AI Redesigns a 244,000 Sq Ft Hospital in Four Days, Are Architects at Risk? & Snaptrude’s AI Agent Launch

I think people who will be disrupted, will be the people who don't adopt AI on the consulting side. Because the technology is now ubiquitous, it's available to everyone. - How are they gonna stay competitive to the other people in the market? It just don't understand how any architectural firm not deploying an AI strategy has, I don't know, any odds. - Unless they are the top firm average architecture firm if they don't deploy AI are gonna be in trouble. Because you cannot differentiate yourself with another one down the street. (upbeat music) - Prick's box invite. - All right, what's up, Gents? Look, no introduction today except for Dustin. I think you have something to say to the world about your shenanigans with a certain competitor of yours. So I think everyone wants to hear it. Tell us. - Yeah, I think it's actually one of the most viewed posts they've ever had. A company posted some fake reviews about us. And I found it so interesting that the person who runs a customer as he used to be the CEO said it was never meant to mislead the market. And I find that hysterical because you have to have intent to create fake testimonials and post them on your website. The intent is to mislead the market. That is the intent. So to say there was never intent to mislead anyone is in itself a lie. That's a factual statement. Like there is two people that were just completely ghost customers and one not real people, not real people. One customer I personally emailed and he wrote me an email saying, I never said that I've never used your software. So the only way those can go on your website is if someone maliciously put them under your watch and I'll tell you how I would have responded as the CEO of a company like that. Let's say someone in my marketing department went rogue and did that. My first response would be that person is fired. But that not being the response shows me that this was a knowingly thought that they could get away with it. That is played and simple. And I saw some people defend it saying they have high integrity. You can't have high integrity where you actively create fake testimonials and post them. So these testimonials were talking about how they had switched from us. So customers had switched from you. So they said they effectively churned and no longer like using edify. And moved to one of your competitors. And this competitor is the same company that put these reviews on their website. Which is kind of an incumbent this competitor to some extent, right? They've been out for a long time. No one switched from us. We haven't had a single customer switch from us to them. It's just not, yeah, it's not a thing. We have customers that have left them for us. But this is not a thing. So this is a, well, it's just shocking, honestly. And then what do you think about this? I'd like to think that we are a very big industry but turns out that we are very connected and have very many friends. And like a tiger only has one skin, right? So be very, very careful about what you say. It will be a friend, a friend of a friend. Just don't, don't, don't do it. And if you mess up, if it's at any moment misunderstood, like go 200 miles and go out of your way to say, I'm really sorry, this is not how I mentioned go on record. You know, I think that this is a very, very tightly knit community. I have no idea of what really happened there, but in my views, it does never end well. No, I mean, you know, there's other stuff on their website about the functionality of our software that I didn't even beat them up about. Well, you know, like misrepresenting our capabilities or capacities, which look, that's kind of subjective, right? So I'm not even beating up about that. It's you have the, the action of creating fake testimonials and misrepresenting it, that intakes intent. Yes, and there is someone behind that that did that. Okay, let's flip this slightly. Do you know see that as a huge compliment to what you're doing? Oh, I love it. I'm not going to have any good. I love it. You would only do that if you, if you, it shows, it shows like what they actually fear, right? And who they actually fear. So yeah, of course. Yeah. P, P, Austin, is the best I, ultimately, and you can be on both sides of it, right? You know, it's, it's, it's very, there's, there's no such thing as bad news in the P, Austin. Oh, yeah, I got so many text messages. It was great. I got a lot from their customers. Like, I can't believe them. Yeah, backfires and it was the con. And there's a certain rule, right? Don't, don't talk about your competitor. I mean, it's, it's, that, that's not what you do. That's just not, maybe that's a British thing now. And I don't know, I'm not clearly not British, but I've been here like 40 years. So we're going to use the best parts of being a Brit. Yeah, and you don't talk down about your competitor. Why would you do that? Yeah, I get like some industries. So sometimes it's quite fun to watch a band to play out. But like construction, largely very professional. And it doesn't have the same effect in this kind of industry, right? You have to be very careful. Because like you say, Alan, everyone is very well connected. Everyone knows each other. Eventually, like, someone's going to find out that you probably, probably shouldn't. Yeah, it used to be a little better. And I think what's, what's been going on lately is in order for you to defend yourself, you have to call out some of the misrepresentations that are being made all over the place. I mean, that customer in particular, you have to download their desktop app to use it. And they market themselves as a cloud company. Like, you guys, this is, this is misrepresenting. It's misleading, the intent is to mislead people on the capabilities of your organization. But like, there's falsehoods being propagated all over the place. And so I would like not to talk to about our customers and only our value propositions. But when you hear these claims repetitively and customers recycling the information they hear, you have to try to at least stand up for your position and like give a countering argument to why maybe what someone holds to be true might not be a factual statement. And then I, you know, I bring up this because like, if a company's willing to post face to estimate fake testimonials, what else are they willing to do? Mm-hmm. Integrity comes into question, I would say, about the brand in its entirety, which is a little concerning. Yeah. Well, it was fun. I loved some of the post people who were like eating popcorn and I was like, oh, to be fair to them. They did respond directly on your post. However, yeah. They didn't respond to a specific question, I asked them. Yeah, of course, of course. So I guess we can try and get them on to bricks and bites at some point and hear from them. Please, we could do another AEC fight night. So anyway, so Dustin, that's where you've been. I mean, we have been, we haven't seen you for a few weeks, but you've been relishing in your LinkedIn fame following your post about the testimonials. And to be honest, I would say there's a complete backfire because now you have, oh, how about this? We use the one that like instigated this whole thing. So yeah, how long will you end up winning? I wish I did. Now, I am the one at my organization that found this. So we have agents that scrape, scrape all competitive marketing material and get a report every Monday morning. Sunday and I are Monday morning of what every single customer or every single potential competitor is saying in the market. And so that's how I stay tuned to what's going on. And that's how we found it. So it went up Friday, found it, emailed the customer, got an email response. He said he's contacting them and then I actually, well, I won't go into that. I tried to reach out to them before posting that and I got no response. (laughs) Oh dear. So that's that. And it looks like we are now in the first week of November. So events season kind of winding down. I mean, the two big ones are over. But I don't know if you guys saw the two big ones. You can name them if you'd like. I actually plan to do that at their events on the same week next year, which is in the description. sane. And I was someone, I think Mike, Mike Powell's, uh, uh, uh, pole out. It was just like, which one are you going to attend? A, B, or none. And the majority came back none. Which was, uh, quite, quite entertaining to us, uh, spectate. So, uh, you know, uh, the, the democratic and Republican conventions, you know, they coordinate so that they never run on the same time. Yeah. Like, guys, who is going to have a conversation? You can't overlap, Jesus. Yeah. I mean, if you, I wonder that this is clearly at the liberal decision, right? To, to do this on the same, same week. So I wonder, I wonder why? Like, surely there's a, there's a zero sum kind of outcome. Um, we have to choose one over the other. I don't know, maybe it wasn't liberal. But even if it was them, when you're trying to shift their out. You saw Sid's post that Autodesk already had it booked. And then I don't know when pro core announced theirs or when they locked it in. So, hmm. So we got that. It's becoming really, you know, I think it's used to be quite easy to know where everybody was going to be. And, um, I think we go to these events more to see the people. And to have a very efficient week, done to, um, necessarily, you know, always all due respect to the speakers and, and the people are put a huge amount of energy in, in, in recording lectures and sessions, etc. But, um, yeah. Now there are so many of them. It's going to be really difficult. I mean, in, in October, November, you could be at, at, at, at, at an event every week. Literally. I mean, we were just talking about that. And I'll slap more in a hotel in the last two months. I have my own bed. Yeah. Yeah. The poor people have to attend this stuff and decide where they want to go to next. So what, what happened in the design world this week, I think there was a few kind of, um, uh, fundraising rounds looks like MEP is a trending area, as well as, uh, some of the bigger rounds, Mac9, dream, silver, uh, etc. It continues to be very, very busy, right? Um, and, and some, I would look back, you know, over more than a week, but, um, it, it's attracting a lot of funding. And some of it, like dreams gets announced, but it has already been in place for 12 months, in fact, whereas, some more are, you know, immediate, small, round big rounds. But yes, I think that the, the, the, the very big trends that we're seeing is obviously physical AI. And then MEP, like everybody loves MEP. So there's really a lot, a lot of announcement and a lot of capital pouring into that MEP space. Um, and that's, that's, it's intriguing me more from a point of view of what's the conviction of capital that's investing, rather than the, um, the, the, the nature of the workflows that are going on. Mm hmm. What's your thesis behind that? Like, there has been an obscene amount of capital in a subset of the design space. Why? So, so, you know, um, a Patrick, if Patrick was here, he'd probably say, it's foam. So what, once one or two have got some rounds, then everybody else follows. And I, follows, and I can't possibly comment whether that's how VCs are investing and just want to match each other's portfolio. Um, so not to be left behind. And so they're all placing a bet on MEP and they're looking for MEP. Um, I think that there is, there's some, you know, deep logic on it, I think, around MEP as a, um, as a category, first has had a lot of, uh, demand because of the data center build out has a lot of demand on it because we need to have high, you know, higher performance buildings. And if you want to get to your net zero ambition, you just have to make smarter buildings, smarter MEP design systems. And, uh, and, and historically coordinating the design across those disciplines, um, is, is, is, is hard, you know, um, we don't have a design coordination system in, we've only got a clash detection system. A clash detection system always seemed to me to be completely insane. Like, uh, space, space, we, we, we, we want to design augmentation and design coordination system, not, um, not, not something that tells us that there's a pie, um, that's going through a wall. And that's, it's complex. That's a clash. And it's like, well, I haven't designed a builder's work yet. So yeah, you know, you're like, you have one thousand clashes. You're one thousand clashes. You're worried about 50 of them. And, and so there is, there is, um, there, there is potential, uh, definitely for radical automation, um, in, in a design system rather than a documentation system. And that's, uh, that's our thing where, uh, this, this is coming from. And it's, it's a reasonable scope, right? And in terms of fees of the design fee and the, the, the, the cost of a building, it's after, after the structure in the facade, when you'd say there's not much left, but, um, MEP is a big package. It's kind of, kind of interesting, right? When you think about these companies, um, they'll be in funding now in the MEP design space. And I think that, like, there's a kind of shortage of people who, uh, who, who are designers within this, uh, within the world that would do this stuff manually. So professional services, design engineers say, and then I guess, like, all of these startups are going to be competing to hire these people, um, and have them on, on, on their company, in their companies, uh, on a technical perspective and helping them build out their product. So I, I, I guess you'll have maybe could result in some kind of, um, battle over the scarcity of talent as these people can't be at all companies. And then, uh, does, does, uh, does the kind of, uh, argument fall into the, the place where, uh, uh, you're, you can, you can only win this game. I don't know if there, if there is a, we'll be a winner of the, if it even makes sense for them to, to be one winner. But can you win this if you can't pay the biggest salaries or attract the greatest, uh, engineers to your, to your, um, organizations? You know, I think the base of design and how to design a building is in every book. You know, you can go to university and learn how to design. And so, um, that's already in, in public domain and machine learns in, in, in, in, of your favorite AI tool, right? So, um, so actually, um, that's not, that reflecting that into a tool or, or getting close to write the, uh, to, to, to, to write the equations that you need to solve. It's, it's, it's, it's available. You don't, I don't need to buy software to do that. You're going to, and, and, and if you say, no, no, no, I'm, I'm going to, I'm going to help, um, by helping an engineer, uh, that knows what works and that collaborate with the client to drive the optimization and understand the compromise, um, between the facade design, the heat gains and the ventilation and the plant room and the GFA's, et cetera, and drive that conversation. Um, then I'm, I'm going like, documentation is so secondary to that. Um, well, producing the drawing is just a prompt away soon. What, where, where is the software? Um, where's the software offer? Yeah. I mean, I think it's an interesting argument against how much money is when, dumped into reading documents. Oh, yeah. Well, that one is puzzling, right? I mean, that one is puzzling. It's like, if the 20 years of BIM hadn't happened, like, the whole point about BIM was document are a derivative of a semantic representation of the design instance. And so you're, you're saying like, I'm deploying AI to read the drawing for which already have a model. And we'd like to get the semantics out of the lines. And then the construction folks would say the model's not conditioned for construction quantities. But that's, that's a condition on top of the model. And, and so to contractively put it in, uh, the delivery of the project, what you do is you say this project will be delivered to the principle of ISO 9650. And the contractor supply chain will contractively agree with the client what information semantic information will be provided by the design supply chain. It's resolved. Right? Now, I've just expressed it in a way that is probably incomprehensible by the majority of the world. I get that. But the idea that the client says their designer give the designer to the to the builders in a way that they can count the objects and they don't have to measure it on the drawing, it's a solved problem. Yeah. And it's not. It's it's not if you the model doesn't contain all the LODs necessary for all the trades to actually in general, you don't put enough money into completing all the designs of the world. But that's really, that's really, I mean, we're going to have somebody in a minute that's going to talk about multi LOD and automated LOD definition, right? So LOD hopping. Ultimately, you the reason why the designers doesn't do that is because it wants to give flexibility to the supply chain that will build to resolve and have a bit of a loose fit to align things themselves, right? If you'd say no, no, no, no, actually, and we see it in facade system or we see it in structural system, if you are, if you are designing like the dome of the Abu Dhabi Luh, you're going to be very, very specific down to manufacturing precision of what is the geometry, what are the tolerances, and how does this whole fit together, right? And in in in a design firm like Berahapo, when we get a contractor that says, I want the full spool files with the cuts and and and the hangers, et cetera, we say, sure, no, no, no problem. You're get get your get your team in, tell us what machine you've got, and we write code that generates those full files based on the LOD what in America, you'd call an LOD 200. Like it's it's it's it's just a computational problem, right? And we've been doing this for 10 years, but typically that's not how the contractor wants to do it. The contractor wants to say, I've got the design intent. We know we need to have the the cabling going this way. Leave it to us to talk to our supply channel. How are we going to align all this so that it's easy to manufacture? It's, you know, it I think they really are not looking at it from first principle and and thinking, oh, you know, the model doesn't contain all the information. The model contains the information that is required and that that represents the decisions that have been made. If we're saying deliberately the supply chain, we make the decisions, then I'm not going to put this information into the model. It's it gets me back to a funny argument. I used to get in with people in crypto, right? So a lot of crypto people would be like, oh, you can transfer money instantaneously anywhere. And I go, but that argument is you are actually against the financial banking institutions regulations. You think banks can't instantly deliver money into any account in the world that they want? No, there is regulations for safety and for fraud protection. And so it's it's these things that go, oh, it's a problem because I don't like it. And thereby I'm giving a solution to circumvent a problem I don't like when the problem is actually, no, we as a society or we as a industry have deemed this to be the best way to operate based upon historical precedents and based upon what we know that some things are better left at the trade level. Some people are better left at the design level. And so I think it's interesting when you get a lot of people outside this industry that come in and then think that they've discovered something that no one else has ever thought about in this space. And all of a sudden they can now pontificate holy shit. I found the holy grail that's going to solve all of construction. And this brings me back to the greatest failure ever was Katera. And I remember sitting in an entire room of investors in Tom Flee's will testify. I stood up and said, this entire company's failing, tire company. There was investors that were in Katera. And they're like, who is this guy? I'm like, she's just a guy that understands construction. It's just failing. I mean, look, you're kind of going going against the startup mantra advice here, Dustin, which is to do hard things. There's hard things. But then you have to have a reasonable path to execute against them. And I think a lot of these people talk about big ideas. They can sell investors that are uneducated and uneducated. I mean, as nicely as possible, I would like to say something more aligned with what I truly believe. As you said before this call, they are good at moving money around. So if you speak their language, they're very likely to give you money, very unlikely to understand the hurdles that you're about to face. And don't know that you're not going to be able to execute against them. And then, yeah, we create this thing where a lot of companies get funded and a lot of companies don't get funded. Where if it was my capital, I would probably have done the inverse in my investment decisions. So that's where I'm going to give, I think the startups have the apology and Dustin, you've been there. Your apology to actually, if they've got enough of a runway and enough of an open mind. So they don't come at it with, I know an answer. I'm like, no, no, I'm going to try to find an answer. And so, you know, same MEP category, if you're going to say, what I think the gap is, is there isn't a modern AI first MEP firm. Because all the MEP firms, all these people, and they've been doing it forever, is going to be taking a long time to rewire them. And on top of that, the current software vendors are going slow at iterating and creating the new stack. So I could come over and say, actually deconsulting experience. I've got a simple building. And I would like an MEP system. Tell me why, and I, if you told us that you can be computed, and that will happen, does that. Tell me why it takes, you know, three weeks to get the first drawing. Tell me why it then takes another two weeks to change it, right? And, you know, if the building is complex, I'd say, the reasons are we need to review it. We need to recompute it. It's an edge case. We need to review all sorts of codes, etc. But if it's a simple building, if there is ever such a thing, then why can't I just have a consulting very, very iterative experience? Can I push there? Can I push here? Can I change it? And I think that, that's the gap. The gap is not the traditional consultants buying a technology is a new type of consultants that saying, I think the traditional guys are not going to move quick enough. And the guy in-house might say, look, actually, massing supermarket on plots and understanding what's feasible on that plot, that I'd bring that in-house, asking for a consultant to do it out of house, etc. I've got all the knowledge I've built 200, 300 already. And then I think that's the, that's the, that's the real opportunity where technology unlocks a different type of consulting or or bring it back in-house to those that used to ask the consultants and now they're saying, well, do it in-house. I think that's a very, that's a very scary thought for a consultancy that we might be slow. Well, I think consultancy is going to shaken up a lot here. A lot of the work, yeah. Yeah. Even from a design perspective, right? Like, we can talk about- Well, even from like a cost consultant perspective on owner side, like, yes. I mean, we all know my thesis as we need pricing intelligence. And if we can have something that can very quickly price conceptual designs and give you a very accurate, like, why are we spending, we're spending billion, the industry spends billions a year on cost consultants, billions. Well, well, I was going to, I was going to, I was going to draw, I'm bringing my point into this notion of the master builder, right? Which I think, I think you made a note, you made a note about, earlier on, on, on, on our notes, but one place tell, we were in Boston last week and Suffolk certainly see that they're like, we're turn of the master builder model coming, coming, coming, coming fast. And they're kind of like pushing that narrative right now. So, I don't know. What do you guys think? Do you think that, do you think that was like- Before we launch, I think we have to give like a bit of context of what we understand by the master builder. And so I think we are at the end of a 500 year journey where Designing. You know getting something built was done by the so-called master tradesman a master builder and what was unique about that situation 500 years ago was that you had somebody that could in their own brain understand the design and what it's trying to achieve in terms of outcome. And how that design would be physically realized and so we talk about the reintegration for example in software of the product owner the engineer the software engineer. And the designers right so PDE is being completely merged and I think master builder is exactly that same idea that after 500 years of building becoming really complicated. The number of parties become really excessively large the complexity of each trade becoming really the it was not possible for master builder to exist I mean human brain as a limitation and the speed up which we want to get everything delivered today you know cathedral had master builder. But they took 150 years so so the ideas can can the computer that was you know and kill point can the computer become the master builder. But can we get back to this idea of really reintegrating the making the designing for outcome and the engineering and I think it's a really really interesting moment. And by our estimate we think that that will generate about 40% gains in terms of horizontal integration vertical integration together will give us a 40% gains and you have a lot of data around. We found the best way to do it without computer and it's hard and it's a miracle but with a computer could you get back to the master builder that's that's what we mean by master builder now discuss whether you see it happening soon never. It's a really good name for something that you know ultimately for example sci-fi Martin fish is talked about reintegrating the the delivery contractually and they develop with DPR the whole IPD contracting so to to to be able to. This I love the delivery because ultimately the reason why with with then exploded and don't have a master builder is is because each person takes what they good at. And then draw a scope around it and then contract it which means that nobody really can improve how how the whole value chain works so IPD is a contractual form but then you still have to be able to do it at a process level. And if all the firms are exploded and have never actually got themselves together then they can say that for example they'll do design for manufacturing and assembly but then if you say an architect is going to do that with all due respect. I'm not expecting the person that is assembling the age back system to know how to design an elegant and cognitively really beautiful space that people are going to love being in right. That's not what they good up so why can I expect an architect to know how to assemble the age back system but yet they happily sell you the FMA service the car right that and so so the the really interesting moment is about knowledge actually could be into knowledge systems that are guiding the design so when you're doing the FMA. You should be able to design a system with a good view to what manufacturing and assembly constraint for that side for that method might be. And then as you move forward and the manufacturing and delivery team are starting to say we could do better etc the design system response and say no no if you change that the sustainability and the profitability of this hotel goes to shit don't do that right because ultimately. That's what you're trying to uncouple is to find ways to assemble better without compromising the outcome and achieve the outcome without making things are super hard to make. I think when the master builder did but I think the modern master builder what someone is maybe what I am looking for I'll say like like what I'm feeling like it's evolving to is a GC that continues to. Self perform more and more of the trades and that you are seeing that more and more I only think that a handful will be successful at that model because it also requires a level of operational excellence to not good time sure you're going to make a lot of money bad times you're going to bleed a lot of cash. Because you're going to have more people manage and you won't be able to diversify that is your self perform components across many different customer bases and many different verticals. So I think that from the from the build side people are looking for hey go here and I see all this savings because they're not passing through a bunch of gross margins to these subcontractors. I think you could do it in like a hyper specific vertical so maybe data centers is a vertical you could do that and I actually know of one GC that self performs almost all data centers. But as soon as you try to do that like even in healthcare I don't think you could do it because I don't think you'll have enough repetitive work maybe on one campus you might be able to build a supply chain that because you're just always noticing that campus but it's that diversification of you know construction is a wavy industry. And so that becomes I think the challenge but a master builder in that that knowledge is more easily distributed I do things where we're having. Yeah, yeah, and you're putting your finger on the really the good you know the the the great points like in the world of master builders. The builders themselves were self employees right they were tradesmen they were individual tradesmen they were the one substance and GC's right so so they turned up to work and and in some trades like electricians that's how it works already right. But in others they are indeed employed in firms and the firm needs to be able to help across because even what cyclical is which firm wins the job what's not cyclical is the industry demand right and we've got some fantastic economic analysis from Technion you know the manual sacks have done if you could desegregate that amount into much smaller firms what would be the efficiency of the system of construction in Chicago and he proved that by desire if you could desegregate that amount you you get about 10% efficiency gain on the delivery of all MP in the Chicago area so this is like real economic analysis but it requires the ability to desegregate the packages across buildings and across firms. Yeah, I'll draw on on the point what Dustin says where it like it's kind of industry specific I think that what I find I know it's a pattern of in construction is everyone tries to paint the industry with like one brush. But it is such a diverse fragmented different kind of industry that like it's made up of many industries within one industry many sectors within that as well and so like if you just say one rule like the master builder like yeah it will happen probably in some areas but in a lot of areas it won't and I think scanning across these sectors to understand what has been learned in one place or the other I think it's it's so true over and I want one of my favorite line actually because you wrote it up is to say that construction is like easier and the reason why I say that is if I'm telling you that tonight I'm flying to Asia you can't tell me what I'm eating tonight because it could be an idly in Bangalore it could be Tom Yanman Bangkok it could be a peaking dark in Beijing or it could be sushi in Tokyo band whatever whatever way you describe it all sounds brilliant I know it's going to be a fantastic meal it's going to be really interesting it's a lot to be learned about how how that's going to happen and how that dish is going to happen and what the ceremony is going to be but you it's so diverse right it's so that you just make that up that is a great analogy just you just with that on the fly or it's someone told you that one before it's one of my favorite explanation you know we we're talking about what 11% of the world economy and as a broad brush and and and I try to say well if you put the whole of mobility together you still don't get anywhere near like 2% of the world economy and mobility will include planes Train. games, airports, railways, cars, bikes, that's mobility, right? And then you say, I will run a car company like I run an airline company, I don't think so. And you said, but it's all mobility. And that's like a small sector. People, let's prefer talking about sushi's, I think they're quite right. Very good. But I remember that one for future. Very very cool. I think I'll tell us going to join us in a second. Quick, quick one sentence on ground break Dustin, as we haven't seen you since, um, yeah. So I thought it was an interesting in, it was interesting because in general, I felt like the message was, we are focused on our core competency of we help builders build, um, and we target owners, GCs and subs. There was almost no mention of supporting architecture engineers. It was, we help you manage the building process, which I think is a great strategy. They have a lot of work to do. Um, I know they're investing a lot in AI. I think that's a lot of rebuilding Procore from the ground up if they want to make it very agentic friendly. So that was the overarching message was we're doubling down on making building better versus, you know, they every year, they're like, Oh, we're going into payments. We're going into something else. And you're like, where is this company going? And that was just, we're helping people build. So interesting. Yeah. I mean, financial products and stuff. But yeah, very nice. Cool. Good one on out of good to, good to see you guys. So how was the launch? Tell us more. Just drowned in it. A lot of conversations, a lot of usage. Just trying to keep up. So tell us more like what did you, what did you announce? And, um, yeah, yes. So I think this will be transformational moment for us because when we started the company, we always wanted to span the entire project life cycle from design to construction. And we believed the incumbents are not really doing justice, uh, in solving the urgent problems of the industry. And with this launch, that changes. And we do it in an AI native manner. So the agent is matured to not just help in early stage space planning, uh, but to actually entire the span the entire project life cycle into detail design structure, MEP, including permit and construction drawings can now happen outside the incumbent ecosystem, outside the autodex ecosystem inside Snapro. And they can happen in an AI native manner, which changes the equation altogether of what's possible, uh, for the industry. Mm hmm. Very cool. As a, how do we, um, like, I guess there's a few few companies claiming along similar lines right about like, um, designing things with, with a prompt and so on and so forth. Like, how do, how do people as an industry kind of debunk or, um, understand like, if I use company A, I'm actually can, I'm actually going to get a buildable design of where I use company B, um, I'm going to get, uh, something buildable, but something is better than the other designer at this, but it's not quite the full thing. And I think there's a lot of like confusion and maybe smoke a mirror in the industry. Yeah, I think given what AI, uh, has enabled, like, there's a lot of smokes and mirrors, but the real proof is on one side, just trying it out on a small activity. Just sign up to these programs. We are also available on, on our website that you can sign up and start using us. Uh, and the second nature is like, who is using this in the market? Like we have customers who do very complex buildings, right from healthcare projects to aviation planning, airports to a lot of mixed use and multifamily projects. They all get designed on Snapro today. And so we have the credibility where people are using this on a day-to-day basis. And with this proof point, we have also published a paper about what AI can actually do on real design tasks, be it space planning, be it authoring, be it BIM modeling, and how the agent actually performs on par with like what a human team would, would get to in, in a matter of weeks. So all of those proof points are available. I think the only way to know whether things work or not is try it. And I think in a matter of months, you'll see like the chips will fall into place because people will use them. And you'll see enterprises are allowed to a certain type of tool and they will outright reject the ones which are just, uh, prototypes. To tie back to, we were talking a little bit about, you know, firms in another sector. But, you know, coming up with more tools. And we talked about, you know, the ultimate customer bringing in the design in house versus the customer continuing to work with these consultants in the, the hope that they would adopt the, the new, agenteic way of working. What, what are you saying? Um, it was the new, with, with, with the new agenteic system. Should the consultant be worried that their, their, their customers are going to say, look, I'm going to use this. Yeah. I'm going to use the cliched statement. Uh, because the technology is now ubiquitous because it's available to everyone. So we are seeing a lot of inbound from owners and developers who have in house teams directly seeing, can they actually like deploy AI solutions like Snaprood to automate their design process? Because at the end of the day, they benefit the most if a project opens a month faster. Uh, if they can predict and control the cost a little bit more earlier. Uh, and architects, if they are not deploying AI and their incentives are not aligned, they are being left behind. So I think this is a phase where an architect is wanting to be a software engineer by building, uh, by themselves. And also, I think their customers for the owners are asking the same question, can I be an architect myself and just eliminate the consultant? So it's going to be like a weird period, but very soon we'll see like expertise will prevail in each segment. Like an owner is an owner because they can do that well. Architect is an architect because they can do that well. And the software engineer and the software company is better because they can do that well. But in this moment, where things are in flux, we'll see each one of this segment claims to be better or the other thing, which they will not be. Uh, it is interesting. I mean, this, this point, I mean, you're saying inbounds in both. I think that's really, you know, interesting to understand, as you say, the only thing that's going to happen is those that don't use AI going to be left behind. But in that world of actually starting to use, who has the highest propensity of wanting to try? You know, that because that's what it boils down to, right? Is, is I'm an architect in a developer? Will I, will I want to adopt versus I'm an architect in an architect world practice? Will I want to adopt? Do you see any difference? I just, this is the same. We are seeing that owners and developers have a lot more agility to try and deploy. I'll give you examples without naming some of the names, people who do a lot of cookie cutter designs. Like, let's take restaurant chains, some of the big restaurant chains. They already have a set design standard and they just want to execute. I want to open 10 new stores, 15 new stores. What is an architect doing over here, taking like 12 weeks to get me a permit drawing set? Can I take it down to four weeks? Get it permitted and if there are comments in the permit, I don't want them waiting around two weeks to address those comments. Just do it within two days. This is changing things and we are seeing not just across restaurants, but like so many such verticals where they are like architects and consultants are actually slowing me down. And on the architecture end, I feel like some of the big practices, they feel that they will never go away. They have this comfort that, hey, I'm always going to be hired because like, you cannot get this thing without me. That's why they're like, I'll take it slowly. I will do it in the next quarter. I'll do it two quarters later and just being skeptical because they use AI to write an email with GPT 3.5 or a four year question. They don't even know what Fable can do. So I think each team needs to come to that realization, but I'm seeing that owners and developers are coming to that realization faster because they think financially. They want to deploy capital and they want to return on the capital. Architects are like, I'm billing by the hour in some form. So I'll always going to get this fee no matter what on an existing project. That's interesting. That's interesting. No, I think it's interesting. I would have thought architectural firms are leaning in more because how are they going to stay competitive to the other people in the market? Like, it just don't understand how any architectural firm, not deploying an AI strategy has, I don't know, any odds? Unless they are the top firm, average architecture firm, if they don't deploy AI, are going to be in trouble. I think we have to, at least they put a slightly different hat, right? So ultimately, the deliverable of the design team aimed a set of drawings. The set of drawings is automated the way with AI. It's really the consideration of one's consultation and what works. So if I am a retail restaurant chef, I absolutely in house know what works. And I will be open to go to an architect that said, I want something a little bit more funky here or I need to understand, you know, because, but if I'm a chain, I know what works. And I do need a drawing, produce that drawing based on what I know what works. And most of the problem will be, I'm opening, I remember these, you know, stores opening in new jurisdiction. And then having to teach their architect, what works. And the value added of the architect was to get the stamp, because only they protected by regulation to get the stamp for the planning, from the planning authority. Hopefully that dies, right? It's, it's, it's, it, and, and because that was never architecture. Yeah, that's actually one of our number one request. All the owners are asking me, if you know anybody who's ready to review and stamp these drawings, I'm just going to hire them on a contract. Just give, give me an army of people. I will just do it all by myself. That's an expensive stamp, if they try and find one. But, you know, if we look at the in in a different lens, right? Ultimately, whether it is in London or in the US, we, we had the Hackney carriage license, right? So the way you took a taxi in London, because to be safe, you knew that this post-grown carriage had the Hackney carriage license, it was safe. And then it became taxis and then they, they were educated, etc. But once you have technology that actually guides the taxi through the city faster, tracks where you are and make you safe. And the identity of the driver is now because there's a in in cab camera, the whole licensing system is a necessary. And it becomes a barrier to better mobility through the city. And, and really, I think we are at this moment where we have to really think about our deeply respected architecture and our thing is about consultation, outcomes, giving customer the belief that their headquarters is going to work as as a headquarter or making circulation work in their retail stores, but not producing drawings and not getting them stamp. That's not value adding anymore. Yeah, let me add to this point, like when is of course this cookie cutter projects, right? In the AI that we have built and some of the examples which I'll put out over the next two weeks, I did a project which HOK has designed and is in construction right now. It's on their website. And I asked AI, you take a first principles approach to redesigning this hospital. And it's a 244,000 square feet project. So it's a complex one from architecture to permit drawings like with MEP structure, etc. It was able to get to 1.5% of the area through first principles. It was able to achieve daylight efficiency. It was able to adhere to adjacencies of FGI guidelines. And it has like 40 other metrics at which it ranked itself and the design that was approved. And we are not very far away. We are about like 5 to 10% away from what the architect eventually approved. And the AI did it over four days. I had to run it in loops. It burnt about like a lot of tokens, about 1200 dollars for tokens. But we came very, very close to what was achieved. And it had MEP structure, architecture, everything. And I think that, you know, this is completely changing the narrative of, you know, let's call it what we're the rabbit killer, right? Ultimately, what people have been saying, you know, we need the rabbit killer. But actually, we didn't need the rabbit killer because rabbit is a computer-added drafting system. What we wanted was computer-added augmented design system. I was on the client side and we were giving 12,000 pages of design specification. And we're giving the architect six weeks to come up with a first, you know, response. 12,000 pages. I mean, you're going to be a efficient reader, right? 12 pages helping you ever absorb 12,000 pages of design. Whereas the AI says no problem. I mean, you're going to have to spend a few tokens as I'll tell you. But suddenly, like I'm designing from day one, and the system is going to say, I'll stop you when this paragraph here, you're not complying to that design requirement or the design specification, right? And that, I think, out of you called it multimodal, well, so we're seeing a completely new generation of technologies that can understand multimodal, whether it's text, program, space, drawings, and we also addressing the central problem of the designer, which is to absorb and manage complexity. Yeah, the fundamental objective was not to design buildings faster. It was the cliched McKinsey report, which I highlighted. I know it's very debatable. But each product, a lot of projects are still overrun by cost, overrun by budget. Can we just focus on those problems? And I think design has a good level to contribute over there. It can actually help you explore more options. I'll give you another example. In the middle of a project, it's inevitable to have value engineering show up. Can you take a project, use AI to explore 500 different options to value engineer this project? And then actually optimize the budget of that project and the time of the project, instead of just doing like the traditional stuff. And I think these technologies unlock that. It was never about to kill Revit or like, replace Revit. It was always about like, can we have more efficient buildings more across every segment, sustainability, costing, timely manner delivery? I think we can like finally address those. Do you have any data about value engineering? Because this is one of my triggers, as you can see. I'll give an example on one of the projects which we are doing. But value engineering in construction means getting this go. It's not value engineering. It's removing functionality. That's not value. There is no one who's ramped on this topic more than me. The biggest problem we have in pre-construction is you start with the design that an owner can't afford. And then you get a price and it's 50 percent. It's not even close. It's 100 percent more than the owner was like 100 percent. And then we go value engineering and then pre-construction last three years. And people are working like, this is where I get to do my rant again. This is why I get confused where people are like, oh, we're solving the hardest problem in pre-construction. No, takeoff is extracting what is known from the documents after you've gotten through all of this pre-con value engineering process to bid on it. The real value is compressing that two to three years span. The only way you compress that is to have instantaneous pricing feedback to design iterations. Yes. I only know one thing. Value framework. I mean, the first thing we developed here in the UK around infrastructure project delivery is a value framework. What's perceived value? And then we try to run that through product design frameworks called Quality Function Deployment, where you are saying, what is the sponsor value in terms of qualities. And then you relentlessly iterate to get quality function deployment, which is a design methodology that allows you to value engineer. Really. Like, I mean, if you got a master's degree, the overall value, not just got cost, that is Under different times, different terms. cost reduction and value engineering, two different target value designs. We've never practiced it once. Never. We've ever in the history of construction, ever. Now, now I'll have, can we? Are you saying that is? Have we got someone else coming in, by the way, oh, NSX. So I thought we'll need to wrap you up, my friend. But yeah, that's, that's a message software is all. Let's solve the harder problems in construction, which is getting predictable buildings, efficient buildings in a timely manner. Very good. Good work, man. See you soon, yeah. Congratulations. Thanks. Thanks. Thanks a lot. Hi, next up, we've got Andreas from nomic suite. Hello, Andreas. Hey, buddy. We are fantastic, huh? You doing well. Well, this is the first in Bricks, Bucks and Bites where, um, one of the companies I led the investment in. Oh, why then, Dustin was so special about Andreas. Um, he and his co founder are exceptional individuals. And they're working on something that's really, really hard. It's a combination of hardware and software to measure one of the most valuable physical assets we have in this entire world. And they're getting to the source of truth about what's actually going on in our utilities. So I'll let him explain his business. Tell us more, Andreas. Yeah, sure. I think you captured it on the surface level very well. Doesn't appreciate the kind words. I mean, what we do essentially is get information that's impossible to capture for great operators today. I think it's kind of interesting to go into that industry and look at what information they operate on is such a black box, uh, and the fact that they've been able to operate is quite impressive, uh, frankly. Um, but times have changed for that industry. So what we do is we, we have our drones. They're drones that can land and recharge on power lines. They're essentially mobile sensor platforms. They let great operators put a sensor up without turning anything off or installing something permanently. Uh, and at the same time, they could capture kind of things like the temperature of the power line, how much of vibrates? What it looks like around it, um, all without really needing people in the field. So our drones hop around and position themselves intelligently. So, um, yeah, uh, uh, fundamental, um, led this round and, um, I'm joining as a board observer. Um, I think the thing that was, uh, most intriguing to me is the fact that these utilities would even let drones land on their power lines shows that they know they need this information because they're associated with them letting anyone touch this mission critical system shows that they already are aware that that information is critical. They do not have access to it. So if you can provide, provide them access, you can theoretically help them operate their entire infrastructure more, uh, efficiently and make better capital allocations decisions in the future of how to expand their infrastructure. So I think that this is very critical, um, for centuries. And how did you convince them to let them let them land returns on their power lines? Well, you know, we started out in Berkeley, my co-founder, me and we were looking at the problem. So we're building the first prototypes. And when we came back to Europe to kind of, you know, talk to all the different utilities here, um, it became clear that there's almost no one looking at that market because they're kind of famously known within kind of VC world as the slowest industry to work with. They're painfully like procurement cycles. So no one wanted to touch it. And I think that made the clear that they were just underserved of technology. And when we presented this to them, I think the interesting thing is like everyone had thought about this. Like there's clearly power. There's clearly something that we can do to kind of extend range of drones. And most of our customers, you know, they said, uh, we actually tried this at some point, but it just didn't fully work. Um, and the, our first customer, Vesnetz, they were, they're a, a German distribution system operator. So you can think about the power in us highways and roads and they're a road service provider. So they, they kind of have the roads of the power grid and they initially just went like, Hey, you know, we really think this is an idea that we definitely want to buy it too. Um, and they let us first test on a deenergized power line. So we started it kind of on low stakes. We didn't go straight to like the highest voltages. Uh, but surprisingly within within kind of a year from that first landing, and we show them that the drone could kind of fly land, operate by itself. And then a month later, we were on energized lines. So it's a lot about trust. We're working very closely with our customers. So we can assure like, obviously they know that not everything we give them will be finished in the first months. Um, but we've really shown them that our iterations feed in our cycle and building something is really difficult to work with like high voltages. There's some weird mechanics there as you go out and voltages. Uh, so working with them, and explaining through how that actually works and showing them, okay, technically, this is possible. Uh, that has been important for us. So definitely some steps there of building up trust, but once you have that trust, it's pretty clear that they're interested in the data. So totally worth it for them to try. Hmm, Andreas, like, so what's the kind of value at the back end of the, like the data that you uncover for them? Like, what is it? Unlock for the grid, uh, the grid line operas. Yeah, I think that the most obvious one that we work with, uh, most of our customers with is increasing capacity on existing lines. Uh, I don't think there's a single utility in the world that has like an abundance of power lines now because simply the, the requirements have changed in the last 10 years. Uh, previously you had pretty basic systems where you had like, you know, one big power plant, maybe it was a coal power plant and burn that coal or in generated heat, spinning some turbines. And then on the other side, you had a big factory. Uh, the reality today is like your live data centers where the, the consumption just swings insanely high and low. Um, so the reality for them has changed. Um, and what we helped them with is very simple measurements. We tell them, hey, this is how much your power line is sagging at the moment because this tells them how much clearance they have on, on current, like a pastor, uh, it's very, very basic measurements. But the challenge has always been they have to assume that it's like the warmest day of the year. It's the most like, you know, terrible conditions for power lines because that will happen at some point. Uh, and they don't have any way to measure it. So then that's kind of the standard operating procedure. Uh, so that's use case number one. What we see a lot of now is, is actually on security and surveillance. So for power operators, one challenge is, of course, you have a massive piece of machinery that's everywhere in your country. Uh, how do you, how do you get to know like when someone comes in and goes to sabotage it? And recently there's been a lot of sabotage, especially in Europe on, on power infrastructure. How do you monitor this? Like you cannot technically, you cannot install, uh, security camera everywhere. And every place they need to install something permanently is, uh, you know, years of planning and approval from the city. But the good thing about power lines is, you know, if we place our drone on a power line, we have a great vantage point looking down at everything below. So see you have a substation or a kind of an end point where you go from over at the underground, you get a great vantage point. It's like, if you try to take down that camera, you die because it's high voltage, uh, it's pretty simply protected, um, and it gives them information about, okay, who passes through is there a car that arrives at 2 a.m. in the night, uh, is there like a person coming in, uh, and we can fly feed this into kind of the security control rooms, uh, at the, at the great operator's offices. So that's some of the, some of the topics that we work a lot on. Uh, there's an abundance of different challenges they could solve with the same things, equipment, but that's maybe the two ones that drive a lot of interest. Yeah, fascinating. Problem. How did you come up with this solution? So, um, my background is originally in software engineers. There's really nothing to do with, with kind of powerbreds in this space. Um, but one I was about a couple of years into working, two years in the working, I realized there has to be something more meaningful to do. Uh, I worked when I started first and I thought that was incredibly fun, uh, to do, uh, and then I stumbled upon a new program, uh, in Norway where I'm from, uh, that was specifically designed because they needed more people to work on what's called smart grids. And smart grids is essentially that you want to digitize how power grids operate because they did never went through the digital revolution. So everything there is analog. And this sparked a lot of interest because obviously you can make a huge difference in how people live their lives. Like power is needed anywhere. And if you are kind of a, uh, uh, a great country with the great industry, the number one factor is like, okay, you have an abundance of energy available. Uh, so I thought, if I want to make a big difference, this is where I can do it. Uh, there's so many problems to look at. But one thing I saw was there's difficulty of getting information deep into those targets. So I thought if there is a strong current, there must be a strong magnetic field. Uh, and we can take that magnetic field and essentially charge the same way ish that your phone can wirelessly charge. And if we have this, you know, you have the distribution problem. How do you ensure that that sensor is where it needs to be? So let's just put it on the drum and and see. And I kind of had this in the end of the back of my mind. And then I met my co-founder, Lawrence and Berkeley. Uh, we were both international students. And I mean, Lawrence is an insane engineer. He's really, really talented in anything physical. And we started working on, okay, hey, let's try to talk to all. the different utilities that we can just to like see if this is even worth building because most ideas are shit and we kind of didn't know the industry too well so we just needed to educate ourselves and we had a moment when we met alignment and alignment are people who work on power lines and we asked them hey you know we're students at Berkeley can we follow you on a day at work just see what you do he did inspections and pacifico which is a bit south of San Francisco and and it was super fascinating because he kind of pitched us the entire idea that we had in line without telling him about what we're going to do is like I spent so much time just like walking around not doing anything because I need to just get to where the starting point of the information capture starts and he said if there's a drone and I could like fly here and just wouldn't be limited by battery capacity I mean that would save so many of my problems because today the drone team spent equally as much time just to get to the same place that he gets to and we were like oh my god we have to build this it's too too attractive nice yeah the light bulb moment yeah and what so in terms of that you know drone network that's that's recharging um you know the I think the right to charge on the line is it's a big unlock but how far can the drone travel away from the grid yeah that's a great question I mean it's it's optimization algorithms that we can do for this because there's always some kind of interconnection so if you can hop you can get indefinitely but for the current drone that we operate with it's about 25 minutes of flight time and that doesn't get you necessarily super far off the directions of the power grid but this is just a beginning kind of the range that we can offer and for us range hasn't really been something we've even optimized for like most drone companies need stop and are optimized for this like flight time but for us the optimization is like how long can it stay outdoors without people there how many times cycles can we go through the battery how many different conditions can weather conditions can you support yeah so I think that's been been our our focus and it's it's definitely something different than than most drone companies but I mean we we see so many opportunities of just discovery through the power grid and and all the energy infrastructure closed to because I mean anything that's interesting is close to the power grid I think that's right I think and Dustin you will have seen more than just the team and and that I think that the ride to charge so I don't know whether the infrastructure team guys have the air rides that's that's the next one they that depends a bit on country actually so the interesting thing is that the airspace that's around power grids are kind of reserved air space yeah and in Europe specifically I mean it is also in the US and they have something called the right of way which is like a clearance area below the the power lines so they don't control that airspace in any ways they have internal policies typically for like we fly helicopter now here so you don't crash in it but all other air aircrafts are not allowed to fly in that airspace just because of security reasons so you don't want to crash into power lines dangerous so that's been one of the the positive things about our cases like we don't have to fight anyone over airspace rights it's kind of dedicated specifically for it because ultimately you know I think obviously the use case of the utility for maintaining their network is nice there's no no doubt about that um um funnily enough my father was in just that's in that sector so so ramming the electric network in Europe that was you know I would travel Europe and he would point at the line and say that one is this one and we built it in this well this government you know anyway so it's it seems very circular but the the interesting thing around you know it's a very big scale up zipline that you know they they fix wings they have had to break through the the the the rights of flying over flying they have to go to Africa to really scale and develop as you talking about all the technology so that it becomes robust I think that she I don't know how many millions of delivering and safe flights and landing before you know cities in North America started accepting that maybe sending a parcel that weighs 300 grams into a you know tantrum track they like not it's not that efficient you know so so so I think that whole aerial market and if you've got the corridor if you've got the power if you've got all weather condition if you've got enough predictability but adding this the the the the market is immense and it goes way beyond knowing whether you're gonna optimize your smog grade or road date that the world etc and I think that's very exciting absolutely I mean I I completely obviously I'm very passionate about this but I there's so many opportunities I mean we we see so many things just alone in the power grid and obviously it's like a lot beyond this and but especially now in the beginning it's been important for us and probably will be for the foreseeable future for the company to focus on energy and there may be there will be spinouts what we do you know other industries but but specifically the drone in the air I think that's that's your and what what are you achieving how long does the drone are you at a few hours or a few days or a few months where where where so our drones are have been now deployed for I think we calculated the other day total of 20,000 deployment hours so in total and that's this year essentially and we have drones that sit with customers today on power lines so we have about 10 of them that are with customers today so they just sit there in charge and they observe the locations we measure you know how much current is flowing temperature so we are in an operational state now and a lot of the challenges now is like this okay we have some networks that we obviously want to improve the reliability because the more you have it out there for you know what happens after 10 months 12 months but but we're at this scaling point where okay we need to really build a lot of drones and they need to be really reliable and we need to make sure that not only do we capture their information and not only is it the drone right it's a lot about the information that we capture and present it to the decision makers like what do you know if we measure this temperature and this current what does that tell you about how much capacity you can actually you know change and how do we integrate that in a way that it's actually actionable and that's that's a lot of challenges now that we kind of solve this data collection problem very cool okay Andreas we need to thank you Andreas thank you so much for coming it was a very fascinating all right to hear about I have a great day I take you I'm okay as well that was a nice twist yeah it was a nice something kind of different for what we do here normally so it's it's great investment resting and yeah I think that as always it's not what it seems you know when it starts you know it's a drone yeah the execution I mean they're they're dealing in multiple engineering disciplines right you got material science you have aeronautics you have yeah every every single type of discipline in order to execute that on but you hit the nail on the head how long can those things be serviceable in the air what weather conditions are problematic and understanding all those things to really scale but theoretically it's a big deal all right guys let's wrap up thank you so much we'll see you in another month