Building Simulation Products for the AEC

okay so we are live nice welcome everyone to the third uh iteration of uh infrared cities community talks and I couldn't be more proud and and excited for this one where we are uh we're hosting Mostapha I'm going to try Sadeghipour I mean, thank you for sharing. That was good. That was good. Roudsari, co-founder of Ladybug Tools and Pollination Cloud. I think you need a very little introduction because of the contribution that you have had on the simulations for architecture, which I believe is probably the biggest in my time, but I'll let you speak about that. But before I start, I have to ask you what I asked the previous two Italians that I had in the show. And in your case, I hope that the answer will be a little bit different. So how do you drink your coffee? I don't drink coffee. Excellent. How do you drink your tea then, maybe? Do you drink tea? I do drink tea. I mean, I rarely drink coffee. My tea is like the Iranian way of drinking tea, which is you boil water and then you... We don't boil the tea. It's different from British, how British prefer tea, but it's kind of close. And I drink a lot of tea bags these days, unfortunately. But yeah, if I have time, that's the way I do it. Like the tea leaves and then water. Do you want to know all the details? It's just too long. Just people can Google it. I think it's okay. I think the tea answer is good enough. No milk, no sugar, no nothing. Just tea. Black tea. Black tea. Okay. So then maybe my second question, I don't know if you, I think I have some connection issues, but I hope you can still hear me. The next question that I would have would be, how do you describe yourself? How do you normally, what is it that you're proud of? How would you essentially describe yourself? um I uh so that's a hard one how do we describe my so when people depend on who I'm describing myself to but usually what I say to people or let me put it this way like when my wife asked me what to tell other people about what I do or who I am, because that was the hard question. So she says something around like he went to architecture school, but now he does develop software for engineers and architects. And then if I want to add to that, we develop software to make it easier for architects, engineers to use environmental building simulation during the design lifecycle, design, construction, everything at this point in lifecycle. Yeah, we don't talk about Ladybug tools, I think. Like, she doesn't, but I think, like, if, depending on people, I think, like, what I'm proud of, probably, Ladybug tools is something that I'm proud of. The community, the product, yeah. I think that's a big part of my identity right now. How do you feel about that? Good. I don't know. Like, I don't think bad things happen because of ladybugs or more, or I've seen more of the good things that happened because of ladybugs. So I feel really good about that. I can think at least of, of like, you know, hundreds of others, uh, researchers and practitioners that have done a lot more harm than you have done by, by giving people simulation simulations. I'm pretty sure like they do not do the easy one for Ladybug. Right, I hope, I hope. We really tried hard on that one. Yeah, exactly. So then maybe let's talk about that. Let's talk about Ladybug because I'm not sure everybody knows about it. I mean, I suppose a lot of people know Ladybug, but Let's talk about how it started, because I remember it starting, but I don't know exactly the details. So maybe you want to tell us. So yeah, sure. So Ladybug was started as Ladybug. And Ladybug started as a bunch of Python script user objects that I had to develop because I had to run studies on large-scale buildings when I was working in Chicago in Adrian Smith and Gordon Gill. And it was just so hard. Like I had to, I had a script in MATLAB that I had to use to export the geometry from Rhino kind of Grasshopper to Radiance, then go to Radiance. And at some point I automated in MATLAB to just run the script to run it, then get the result, then put it in Excel. Then I had another thing, read that back, then visualize it in Rhino. Just when I was doing that, you know, that was one part of it, which was like, God, this is so painful. And good things happen. You know, Gulio, if I pronounce his name correctly, put the GH Python component out. And that opened up a lot. Grasshopper was more stable at the time. And the office was Rhino, Grasshopper-based mostly. And it just, like, all these things came together. That was one part of the ladybug becoming ladybug. The other part was I was doing a lot of optimization and machine learning at the time and, you know, putting all this stuff together. And I realized that it's more misleading than helpful. I mean, that's a longer story. But the reason for that was I realized people don't know where the sun is. You know, like I haven't teach them like they didn't have the chance to learn about basics. They don't know the basics. And I'm just giving them something as a result that instead of helping them as an insight to be better designers, to design better buildings, it basically stopped them from thinking. I had people with years of experience coming to me and say, hey, can you optimize this for me? I'm like, optimize what? Just optimize it like you did for that building. I'm like, what do you want to optimize? They couldn't even verbalize that. And I'm like, oh, this is wrong. This is not helping the whole process. And that was the time that I realized, oh, we need a better way to talk about basics, like where the sun is, where the radiation comes from, how many direct hours of sunlight you get in this building, that kind of stuff. And that became Ladybug. That was for me the starting point. And then I shared it with some people. They liked it. I shared it with more people. And then it became Ladybug. Chris joined, brought all his expertise in comfort, thermal comfort in his interest. I already had Honeybee, but I didn't release it for about, I think, two years. I think it was only Chris and I, we were Honeybee users. And the reason is interesting because how we opened up the conversation for me was it wasn't ready. I was worried that people are going to use it wrong. So it didn't feel right. I had everything. I showed it to Chris. We used it. I was just like, this is not how it needs to be. And then I rewrote it like twice. And with my knowledge at the time of like, you know, Python and everything, it was just so painful to rewrite everything. But again, I did. And then at some point it felt ready. And I felt like the community is ready now to introduce, okay, now you can run daylighting too. We talked about basic stuff. And then we moved to energy. And Theodore, at some point, joined and helped with Butterfly. And I asked people not to ask questions about it. But it's good. There are so many other solutions now that people can use, including your solution. Who said not ask questions about it? I said. When I posted it, I said ask questions about everything, but not Butterfly. uh yeah it's it's actually it's an interesting project and then um chris also had this urban design uh dreams and hopes that when that dragonfly came around Theo helped with the spider, like the web-based stuff that we didn't really market and talked about as much, but he did a great job. The GBXML viewer that he developed, I think, still is the best GBXML viewer online, free GBXML viewer out there. So that's kind of like the whole story of LadyBug2s, which happened over years. I think this is a beautiful story of community. I think that's exactly why I think it grew so much, because it was really... It was pure. It was really addressing a problem. It wasn't sort of like, you know, hey, what can we make that would be successful? It was more sort of like, yeah, I need to solve my own problems. I need to actually solve my own needs. But like, hey, look, I've done something that is very useful. I think there's a bit of credit to go to, you know, the employer or whatever, like, allows you to essentially, you know, right I mean I did it all I mean it goes but I did it all on my own time it was uh it's an interesting I did some for the projects that we had to use it but the core development I i was single at the time no kids uh chicago I moved to the neighborhood which is close to the to the only startup at the time that I knew which is like 24 hours I didn't even have a wi-fi at my place So after work, I would go there. It's in Old Town, Chicago. And it's kind of like after midnight, it becomes like a library because there are a lot of medical students. I don't know why they come there and study. And I was developing the tools there with the Starbucks. And I released the first version there from the same Starbucks. And then I would walk back home around like 2, 3 AM sometimes. And it's funny because this street that I have to go down, it's an old town and there are bars and, you know, like, disco, everything, like, on the side. I would walk, you know, and people are just half naked, half drunk walking. I would always see, like, you know, everything was kind of grayscale or everything was colored. That was grayscale walking. And there was a subway who I would stop sometimes to get a sandwich or something. It was funny. The guy asked me once, just like, what are you doing here? Because all the people who would come, they were drunk, and they were just trying to get something. I would get my sandwich to eat something to go back home to sleep. That's how it happened for the most part. Of course, they get the credit of being open to use these new solutions and technologies at the time. I gave them a lot of credit. I've been thanking for a while for allowing me to publish. But I've now realized that I should probably thank the guy. GABRIEL SANCHEZ- Yeah, ASGG. And my co-workers at the time, Anthony Viola, I think, he was a big part of testing and giving feedback. He was one of the first users before it becomes a thing. And we built that. I don't know if you saw, there was a box that we would move around. it will run some things with ladybug and that's also quite sorry because it was so like everything that we were using was so cheap he paid for everything I think and then uh we had to wait for the whole offices to go so they all turn off the lights so it can work otherwise you know it was like it was creating noise so the thing wouldn't work no way no way yeah very interesting yeah like a single life I want to move on to the next one because I think that, you know, there's one thing like this beautiful idea and then, you know, you are essentially working on your own solution. I think for me, like when I remember, I remember seeing Latify at the same time that we were developing things, you know, also Fosters or having sort of like similar ideas. And yeah, to me, it was like really you know nice to see something coming out becoming like a community rather than something you know there's like you know foster partners using this or this you know using these two but nobody gets ever access to this I've actually commented that to to host himself I guess once I told him again it could be like you know your tools that are developed here could be actually you know and pushed to the community and really grow and I think partly so that community collaboration and sort of like you know being able to you know take something like a bit of knowledge of python that you have with some you know deeper sort of environmental engineering knowledge and like you know getting out something and saying to the community hey let's develop this further together right and I think I remember there was a lot of like you know satellite people developing Ladybug with you and with Greece and really sort of like maximizing its capacity. For example, Theodore, funny story is like we actually met, I've met Theodore now, he's my co-founder in for a city. Right, yeah, I think that's actually part of the, I'll get back to what you said, but I think one of the beauties of Ladybug Tools community is like how many people started offices because Ladybug Tools? or wrote papers because of Ladybug tools. And like, it's businesses that because of Ladybug tools, people met on the forum, you know, and that always gives me like, makes me emotional. I think like, that's one of the best parts when I hear people like, I have a job because of you, you know, like, or I changed my, and when they say you, it's basically Ladybug tools. They don't understand. It's only me, but it feels like, and, Yeah, that was big. And I think it's kind of like, you know, mostly I got lucky because now I know a lot about open source projects. I know about like community. I just learned it during the process. But I think there were a few things that I did right. And then there are more things that I got lucky and just worked out. And one thing about community is just like Ladybox is unlike a lot of other projects. We never open sourced it. it was always open source. It was always open because I looked at it as a scientific experiment, too, at the same time. So I always thought, and I still think, like, any scientific tool, you should be able to go and see everything. And it's one you have to make sure, like, is valid, right? You have to make sure, like, this is what you really want. And I know a lot of users don't care, but I care about those people who care, right? You know, those small, like, early adopters, the ones who care. And the other thing is that you should be able to extend it, customize it, extend it, build on top of it. And this is something that open source brings naturally. And it worked out well for us. Again, people ask, why did you open source it? I'm like, I never open sourced it. It was just open. Always. It was just how it was. It started it open. And then I got lucky in a sense of I found a co-founder. I mean, I know Chris because of Ladybugs. We started the company because of Ladybugs. There was no way, you know. I would meet Chris otherwise. Even all these years we have been living in the same city, maybe for like two or three months when he was an intern at Thornton Thomas City in New York, I was working there. Otherwise we never even lived in the same city. And yeah, Chris came along, he was like a force of nature. He is still a force of nature. I did a lot of great work. Abraham helped us a lot by picking it up, teaching it, and providing feedback. He was our semi-automated test for a long time because we didn't have automated tests. So we would release something, and then we're like, OK, Abraham is going to get back to me tomorrow morning with a lot of bugs. Yeah, the other thing, like, because when you're teaching it, I think for me, at least, education has always been, like, the number one, you know, funnel for, like, you know, testing ideas, developing further. Like, I see, for example, in my work in IAC, you get something out to students and they just take it and then they sort of, like, explode with it. So I'm just trying to get to the point where you, you know, I'd like you to discuss not only the benefits, because I think we can all agree that Ladybug's community, Ladybug as an education tool, Ladybug as a scientific collaboration, it is obviously one of the best examples that we had in the industry. We started from someone, from some really great idea, a lot of people jumped on this idea and helped. Even if you look at like, you know, Energy Plus or Radiance or like all of these tools, they were actually developed as open source tools coming from the Department of Energy or other, most of them with grants and funding. But like still there's like people with a lot of passion, like bringing tools for sustainability and sustainability how the industry moves and the business around it. So like, you know, when we talk about software tools in architecture, it's not like really a very rainbow discussion. It's probably like, you know, there is a lot of, there's a lot of like frustration. There's a lot of, you know, there's a lot of like barriers. We are fragmented. And, and, you know, the thing is that if you, if you go to talk to, let's say the transport industry, right. And say, batteries or I don't know, whatever, like electric vehicles or whatever, like maps or any sort of like, you know, aspects of that, you have standards, you have like ways to, you have to be able to plug things together to make them work. And I think we're not doing that in architecture and engineering construction that much. And maybe that's also the pitfall of all of that is essentially that there is no value, defined value, you know, when I'm making a building and I'm delivering that building, I'm not necessarily responsible or sort of like compensated for making it more sustainable, making it really, you know, and that's sort of like, you know, in that way, like simulation products can also be less. I think this is changing, this is probably time to change. But I think, you know, I'd like to hear your thoughts on this. Right. So like all the valid points, I don't know if I can add more there or much more. It's just, yes, it's a challenge that you basically, the incentives are not aligned necessarily. Like the person who has to pay for it is not the person who gets it. things out of it at the same time. And that's where I think regulation are the ones to help. Basically I think like people needs to be, or the industry needs to be forced to use them. It should make sense. It should make financial sense for them or like somehow they should be able to get things back. One of those things that I don't think I can solve, and I'm not necessarily put a lot of time and energy on, you know, thinking about them. And that has been my strategy with Ladybug Tools, with Pollination, with most of the things that we do is just, I, you know, I used to do the work, like, in a sense of design, construction. I used to do construct before coming to the U.S., like, you know, being in the buildings. Then now, like, at some point, I made my decision of, like, you know what? I'm going to go one step back. I'm going to empower other people to do more. I'm going to help them. My thing is I want the people who trusted me, I want that innovators, early adopters who took the risk when things didn't make sense and used it and tried to push things forward look awesome. I want them to look exceptionally smart. And if I can do that, they are the ones responsible to figure out the next steps. I'm like, OK, how do I go against this legal issue? How do I go start this initiative to basically make sure people see in the lifecycle of the building how this can benefit everyone? Those are the things I basically decided, made the decision of, Hey, it's not my concern. Things are going to move in a better way. And to me, it's just like things always move in a more efficient way. Like we go through efficiency, right? The things that are more efficient will survive. Uh, and, uh, I mean, of course, there are other things. Efficiency is something that is relative again. But that's my goal. I'm trying to empower people who really can do much better with the things that I do. And I think bigger picture, some questions like that, I leave it to other people to figure it out. I'm here to help them and support them in any way that they need to look smart to be able to get that done. I love focusing on the early adopters. I love it. Let's talk a little bit about to pollination like so you start doing like and then you say okay this has to work on the cloud and it makes sense to me absolutely and uh yeah like the journey there like and essentially how do you see the future of all this you know we're we're focusing on a little bit of a different angle of like we want to make these tools fast with ai so we're focusing on the ai component of that But yeah, I'd love to hear your thoughts for the future. So I'm not, again, like that's what I'm doing. I'm not that futuristic. I'm very like, you know, in our team, I'm a dreamer, kind of like I have like hopes and dreams. You know, a lot of things that I've built, it starts with a dream of like, I want the world to be like this. but I don't know what happens in the future for everyone. What's the dream that you have right now? That's perhaps what we're looking for. Actually, okay. We will get back to your previous question because that was an important one from Ladybug to Paul Nation. Or maybe actually I can tell how the dream changed. so with ladybug the dream was for me was as I said because of the pain point it was just like I stay in the same environment I can just do it all you know I can move a slider and then I can see the difference I can see the change I can make the decision I don't have to leave this rhino grass every month I can do it all and I can ask questions and it doesn't stop I don't stop like oh in this environment you cannot run this study and that's part of like the the idea behind Ladybug that a model, like an analytical model is a model that can work for daylight and can work with energy. They are different, but like you should just that those nuances is the responsibility of the tool to resolve so the user can use it in different ways. So that was a dream for Ladybug. That's where Ladybug started. When we did all those stuff and then at TT Design Explorer happened, parametric studies became a thing because it was easier, you could run faster. The dream more started to change of, I could see people who don't know Rhino and Grasshopper really struggle. to use this stuff. And we had these people who would see them, and I talked with a lot of mechanical engineers at the time who would say, like, we love this stuff, but I just don't know 3D. I just can't. This grasshopper, like, you know, explodes my mind. I can't get my head around this. the idea for us was, okay, this thing, we even have a slide, like this should go free out of the jail of Grasshopper. At the time, it felt like a limitation. What initially felt like, oh, and is the reason of Ladybug Tools exists is we have Grasshopper. But at some point it grew out of that. And that was the time that the dream was, okay, everyone should have access to this technology. So we wanted the best of Ladybug Tools, what it can do to be available to everyone. And that was when pollination came around of like, okay, what if we make this available on the web as a headless technology that everyone from every tool can use? Turned out the cloud computing itself is not a selling point and it's not a real pain point of the industry people talk about. It's more of a, you know, like a painkiller versus a vitamin. It's more of a vitamin, like people won't take it. So like the ones who actually have no pain, it's kind of like for luxury, you know, like what they call it in a health drink, whatever, that kind of thing. Then during that process, then that helped us. And that's kind of like when that dream changed, we start talking to more people who we wanted to be able to do this stuff. And we realized the main pain point is actually happens much back. So if you have preparing geometry, adding metadata, running a study, visualizing, getting the result, visualizing, understanding it, right? Getting some insight, taking an action, coming back and go through this cycle again. We were just like, because we were coming from the background of like, oh, geometry, we fixed it. Like, no problem, like geometry, you know, like we extracted, we just have a script. We thought the problem is like, oh, you can't run it fast enough. You can't run it in scale enough. You cannot visualize it everywhere. Then we realized, no, wait, like that's like a small fraction of the community or the users that we can help. A lot of people are stuck. They can't get the geometry out. They spend like 80% of the time, 50, 60% of the time of the project just redrawing the whole freaking thing that's already available in a digital format to be able to say run. And then because of that, they have no time to think about running more options or running, you know, like all these things that we think are good. So as a result, and that became now like big part of pollination and pollination used to be called pollination cloud. And we call it now pollination or pollination solutions because we realized there are other solutions we need to build for people to get people there. So it's kind of, we said, okay, let's just stop on the right side of like inside all this. Let's just try to fix the thing that we can fix, but no one gets to it because it's really hard solving the geometry problem and interoperability. Yeah. The other thing I should say was the thing of like this isolated thing of You know, people are, oh, we use Equus for certain things in the office because we're good at, we have these people. And then we want to use ISV for those things. And we use Energy Plus Open Studio for that thing. And every single time we have to redo, remodel everything. And then the dream thing became for us, it was just like, what if you have this central schema that we just built? Validate, figure it out. And then we serve all this other, you know, fall formats. And that became HP JSON and pollination and Rhino and Revit plugin. And I think now we are the only tool that it's kind of became the Rhino for building performance. How Rhino is, you know, you go to Rhino for geometry, you fix and you go somewhere else. That's exactly why I used Rhino in the past. I still use Rhino on a daily basis for prototyping, testing. It's just the best platform for development still out there. Even like it's pretty old, but I think like What they did with that ideology. And that's part of like what Pollination became now product. So we have this Rhino Revit play and we are very focused on geometry. But we have this thing of like, you make the geometry valid and we guarantee that it will be exported to all these platforms. We know if it doesn't, if there is a bug, we'll fix it. If you can't fix it, you get all your money back. No questions asked. Because if you trust and spend all the time to use my platform to fix that geometry, I cannot fail you. I cannot make you look bad in the office because you took the risk. Again, that's my thing. If you take the risk, I want you to look the smartest person in that office. I want everyone to say, look at that person. They took the risk. Look at them. They're awesome. That's my job, to make you look amazing. And that's kind of packed with the users. one thing that we are building we don't talk about it much because it confuses the users or customers I should say we are focused on like hey rhino revit plugin build your geometry great but we're building a full infrastructure that makes it much easier to develop and write custom solutions that are both cloud native and they can run locally And that's something that has been a dream of mine because I want the next office, which starts being like the building environmental office, I want them to get from a start to the first customer to the first solution in 10 days instead of 10 months. Because we have done the whole infrastructure. It's just like how AWS, how Google Cloud Platform has done this. You can build cloud solutions because they have built the whole freaking infrastructure for you. It's just so valuable. it's hard to sell no one wants to pay for infrastructure so we can like consider the infrastructure even I i I would like to say from what from what you just you know I like the most from what you just said like you know I want to make the guy look uh really cool and smart in the office and I think that you know this is it's a very different incentive but like actually It is about the people more than it is the communities rather than, and I insist on that because you mentioned Grasshopper. What is Grasshopper? Why is it a good prototyping tool? It's essentially just another version of someone, sort of like David in this case, just sitting down and saying, no, I'm going to fix that problem. Genitive components is going haywire. Bentley can fix it. All of this can fix it. And he says, I'm going to fix that. I'm going to make something. that's going to be mine and it's going to be super cool and that the community that then created that software was like yeah and then even if you if you know if I look at speckle for example which is like you know I the way you're describing is essentially what speckle does for like big data and things like that so it's dimitri and and um and the team like that make the community the community essentially make the person we also believe that a lot that you know that's where we have like everything around it for like the host of like uh, infrastructural component that you say, you know, all of the connectors, all of the, you know, plugins, I think open source and accessible, so everybody can develop and further take them. You know, we have an API, you know, because we, I totally agree with you. The problem that we have is, you know, we've got tools, we've got singers, we've got sort of like, you know, powerful simulation tools, but if I have to take 10 days to prototype how I go from, you know, open studio to energy plus and from energy plus to IS. I mean, I'm like, you know, I've essentially, I need to have a dedicated team developing essentially prototypes. But I want to start from there and take you to, you know, this idea of integration of climate simulations, right? Like, and then, I mean, I couldn't think of any, someone has dedicated more time into this. I want to go into the side of like, no, how does that, really affect architecture? How do we meaningfully make more sustainable buildings? Because I remember in my work, in many high profile projects, doing this more for sort of like, you know, let's say, I wouldn't say greenwashing, but I would say at least sort of like, you know, saying that, you know, we tested this option, we tested this other option, and yeah, this is the best. But of course, we didn't really We didn't really design with performance, but we have really great tools. I've developed over the years for my work a lot of tools that essentially I don't think they were actually used as much because many times people just want to put a stamp on something. So how do we go forward? What we have to do in terms of integration and climate simulations to really meaningfully impact. Because right now we have like a heat wave and it's going to last like, you know, I don't know how long. And, you know, the world's not going to get any better, if you ask me. So what do you think? I think it's like what you just said in the last section. It's just, it's people. It's about people. I don't think tools, tools are part of the solution. But the main part of the solution is going to be people. I mean, I have this quote from the book Too Big to Know that says, when a network of professionals work together at its best, the smartest person in the room is the room itself. And those smart rooms are the ones who make the difference, make the change. And I saw LadyBug2's community like that, that you get these people in a place that they can fail fast, and it's okay to fail fast, and it's okay to recover, and people can talk to each other. And I mean, I actively, one of the things about LadyBug2, as I say, I think I did right, I actively watched people who were ruining this. and I blocked them from the community. I mean, not in a, like I blocked them, I talked to them, like some of them changed, like, but if they didn't want to be part of the community, if they wanted to make unnecessarily noise, they were out because that's the value. I think the future is also, again, My job, I think like, I mean, you know, your job was that, but I think like we need people like- Happy to be told by my job. No, no, I'm sorry. I don't want to say what you need to do. You should, you should. I think our job is to build the solutions to help that people who stand in the office and say, hey, we should do this better to be able to do it better in a reasonable amount of time and within an affordable that is easy to to show the benefits to students. Like one of the first time I wanted to use simulation or I didn't use it, but I had my first project that I wanted to use some, you know, natural ventilation and thought about, you know, those arrows of hope that you draw, like that kind of thing. And I remember the professor at the time who I showed my project to told me, what's the percent how much better is this and and he said it not in it I think it's a very good question but he said it in a very like uh diminishing way of like are you an engineer you're an architect why do you even talk about this but And I said, I don't know. And he said, if you can't calculate, why are you putting it here? Why are you shaping? And I spent like about six months after that trying to run one CFD simulation. I didn't even know it's called CFD. But that's my thing, right? Like, I don't want that person. I want the next person who is like me in the office or like you, like goes and says, we have to do better. I want them to have the right tools. And they say, okay, okay, let's do better. We give you this. How do you do it? Then you can tell them, this is how I have these tools. I can start. I don't need to build the infrastructure. This is our building. Let's do this three, four studies. And the more we do, again, like, as I said, we are in the, left side of things right now, like the geometry, metadata, unfortunately. But as we move to the other side, there is more insight. It's easier to understand. It's easier to share. It's easier to help people to make the right decision, to make the right action. And if there is one thing about this large language models that I like, I think, like, this ability to make things more accessible and easier for non-experts to understand the risks. But I think if you can use them right, they can help a lot on that side of the giving insight, helping people make the right decision, analyzing the results. So that's my thing. I think for the future to be better, again, My job is to build the right tools for the people who care. And if I empower them, I believe they will get things done and they will make the change in the right way. So you're making the room smarter. I'm trying to build the room and help people. Don't be shy in the room. I call it frictionless and with low inertia. Like you need an environment that's safe to fail. I think that's one of the other things I think like asking like one of the things that I think I did right about Ladybug. And Chris told me about it as something that he liked was, I would make mistakes when I were building and people come and like, oh, this is wrong, something. And we'd go, oh yeah, that's a bug, I'll fix it. You know, it was just, I wouldn't go back and say, no, you are wrong. I did. I want to, I want to, I want to catch that point. And I want to ask you something for me has bothered me for a very long time. So I remember when, you know, I was in a job interview, like, you know, after my, after UCL, after sort of my first sort of degree. And I had just built my own CFD software. So like I just, you know, wrote code. was coming from Stam's sort of physics engine for staining fluid. But I had developed a whole series of solvers. And then I had someone asking me in an interview, which was the lead of environmental team, saying, do you even understand the physics? And I was like, I think I do, to a point. Now I have to write the solver. So I understand to the point that I can do that. The reason I'm asking that is I very often see and hear climate. experts or like people who are like you know invested in a in a specific topic like cfd experts like I've been doing 20 years my cfd experts uh you know come to you and say oh but this is not accurate enough this is not uh you know it will never sort of be you'll never be able with cfd I've even heard that you know ask people asking when I was in my phd you know with cfd you can never understand the wind I'm like yeah but like How else would I understand wind comfort if I don't do CFD? I mean, there is no other way unless you expect me to build a wind tunnel for every design iteration I have. So how do you feel about, and this is, I want to touch upon the point of like accuracy, you know, really sort of like eyeing models towards insight, which was what you were talking about. Because I think, and then maybe to sort of like circle it, it's like there's cases where if you ask someone who has been running simulations for like 20 years. They don't even ring around the simulation to give you an answer. They say, hey, this building is too close, et cetera. So eventually, we will perhaps not even need simulation. Well, if you get, like, that's assuming, like, a lot of people have that, spend that time of, like, moving from data information to kind of wisdom, right? Knowledge and then wisdom. Like, you have the wisdom at that point. You just look at this. Like, I know. This is the thing. Okay. So, like, this is a very interesting topic. And I think most people think I don't believe in simplified methodologies or, like, this. Because, like, our brand is... You know, like we just go for validated engine. We go for like, you run the simulation, you're in on cloud to run faster. We don't simplify. And the reason we don't is not because I don't believe in it. I, the reason is because there is this Venn diagram that I have in mind, you know, simple solution, beautiful colors and uneducated user disaster. Again, I say it again. Simplified solution. It has limitations, right? Beautiful colors and uneducated user. Disaster. Disaster is in the middle for our listeners because we're not going to necessarily be seen. Yeah. I, yeah, if you know what are you doing, use as simplified as possible because you just want to get this, you're just like healthy. Okay, this doesn't look right. Let me change this. If you understand the limitations of that solution that you're using, I mean, for CFD, let's say like if any solution that gives me real time, and I've said that to several people I have talked to, I want the same thing on top to say percentage of, you know... I don't know, reliability. Because they say, our algorithm is good as long as your geometries are X and Y, because that's why we trained it. In this area, the percentage of like, it's probably like 50% correct. In this area, it's 90% correct. Because if I'm using your tool to only understand, you know, like the wind speed in this small geometry that they have, I want to know, like, is your tool trained well to do this? And I have to have the intuition of like, if the result is really wrong, I look at that, that's not right, you know? The problem is, how do you get to that balance? And I love to get to that balance. And that has been one of the questions that I have. And my thing is when we build the infrastructure in five to 10 years, now that's the time that, and I think things are getting better and faster. So maybe it's not five to 10 years anymore. but there will be a time that maybe you can have you can train a model to be that thing of like look at this thing as ah this is wrong this because this have you checked these five four things if there is someone like that next to you because again I can tell you like I i I don't go to many uh presentations or um like what's that the student things anymore school but I see here and there some presentations online, and sometimes it's just so painful to watch of someone, oh, we ran these parametric studies for distance and we made this decision. And the parametric is all about geometry and how things are changing. There is no window where it is. The thing that they're using to run the studies doesn't even do the shade calculation. So it's basically surface area calculation. Like the minimum surface to volume is going to win. And they're learning hundreds of this. And they make the wrong decision because it has been easy to run. So those are the things, again, like going to that Venn diagram. And I have seen a lot of these disasters. I call it, you know, what's it called? So before these things, people were at least as not confident. I'm trying to find a word not to be so aggressive, but it's kind of like, you know, dumb and confident at the same time. That's what happens. It's just like, a computer told me it's just like come on man like that's the thing like uneducated user that's the problem you know like we have to educate and that's kind of maybe that's why I'm very interested in like on on the people who care and they know and they spend the time because they are the ones to control the thing make sure the disaster doesn't happen and maybe when we have more and enough of those people we can have more um simplified solutions I can can I add one more last absolutely So the thing is about in this Venn diagram is one way is maybe like it's four. It's just like simplified solution, colorful, and then doing a lot of things. And then uneducated users. So you can simplify solution that does one thing and does it well. And because then because of that one thing and because you focus on one thing, you can have more people to use it because you just remove a lot of uncertainties from the solution. So you know it will be good for the single thing they're supposed to do. So the part that it just bothers. So if you, if you have a solution that's real time fast and only does UTCI, it is only trained on like certain type of buildings and just like tells that to people, I want everyone to use it, go use it. Great. But if you see tools that just, they're fast and they do everything and just like, it's just like, yeah, like being careful. So going back to the event diagram, we will be trying to do an infrared at least to overcome the problems, of course, like, you know, knowledge base, which is very important to overcome the problem of the uneducated user. Like, I do think that, you know, LLMs and the AI work is really important there from what I see, what theater is cooking. I'm really hopeful that we will be able to have a solution for that. And I also think that, you know, On the side of what sort of confidence level you have to answer questions, I do think absolutely what you're saying is critical. You need to be aware of what accuracy your model has. Every model is inaccurate. Every model is wrong. As Giovanni said in our last podcast, every model is wrong, but some are useful. So I think someone else has said that before. So I think this is exactly where we should be putting the effort to understand the balance, as you say, between being useful and being right. Because you can be right and not be useful. And for that, I want to bring one more thing before we have to move on. But I want to bring one more thing that for me is really important. I think one of the things that Ladybug has done for me is hugely important. is to give access to these simulation tools to a really much bigger environment than you know the let's say computational design teams of high-end offices and I think that was one of the problems that I always had you know I'm just doing this for only the 0.1% of architecture of the world. 99.99% of architecture is bulk designed, bulk developed buildings. If I've ever built a house, I've never managed to do that. I was thinking of building a house at some point in Greece, and I wanted to actually put a genetic algorithm with Ladybug, with sort of like, you know, kind of like view analysis, and just make that decide on the most simple geometries ever that I would be able to build, which is like two or three boxes, on where I should put this box. And, you know, what I'm trying to get with this is, if we want to empower not just, you know, the very clever and, you know, very well paid consultants of the high-end offices, and we want to empower everyone, we need to also consider that there will be uneducated users. There will be users that will not have any access to technology. They won't be able to do any Python scripting and whatever. They might not even have access to computing. So when we're talking about the 40% of the 99% of architecture, like, you know, then the 40% emissions that they generate, we have a serious problem to fix. Like, you know, we are architects, planners, we're like hoping actually architects and planners to develop sustainable solutions. It's our responsibility to give these tools at the hands of everyone, not just the elite. So I want to ask you, because I know, for example, you know, if someone gets a Or now if someone gets on pollination, they can probably afford to do simulation much cheaper than if they had to pay two people having a whole team of experts that can connect the different solutions or pay really expensive licenses on other things. So what do you think about this democratization of tools? Well, I mean, the premise of the idea, I'm all for it, but the way I see it is a bit different in a way of, and, you know, the things that you said is like very correct, like not everyone has access, you know, you don't want to have a solution for only like the small, uh fraction of people because they they have the privilege of you know having the you know the financial ability and all that and I think like one of the good things about ladybugs was being free in open source and everything which has been that and pollination they have a free path of using pollination when you still need a rhino license but yeah like we have the uh I mean from the web it's fully free we give everyone three hours and that but To me, the distribution of that expert knowledge to the masses should be done through a filter because giving access to someone who doesn't know anything about this just to be able to run the studies is not going to change to do much. it's like you're showing me a cd scan of like a something and telling me like it's very good just like I mean my wife is a doctor she just like sometimes she's an er doctor sometimes like she checks things on her phone and interestingly enough telling me like wow look at this like okay just what you know I mean now you know she has she educated me a little bit so I have a better idea a little bit better idea but that's my thing so it should be filtered and that's that's also something that I think with ladybug tools one of the things happened was that people that what you call computation or design tech people were the ones who picked it with something like human ui they rafted into something that more people can use and they're exposed it to them like you know again like a purpose-designed tools for people to use for certain things. For pollination, I said, like, we don't talk about it, but we have built the framework. And there is a concept of pollination apps that you can basically write a little bit of code on top of the whole infrastructure they've built. And everyone can use it. And technically it's free if you run like up to three hours. So yeah, I'm all for it. But I think, again, I'm building the infrastructure and I think the change comes from people who care and know about the topic and they have to package it and share it with the larger audience. And we are building, and I'm just so lucky and privileged right now to have the whole infrastructure provided by, we use Google Cloud, for us to be able to build on top of that. And the way to give back is for us to do the same for environmental building simulation. So everyone who wants that, the next generation, they don't have to start from scratch. Yeah, I love that. And maybe it's time to also... tell the world that we are building an infrared app for Polynesia. I love the Polynesian apps idea. And we've been cooking that for a bit. We'll see where it goes. But it's something we've been trying to do, which I think we should continue. And so maybe to try to come to a close, because I see the time is where we're getting. I would like to get some of the questions. uh before we go there oh we have messages and questions I didn't see them oh sorry yes we I'll I'll bring the questions up in a minute but uh so you know you again like maybe I wanted to before continuing I wanted to to say you know like talking about the city scan or the or the x-ray uh eventually there will be a tool that will you you won't need radiographers or like, you know, radiologists. I agree. And there are already tools, actually, they're trying to get there, but it's still, you need that human, you know, if you give it to me and then I, that tool tells me I can make a decision for that. Exactly. Which is essentially, and I think that's what I would like to ask you and, and sort of to wrap it up is the reason, you know, if we went 30 years ago, 40 years ago, we it from a command line perhaps but like the people who made the energy plan would say but this is all wrong the way you're doing it you you don't understand you know they just make little zones for your architecture problems this is not why we did whatever and and then they would probably laugh at the way we would be using it uh but eventually what happens and I always bring uh as an example when I'm talking about cfd this is my expertise right so I'd be quite a lot of time. And the thing that I always say is that, you know, you can't now understand, because you can't run fast, see, you can't understand a little bit by, you know, looking at the form, what will like normally happen. But like, if we take the example of solar radiation, the access to solar radiation tools allowed architects to have a better, much better understanding of like, you know, how exposed the building is to the sun, even like feeling of hey I need to shade these parts of these parts etc so going again from just I'm just able to run a simulation get you know beautiful colors and the performance map to understanding is something that takes time but eventually what I believe will happen which is also the question with lm ai and all that is you can move out of the I have to model it myself And I have to run the simulation and know the parameters. I have to acquire the domain knowledge. And if we could tell architects more about how to combine spaces, openings, HVAC equipment, and materials to make buildings more sustainable rather than teaching them how to run a simulation, it would probably be a better use of the time. So the idea is, perhaps for me, that we're going into this sort of like in abstracting a little bit from the models and getting into a little bit more sort of like the knowledge base and understanding essentially, you know, insight and how can you be action upon those insights? So like, you know, using them and coming back to, I say, what's your, perhaps your takeaway, you know, years of contribution to, you know, getting architects to access it. And also now with pollination, you're creating, as you say, the infrastructure, the infrastructure will be there soon. We'll be able to build apps. Perhaps someone will come to Pollination and build an LLM that's going to take all of Pollination's sort of simulations and be able to, you know, come up with insights. So all of these eventually will happen. And the question here is like, what's your, you know, what's your takeaway from like, no, what else can we contribute there to make this happen you have so many things let me let me start one we are not going to use all simulations of pollination because our terms of service doesn't let us do that like people don't worry no we're not going to use your data to train anything if we need something we will write it so that's not going to happen uh now large offices are going to cancel so let's clarify that I see a lot of parallels, and I think I'm in agreement. But as long as we have guardrails, I think things can get, you can ask more people to use it. I think we don't have much time, and I see there are questions. But to me, it's like driving, like the automated driving cars and everything. It's kind of like now exists too, but you still need to have a driving license. And going back to how we used to drive the car and how now you don't have to change the gear anymore because we have a system that understands good enough that basically for a typical driver, it can even change the gear and it just makes sense. But it comes through a lot of iterations and doing all that to make sure this automated gear change is not going to kill someone or it's not going to get your car into an accident. So to wrap up, I think You know, you learn a lot. There are different sides of me now. Like it's about running simulations about someone who develops a product. I think generalistically, I think one thing is not, and I didn't, it's not something that I came up with, but the thing that they say, we tend to overestimate what can be done in a year and underestimate what can be done in 10 years. It's very true. I've learned that natural experience of just how true that thing is. how important it is to have that tenacity of doing the same and have a team who has that tenacity to just wake up every morning, own the problem, do the same thing. We did a presentation at the hackathon recently about the hackathons we have done during the last 10 years. And just like we were doing the same exact, it's not the exact thing. We're trying to solve the same problem again and again at least for the last 10 years in the hackathons and add five, seven years before that. There's a compound effect of doing the same thing again and again, and have a team to be able to do that. Not following the shiny object, but at the same time, look around, see what's available, bring it in, but just like be committed to the problem that you're trying to solve. And then get your hands dirty, have a skin in the game. That's something that's very important. I'm looking for more individuals and offices with this kind of approach more than like this, you know, LinkedIn optimized approach of making noise and making a small noise at the time. So the algorithm picks you up that you're the best noise Like, we need more signal. We need more people who just committed to do things. And before we wrap up, I have to thank the whole LadyBugTools community. before I forget, like the ones who have done this, like came to the forum year after year, help answer the same kind of questions. And then the people who contributed, of course, Chris is the bigger one. Like people know Chris Mackey, co-founder of Ladybug Tools. Without him, many things wouldn't happen. And we have a list of people on the Ladybug Tools about page, people who contributed. Mingbo Peng, who works with us now, Antonello, Mikkel, Abraham, all the people who I don't want to forget. Antoine joined and did a lot. Theodore helped with all the butterfly stuff. Theo, you know, like, that's the thing I see. I think it's just like when you say, what do you see? It's kind of the same thing. I have been doing the same thing forever. And I hope at some point I can say, okay, it's done. Now I can just go do something else. Because it's kind of, you know, it's just like, it feels so, it's so boring at all from the perspective that I do, maybe because I really care, I deeply care about the problem. But then from outside, people can look and just like, how hard it can be? Why are we doing the same thing for this long to just help some people to run simulation? It should be solved already. It's a question that bothers me all the time as well. Looking back in time, I'm like, why the hell are we still stuck in the same problems? Availability, speed, insights and accessibility of knowledge. But we need to solve the problem, so we have to continue working on it. I'm not questioning that. I love what I do. I'm very grateful and I'm very privileged. I just got lucky, you know. I think you also put a lot of work. So on that, I think we're going to, you know, we're going to start with questions and comments. I'd actually start with a kudos from Alexander. I'll just like read it here. It says, Dear Mustafa, many thanks for facilitating the process of building simulation and making the user interface appealing to activists. So I do think that, you know, here's an example of like how, you know, the user interface does matter. It does matter that it is simplified. It does matter that it is essentially, you know, accessible. Do you wanna add something here or? Just say thank you. Again, it's not me. It's the whole Ladybug Tools community who made it. I got lucky to be the one who started it. So thank you. Thank you for using it, I guess. I think it's not just luck. And then I think Alexander probably has a more sort of deep down question. The potential to simplify the simulation process of double skin facades specifically in the future, so like adding a set predetermined so input output variables like simulation. Yeah, so one thing is just anything LadyBugTools decisions is Chris right now. I'm more of like a product. But I can comment on them. The reason I opened that with that is just like I'm not the one who makes the final decision here. Chris Mackey is. So the thing is, again, like the answer to Can you do that? Basically, if you're successful, you will be able to do that yourself. So the answer is yes. And with some help and support from us with building the infrastructure, you should be able to do that. You need basically a research that simplifies that because we are not a research institute to simplify it. And then if you have that, we have a simple way, we can help you with all the things of like, this is the interface, this is how you put it in, this is how you get access to very accurate engines if you want, or simplified engines if you want, because as I said, like the same kind of geometry schema is translatable to all these different formats. Excellent. I would probably go with Martin's question, which is, what is the best way to take into account thermal bridges of a building when making energy analysis? So if you're planning to integrate them in the future in some way with the calculations. Well, we right now have Therm. And again, this is one of those things when I talk about it, I'm a little bit uncomfortable because it has been a while since I ran a real energy simulation that I did all of that. But what we suggest people to do, again, like post it on the forum, Chris will answer this. I know, and I just don't, I'm just out of that. I'm a product manager right now. I can say it, but probably that's not the best answer. Ask Chris. What's it on the forum? Chris will answer that. Or maybe invite Chris at this. I was exactly thinking that we should at some point. Chris and Mingbo, they're the core scientists there. And Mikkel is our ratings expert now. They are the ones. I'm the product manager. Ask me product questions. So I'll ask you Kaushik's question, which is probably more related to a funder. So how did you find collaborators in Initiative and how did you make You know, revenue. Nice. Okay. How much time do we have? How much time do you need? No, we can talk about it. So the first question, collaborators with Ladybug was, it was very organic. People just came around and said, yeah, I have done this. What do you think? And then if it was useful, a lot of times we help people develop their own solutions. We never put them as part of the core libraries because it didn't feel something that we want to maintain and everyone can use. But I never walked around and tried to find someone. It was always like, you know, someone came and said, Hey, I have this thing. But we kept it open. We made it easy. We didn't make the process hard. And we are still supporting. Right now, we are emailing with a group of students from Iran who have developed some stuff that they think they want to add to the core libraries. But again, Chris is now the right person to talk to when it comes to Ladybug tools and what will be added there. How do you make revenue from an open source product? So there is no open source product. there is an open source project. I want to be clear about that in my mind. Your open source project is not your commercial product. And confusing the two can be disastrous if you are doing open source, just so you know. So that's why we are very clear. Ladybug Tools is our open source project, and Paul Nation is the umbrella for our commercial products and they are different because for a commercial product you're going to make profit you're going to make sure it's stable it solves the problem of the customers that you have it makes them happy it's very reliable you're not going to take high risks on your product you're going to take low risk or medium risk with like a high return then you have Ladybug Tools project, which is an open source project, and that's the place that you want to take risks. You want the innovation that happens. Again, of course, we have a productized version of Ladybug Tools, which is like the Grasshopper plugins that we keep and they're maintained and they're very stable. But for the core libraries, we do crazy things sometimes. They never make it to the production, right? And I think the idea of making money out of open source itself at this point I think it's not a healthy idea there's a lot of documentation out there I I'm a fan if you google for uh there is a paper uh I forgot the name of the writer but it's the title is something like your open source project is not your product there is one and the other one is there is no open source commercialization or something like that these two papers helped me a lot because we had this big question of like how we are when we start pollination is it going to be a version of ladybug tools which is paid or is it a different product I'm very happy that we decided that it's a different separate commercial product I don't want to make so I don't have to make money out of open so again that's why I said how much time we have because I can we spend a lot of time on that I can just go through all the different things out there of like oh you make this open source and you have a you know a premium version that people pay for look at this product look at mongodb look at this and then I can talk about all of that and their licenses and how but that's all that's a lot of details where I stand like the short answer is don't make your open source project your product and I know like there are the what is the name of the company the of the open now they I'm interested to see how it happens I think like that's a branding mistake because your value proposition shouldn't be being open source I could go on for ages about that. I don't remember how many times I've said to my students, open source doesn't mean free. That's different. If we were able as a tech society to adhere to essentially being able to be open source and and sort of like respect IP, we would be having a very different conversation about AIs, sort of like impact on IP, you know, there's like artists suing, you know, generative design. Essentially, we're talking about something that means a series of podcasts. Right, that's a series I would love to talk about because we spend a lot of time on that. I think it's a very important topic. It's an interesting topic. And my take is the easiest solution is just to separate them. I still am a big believer in open source. Of course, I just don't have to say it. The name of the company is after an open source project that we still maintain all these years and we do. But if you want to sell something, make sure you just separate it. not to hurt your open source project. I'm very concerned about open source projects get hurt because people want to make money out of them quickly and without thinking again of how to do it. Yeah, I couldn't agree more. But again, you wouldn't just take someone's book, print it, and then go sell it. Therefore, we could potentially think of a world where we don't just take someone's gold, put it somewhere else and then sell it. But I think we unfortunately have failed that part as a society, for now at least. Let's hope that in the future this will be- Yeah, I don't think that will happen in any future. There will always be people who take advantage of the things that are available. And always you should think about the worst case scenario and then plan for the best case scenario. That's my strategy. i said like I'm a dreamer I dream for the best everything works well but at the same time I'm not naive I know like I i I know the worst case scenario will happen as we do that and this is something I learned the hard way during the process so um yeah lovely um I'll just like bring up the last sort of like question it's more like a request to have an escape button? I think we have it. We do have it. The problem is Grasshopper is single-threaded, so the event doesn't get picked up by Grasshopper. It's hard to make it make it happen like I mean like we do a hack for the new version of fly that we made so there's a file that you can remove and that stops it uh when it's good but I think grasshopper 2 I haven't tried it but apparently it's that's multi-threaded so you should be able yeah I think there's a multitude of problems with you know having a whole cad software uh running and then on top of it a whole sort of like computations and uh platform running and then on top of it an integration of simulation so I'm pretty sure like you know if if there's a there's a lot of like right the escape button would be good but like you know it's it's also one of the reasons to me for me like I want to sort of close close this because we're like a little bit of a credit over time that uh you know uh I do believe that we need a new um you know a new sort of like uh many or not one many new sort of like uh ladybugs many new approaches like that uh because we need more communities that's exactly what we do here this is you know exactly what we're doing in this in this forum in this podcast to try to create a community out in french city we do feel we have something to offer as well to that uh and that is essentially why why are we doing this why we're having these talks we want to continue the conversation I want to thank everyone who joined online. And I want to thank you for joining, for being here and answering all these questions, but more for the incredible work. I do honestly believe, I'm not saying this lightly, I feel really, really proud to be sort of like having you here because I think the whole community owns you a lot and the whole Ladybug team. No one owes me anything, if that makes sense. No, I mean, you owe yourself to basically take what you learn and then try making things better like that. No one owes me anything. I always say like I did it, the project was open source, I'm glad it worked that, I'm glad everyone who made a business out of that, I love it. Just the best thing you can do for us is to just take this thing what you learned, be that person who makes the change. Because again, I'm back here, I don't, I'm not, you know, where I used to be. And the thing that gives me hope like, you know, is part of like, we are empowering people who care. make the real change in the industry and yeah like I've been I've been lucky to absolutely absolutely I think the industry needs that change because you know there is there's there's the hope I do have the hope that we will continue lots of others like you know face and theater and you and everybody will continue and there will be like a lot of new you know people I see the students coming out they're like super eager to to change the industry but I do think we also have a responsibility Because we are responsible. It's slow and steady. I'm on the slow and steady side of things. I think we are responsible. I think things take time. I know things are really bad. There's a wishful thinking of let's do it faster. But I'm going to do my best. I hope everyone do their best. But I'm on the side of things are going to take a long time. But I think overall, if you're in the right direction and we get lucky, Yeah, I think good things are, I think, like, I have a t-shirt that says, what? Naturally optimist. Something like that. I don't know. It's just kind of like, I'm a very pessimist optimist. Doesn't make me a realist, people. I'm a pessimist nihilist. I don't have a t-shirt, but they call me a pessimist nihilist. Yeah, it's from a Turkish brand. It's called Katan, C-O-T-T-O-N. Or how in Turkish? T-shirt right here. Instable, they call it chaos. We call it haunt. I think that's what it is. Yeah, right. Let's just keep... I like that, what you said. I'm glad companies like you exist. I'm glad other companies who care are around Yes, there will be noise. There will be companies who will come and just promise everything is solved with a click or something. And maybe, as we said, at some point things should get easier and we can do much better things. But yeah, I'm glad to see more people around who really care and who would really wake up every morning trying to do it, trying to get, and we are here to support them. And I think the same, like with your company, you are here to support them. That should give us a better future. That's all we can hope for, right? I think we should end with that hopeful before I start saying something. Thank you, Mustafa, for being here. Thank you, everyone. Tune in in about a month's time. We'll have the fourth iteration where we'll probably move out of the simulation and planning and go into more AI. Thank you for being here. Lovely to talk to you. Thanks everyone. And big kudos to Oana for organizing our CPO. Thank you everyone. Have a good day. Bye.

Note that transcript is automatically generated. Some names and factual information may not be accurately transcribed.