one of the things that struck me and given also some demographic information around fertility index, at least where Singapore is, we may go extinct at some point. So the problem I wanted to solve is, how can I do more with less? And this is something that a business owner wanted to figure out in the very competitive world. So that's what drove me. And so I wanted to figure out a way how I can contribute to solving this problem. So I thought automation AI robotics was the way to go. And in that regard, I wanted to build intelligent machines that could observe like a human being and to be able to think like an expert and act in the context where it is put to achieve a mission. So that got me into thinking and gathering the talents for me to build robots with AI capabilities to achieve those goals that people who need more people to get job done, but can't find them and get the robots to actually step in to help the business move along. So your talk, what industries are you talking? The obviously construction being one of them. Yes. So from an industry perspective, construction is an area we play well in. Other areas include facility management and in surveillance and security. And what kind of capabilities, like what is the service of the robotics that you are building? Like what is the service that they're providing to these industries? - So from a construction perspective, we're looking at surveying, right? So one of the solutions that we do provide is how does your construction progress look like against what you design? So we are able to take rabbit drawings, beam information and we are able to compare that with the site, the construction site, where we have scan using both LIDA and terrestrial laser scaners. And this information that we have can be around one centimeter to perhaps two, three centimeters, depending on which kind of solution we use. And then we can superimpose that with the beam. And from there we can get an idea what how much progress was done or whether that construction that was done was aligned with the design and how well it aligns with it. So there is some degree of opportunity to address if something has gone wrong, we can pick it up. So we provide a solution that does that. And another aspect of it is our robots are able to carry their camera system that allows us to enforce safety conditions while construction is going on. Like identifying if somebody is wearing a helmet or vest or if somebody has fallen or they are stepping into areas that are not supposed to be and we can send that LID right away. So the enforcement of safety measures in construction, our robots can roam the construction site and provide that information. While they are roaming, they can also provide security like if there's somebody who has intruded and not supposed to be at the site. We can pick this up and we can lock the video evidence to show what's going on. And because we have the scanning capability, we can do pretty much quite a lot of things. One of the major thing is reconstructed and provide a photo realistic, photo clear situation of the construction site. And then we can also provide something that is unique. What you are able to do for Google, like you are searching for work, because we have scanned the construction site, we are able to tell you if you are looking for a project where the object is located, because we have scanned it. And pinpoint where it needs to be. We can also provide information like there was a fire extinguisher that was there, but it's not there today. So where is it? So these are some of the things we can do with the data that we collect. Other aspects of it is we can also measure progress. So if someone is reporting that this is done, and we can verify it through our scan. Mm-hmm. Yeah, nice. Interesting. Is it primarily science inspections for now? Right now is for site inspection. We also do facility management through inspection. So we are able to identify if there were sprawling on the wall, pieces of things or bricks are fallen off or they're cracks. OK. You can identify things like if there was a walkway that was too narrow.
and then it is a violation of safety because the emergency stuff cannot go through or there were unattended items and not supposed to be there, we can pick it up. We are also able to look at meaters as needed and identify if they are in a correct meaters setting. So these are the different facility management for maintenance that we use to robot to do. Example would be if there is a large facilities and you are certainly trying to figure out there's enough, not enough people to do the inspection, we can get the robot to do it and it's able to use, we can use different forms whether it is on wheels or whether it's on lakes or caughtropet to be able to navigate the environment, take the lift of its own as well, go to all the different spots and you can do it a few times a day if needed. I think the future will be around, maybe I unpack it this way, so when buildings are designed for accessibility is where robots will work very well, although not every accessibility route is suitable or you cannot it may not cover every part of the building but it's a good start. I think moving forward in the future construction will be looking at robot-friendly buildings, so that would certainly change when we think about how it's being built or design and built and certainly how it would be maintained so that other robots that do surveying or inspection, you may also use robots that are for cleaning and for delivery autonomously throughout the building. Interesting, okay yeah I'd like to go into your thesis of the robotics and say just quickly on some false that you've experienced of how you are implementing these robotics onto construction sites, so what have you learnt so far about industries, whether the industry is receptive to the idea skeptical, like yeah just speak to that kind of. Yeah sure our relationship with construction companies is a hit and miss, so what I mean by a hit is because they are urgently needing people to execute specific tasks, whether it's for inspection or whether it's for survey, they need somebody to do that and they need to be able to collect data on a consistent and frequent level, so our robots and our sensors that we put on our robots to navigate construction site is very useful for them, so that's a hit for them because that lower that manpower requirement or they can reschedule people to do other things that are urgent. Where is a mist is actually what I spoke about earlier having a glass house perspective of things, sometimes if you have too accurate data collected or you have too much information, construction company may worry because how do they justify to the financier or the property owner if he doesn't meet the specification exactly, so the idea is there may be some tolerance for error but if I can pick up an error regardless of how much you can tolerate it, you have a sense of something is not right, so that's what I mean by a glass house perspective on our construction site, it's just too accurate and so as a result of seeing everything what it is not sometimes give a psychological impact to those financiers or property owners and you know that hey this is not done right, while it is perfectly within the error of tolerance, so it becomes unforgiving and so this may be a contentious area and therefore you know companies that are in a construction if you say no this is too much for me, I don't want to use it, I still want to use a person and then negotiate if something goes wrong. That's interesting. And sometimes it takes more than just a bunch of technology to navigate this field, you know, there has to be some sense of wisdom in how that is done because if you use it as a governance and penalty then it's really going to be tough to getting the project done, so one way to approach this is not do it where you space out the inspection or you are checking the site too far apart, so the advantage of having this is actually to do it every day almost when the construction team goes home, let the robot do its thing, so the next day you get a report and that's what we do, we give a report where the items are missed and you can correct it before it goes too far and that's I think is the key right to see it as a part solution provider into making sure that if there's an error made is caught at the right time. Interesting, I'm curious, when you're thinking about this and as there relates to the labor crisis and the reasons for you doing this, why did you pick site inspection and not go for something like a bricklay or painting or some of the other use cases that we see throughout the industry? Yeah, so when it comes to painting and it comes to bricklaying, these are requiring robots of a considerable size, so in the context of trying to use this robot say in painting, you really have to make sure that they paint it right, while there are many cases where a flat wall may be possible, given that the design going forward, the seldom you find a structure where robots can easily navigate and move around or even get set up, so that difficulty is something that we didn't think we want to solve today and then when it comes to bricklaying and so on, in the context of Singapore is always prefab, so it's done in a factory-like environment and then it is then sent out to the site to be essentially assembled. So the technology for this to do bricklaying in Singapore context may not be feasible to be a long run business, a long term business, perhaps in bigger countries where you have a lot more space and then bricklaying becomes useful. Not saying that you cannot be used in Singapore just basically means the context where a construction happens is different because of prefab in Singapore. Yeah, very good. Okay, cool, so people now when they're thinking about robotics and particularly inspection and reality capture, you actually see some of the big reality capture platforms offering like a robotic inspection service as part of what they provide. So why do people pick the construct or what's the over like other robotics companies that exist within the ecosystem? Sure, so we are picked because our solution works and the key about our scanning solution, it is very accurate. So that's one of the major reasons. The second reason is because with the data that we collect reality capture, we are able to do a lot of analytics on it and we can use AI for the analytics. As I've mentioned, we can do things like reproducing it photographically accurate to a millimeter. We are able to search objects within it. We call few shots search. We are able to measure it from a digital perspective that is accurate without having to go to the site. We can take progressive scan of the area and then show that scan built up. We are also able to essentially detect what change. So one of the interesting thing about change I will segue a little bit is not only can we determine a wall, we can also determine if the tree was growing and how much did the tree grow. So volumetrically, our analytics tool can do that and it gives us information that is useful when the user wants to do planning. The other aspect of our solution that is interesting is that you can buy the solution or you can buy a service. So sometimes in construction, you may not want to own the equipment, you just want to have a service and you want to have the data and use that data to do a report and that could be a lot cheaper. In many cases, if they want to own the equipment means that they are a construction company that can use many times over. If they were just doing a project and far in between, they can engage our services to provide the solution that they are looking for. Very cool. So your CTOS from PIXR, right? Yes, it was from PIXR. He will go
under Steve Jobs. Yeah, that's cool. That's a cool story. He worked on Wally and Toy Story 3. Oh yeah, he worked on like 10 plus movies from Pixar. That's pretty cool story. So question number one is how do you attract someone like this to your mission? Well, I think the longest short of it is that I need to have a vision that people feel that they can be part of. So one of the amazing thing working in my CTO is that no easy talented and is very smart and he has world-class experience is that he figure out that our mission and our inspiration to do stuff using AI and robotics for society, jobs. And we think we can build something exciting and contribute to the world. I think that's one of the major things that kept us going at it. I think the second one is when we see the smile on the customer's face, right? We work on very hard so we work on very hard problems and we have to find solutions to it. And these problems that we solve are not easy to be able to get robots who understand this environment and then to perform the task autonomously is not easy task, especially in unstructured environment. So with him, we were able to pull like-minded people and because of his talent, other talents want to work with him and under him to learn from him. So this allows us to build a team that can tackle these difficult problems that the society and the industry face. That's very cool. How was the transition to going from building and animating virtual entertainment kind of, we're asking robots for arguments like real ones. Yeah, so if you think about it, it's actually all mathematics. So if you are looking at building an animation, most people think that you actually were animating a movement from one cell to the next cell, but if you're thinking about doing that, you probably take forever to get an animation movie done. In fact, what is actually true is that the animators just do a spot of animation or the character and then they show the end of the character. So for example, the character is supposed to walk from A to B. It's not easy for you to animate one cell at a time, especially at 24 to 30 frames per second, then you have to do a lot of animation. So what actually happens is a program is written that transform the start of the character walking and then automatically bring it as if it was an animation to the end. So they actually write the code to do this and there's physics involved and then mathematics involved in order to show it moving naturally or where a hair is wet or hair movement and rest of it. Those are not animated, but in fact program. So when you jump into from discipline from graphics and development and building research solution for this, the algorithm for this, and then we want to actually do reality capture in the real world, the principle do not change. So with this capability, my CTO and the team is able to calculate based on picture taken, lighter information taken, including the IMUs. We are able to recreate this and then manipulate this into providing photo realistic perspective of it at high accuracy. And we also use 3D Gaussian splatting in order to achieve the natural perspective of it and then to be able to fly through. This ability to fly through after we scan is very key. Robots only need to see point clouds, but human needs to see photo realistic perspective, not from just a picture point of view, which is a flat perspective, but also about a 3 dimensional perspective so that you can move in and out and see it from different perspective. Those capabilities are both required when they have grounded themselves with knowledge and mathematics and physics. How do you make sure that companies don't get, or how to push your advice to people who are implementing or thinking of buying robotics on a construction site? How do you make sure that they don't get burned? They actually actually works, like the rollout is successful and they are fully advocating for robotics. Yeah, construction company needs to be very, very clear about the outcomes they want to have. I want to state it out front, not all robots can do everything that you see on YouTube. In fact, they are very, very far away from achieving any useful function in the real world. So we've got to make sure that you don't get sucked into that height. So making sure that you have very clear outcomes. Second, I think what needs to happen is you need to have the vendor, the service provider, show you on site. It's one thing to show a bunch of power points and a bunch of videos. It is very important to have the equipment seen on site in especially your site, right? So sometimes your site could be a flat land, but sometimes your site is actually a tunnel. Sometimes your site is actually a bridge. So the sites are all different and not all robots can do all sites. So make sure that the service provider is able to demonstrate it clearly on site. The third area is the robot has to, you have to work at setting your expectation to have very clear outcomes when the robots are being done and that you have an opportunity to have the service provider hold your hand because most construction company do not have a robot's department. So as a result, they really don't have anybody to turn to if something went south. So it's very important as part of their service is to have the robot service provider hold your hand and guide you through. So that's the third aspect. The fourth aspect is once you get involved and you are fairly knowledgeable about how robots contribute, you really have to think about your concept of operations, how robot plays into the construction process flow and when they come in and when they do not come in. So it will impact on the business re-engineering perspective of how construction work processes are done. And then I think the fifth area is figure out to be a partner to service providers, robot service providers because the feedback is very important to get the robot to perform at the fidelity you expect for your next run. Robots can be tuned, algorithms can be changed. The robot company are likely to have deep expert knowledge in the industry they are in, especially in the construction field when they have so many tasks to be done and there are so many materials to look at. So it's very critical to be able to have a feedback mechanism. Some things can actually be tweaked within the time frame of your construction site. Some things is beyond it and so creating that relationship means that you are able to feedback and for them to improve. So the next time you use them, you have a better product. Very helpful. Thank you. Okay, Chien, zoom in out a bit then. So when we pull up before you say something about something's happening with AI, power and money and it might affect everyone, whether they like it or not. Tell us what you're seeing and what you're thinking about in that realm. Wow, that's a big one. So I think there is a new world order happening when it comes to AI power and in many cases, how that would impact us. So I would like to touch on a couple of big ideas. The first big idea is you need data for AI to work very well. So that's a big idea. Second idea is actually there is going to be a new currency in play because of AI and when I mean by the currency I play, taking the idea of AI providers and AI users, there's also the whole idea of AI infrastructure. So I believe going forward, there's going to be an AI token idea that's going to come about. So what I mean by this is large data centers will be built specifically for AI processing and these large data centers you're talking about billions and multi billions in terms of cost and they will also be using a lot of water and electricity. So because of this, not many players can handle this cost and the ability to build that skill. So it's likely you'll be the big
type but also governments. So when government get into the play or building large data centers that has these capabilities, what's going to be impacted is this infrastructure is gonna sit there and do nothing. It will then create AI service providers. And with AI service providers, it will be building application agents that rest on the AI infrastructure to enable it to be useful. And there will be many that will build this. And the VCs are actually funding this to happen. So agents, AI agents are being built by different companies, people are figuring new AI algorithms with so very complex problems. And then you have AI users. So users, AI users will be paying for the AI service providers who in turn needs the AI data center in order to get the application working. There is going to be a new token economy. So a new currency is gonna come up because AI users, whether they pay in Fiat or they're gonna pay in tokens or crypto, whatever the case be, it is going to be requiring payment to this service provider in order to get the service that they want. So with this cycle, it basically means that it's going to be a divide. Those who don't understand AI and cannot use AI or have no clue how to start is going to be left out. The AI users and AI service providers with the AI infrastructure is going to accelerate at very high speed when it comes to innovation and progress. What used to take a while for someone to build our PowerPoint can be done very quickly today. And with AI allows it to give ideas that the person is not thought of in higher quality. So imagine if you do not have access to this, you do not have access to a computer or the internet and you do not have access to this ecosystem. Or worst case, you don't have money to get access to it. The consequence of that is you will be left out as a person versus somebody who is augmented and powered by AI. That divide is going to just accelerate. So that I think is going to be a deep divide that's going to happen. But AI actually just exists in the digital world. There is going to be a need to figure out how AI with all these advances is going to connect to the physical world. Therefore, physical AI is an expression of the digital AI. And as the consequence of that, there is going to be robots that will be significantly faster and significantly able to outperform and in fact are super human in its capability to engage the physical world. So if we are not on top of this, whether it is in the digital AI or the physical AI, the impact on humans is that they will be felt disenfranchised. And this is something I think the governments will eventually have to struggle to figure out. Now, I'm not going to jump into how AI is going to be used in warfare. There is completely another subject. But I'm looking at how digital AI and physical AI can be used in terms of tech for good. I think that's where it is. The question has to now happen. What are the gut rails that we put when we are thinking about using digital AI and physical AI in our world, whether it's construction or other industries? Now, one of the major consideration is is there a kill switch for this? I think we are on the kill switch. There is no easy way to shut this down. Nor do I really want it shut down, because the benefits hopefully will outweigh the disadvantages that may come. So while we don't have the kills switched anymore, I think it is beholden on anyone who is in this field that's using AI and physical AI to make sure the gut rails are published and make it very clear to the end user what those dangers might be. So that anyone using it is conscious of the dangers that you might pose while enjoying the benefits that it provides. Wow. So what do you want back? [LAUGHS] Yeah. So let me just correct me, because I'm just interpreting what you said. And I'm trying to translate into my understanding here. So when you mention AI service providers, obviously, you have big-- you mentioned like the AI data center being built, big tech and then government, and then working together to deliver these data centers. Then on top of that, you have the AI service providers. So with these providers that you mentioned, would they be the big names that we are already familiar with, like Amphropic, Google, and-- Oh, yes. Those would be the ones, those that provide the applications. Like Microsoft, right? So those who provide those applications have significant influence and impact in our daily lives. So I mean, like OpenClaw, if you use that, is able to jump into any Clore, StarClaw, right? Is able to jump into your phone and your desktop and help figure out your life. So what you have actually published is your life just when public. No matter how you think about it. So the danger is, is there any cybersecurity around this? The short answer is AI service provider providing that service as fast as they can for two purpose. They want to get as many users as possible. And then they want to get the investors to help them pull free and scale. So that rails for security is now left as a after-dog. The consequence of that is, if you think about it, if your data get hacked or somebody gets hold of it, or there's identity theft, or they screw up your life and do something, it's hard to recover. Yeah. It's very hard to delete a digital trail. Yeah. And I think also some of the more recent models, particularly Mythos by Claude, who calls some major security concerns within this back big organization, sorry, Mythos by an anthropic. And they held off releasing it, right? Because the impact could have been so profound, which is extremely concerning. Yes. I think about it. So then, not the race for humanity, I'm probably exaggerating slightly here, but the divide of in humanity becomes those who can afford to use this artificial intelligence provided by some of these service providers, versus those who can no longer afford to do it, because one, they can't afford from a financial perspective to a technical perspective. They don't understand it. And free, I guess, just general access because they live somewhere different or any other reason. Yeah. Exactly correct. That's why you read in the news. The data centers that are being built, it's talking about hundreds of billions of dollars. This kind of money is not easily available. How many GDPs or how many countries can group together when you're spending a few hundred billion dollars? And so only the ones who are excessive controls it, they will a lot of power. Yeah, fine. Because the spending and capital required is just so extraordinary. Yeah, no one else can match that. Thereby, their divide just goes on to this ridiculous scale where no one's ever catching up at that point. And so then, yeah, they can then effectively it becomes this ultra monopoly between these companies where they can-- well, I know that there's obviously competition within it, but if they all work together to increase prices and restrict-- yeah, it becomes quite a utopian kind of-- dystopian full start going from my head to thinking about this. I mean, simply put, it's just a grab for AI to become the next power level to control. And then the danger to it is if it's connected to the physical AI, right? So anything physical connected to this intelligence is something to look out for. That's what. OK, so what's the space? Don't panic too much. Last question, because this one generally really for people who are particularly listening to this about the elevator. So remember the elevator energy where it kind of unlocked vertical movement through buildings and the world. Tell us about that and how we then-- how the stand translates to horizontal movement. Yeah, so when you think about elevators that was before the elevators were around or invented, buildings were maybe five to six.
floors and the consequence of living on the top floor is anything that you need to bring up, you gotta go through the stairs and hopefully the stairwell is big enough to bring up your so far and stuff, right? And then the only thing about it is nobody is interested to climb up and bring stairs up. So the consequence of that is actually the first floor is where the shop floor is. Everybody crowds around it, you make business around it, nobody wants to go, I'm much higher. So the value of the shop floor or the building space is actually on the ground floor. So that was back in the 18th century. But when elevator was built, what changed profoundly, changes, people now can get to the top floor much easier. The consequence of that is it means that the top floor is now not having those problems, you can use the elevator to get up and move things up. And because you're higher, it means you can get higher, then buildings get much taller. The higher you get the taller it is, then you get a wonderful view. That's why today buildings with great view is the one right at the top and you can see above everybody else and they have no issue getting to it. And then the noise that's on the ground floor, we've now all the two motifs and transportation that's around, it basically means you are closer to heaven than you are on the ground. But actually, if you think about the elevator, it is actually an autonomous vehicle. Because when you get into the elevator, you press the button and you get from the first floor to the 50th floor without thinking much about it. And while there, you're talking to your friends around you or other people around you. But maybe these people don't talk so much. But anyways, it's convenient. So it is an autonomous vehicle because you press, I want to go there, it brings you there. So the idea of now looking at autonomous vehicles, autonomous and advantage that comes with living up there is like is, is as if you took that vertical autonomous vehicle and then you move it down horizontally. So now you imagine you have a world with autonomous vehicles that allows you to move anywhere easily and where we more and zoops companies like that. And of course, the robot taxi that is coming out. You can think of it as your time is saved. Because when I get from A to B, you can do something else. And it changes the mechanics of getting to places, but it also changes the economic value. Just like high floors are important. It now means that you can get to any part of the city conveniently. And therefore, there's no one part of the city that's going to be more important than the other because of ease of transportation. Today, we centralize into central business districts, we centralize city in a certain way. So with autonomous vehicles coming out, we will change the design of cities. There is no one spot that you will find. There will be too heavily traffic because you can get to any part of the city just as conveniently. So the same idea as a elevator changing the building, the horizontal autonomous vehicle will change cities. Interesting. And then also, I suppose, somehow time, could we tie this into your point earlier about designing buildings for robotics as well and making I guess buildings where elevators are, but making buildings usable by robotics and not just humans. That's right. So to extend the idea is that when you can move around horizontally, you really change the idea of car ownership or mobility ownership. Because you just want to get to one place or another place safely while doing it, you can do multiple tasks and doing other things. So in the context of how robots play that role in buildings that are accessible, you almost can imagine a world where the robot does most with your stuff, right. Taking out the garbage because it's able to move through the building because the elevator is now accessible for robots. It basically means the robot is not only able to move horizontally but also vertically. So if you give that freedom and you build the infrastructure for a city space that allows for this, when you have a COVID, it's no hardship, right? Because you now have robots doing stuff that otherwise you can't easily do. Now most people survive in COVID because they had internet. Imagine you didn't have internet. Wow, you couldn't work from home, the work will have come to a stop. So the internet pretty much allowed people to work from home. And then you have this whole work from home thing going on. So would it be wonderful if there are people who are having accessibility issues to have that physical reach that otherwise they couldn't and they will have to be dependent on caregivers and so on. So these aspects of where robot can play a role in tech for good is they can even be a companion of sorts, right. Because even a more and more isolated lonely world. So if robots can play a role and open the person to a world they are unable to be part of and we can bring the world closer to them, that's important. Because we don't live only in a digital world, we need to have a physical touch. So while this is not a substitute for human touch, but at least there is something they can feel, there is an interaction they can have in the city they live in. So much to think about. Okay, very, very last question I'm sure. It's kind of open open, but if there was one thing you could tell any construction leader, senior profession or within the industry about the next 10 years, what would it be? Well, in a new 10 years, I think construction will be pretty much the same. Where you will differentiate and where construction companies and service providers, a company service providers is data. So you probably need to think about how do you collect data from as many touch points as possible to be ready to use AI to help you use that data, that digital information for decision planning, situation or awareness. Because that data is going to impact when robots come into play and you can then connect that knowledge into a physical robot, the physical AI to execute the plan. If you do not think about this today, what is likely to happen is those companies that have data construction companies that have data ready, plug into AI agents, plug into physical AI, they will outperform the construction company. And of course, and of course, to add one more ingredient to that would be the access to the AI service providers. Absolutely. So you have this token economy of the AI. So that's why the big boys are betting everything on it. Because the impact is not just construction is every facet of our lives. Yeah, very cool. Wow. Chin, thank you so much. I really feel like we could have had like three four hours, but time, time is restricting. Cool. All right. Chin, can you tell people where I can find out a little bit more about your company deconstruct and also about you? Sure. You can visit our website, deconstruct.ai, or you can email me,
[email protected]. We have our products and our services publish. And if you have any specific projects or ideas that you want to work with us, we are happy to be your AI robotics department. And we'll figure out to build a solution that not only works for you, but also for the industry. Love it. Thank you so much.
PodcastTranscript AI