Building the Open Metaverse

Digital Twins in the Metaverse

Keith Bentley, Executive VP, CTO & Co-Founder of Bentley Systems joins Patrick Cozzi (Cesium) and Marc Petit (Epic Games) to talk about digital twins, interoperability and open source, and his unique career as a technologist and business leader.

Guests

Keith Bentley
Founder and CTO, Bentley Systems
Keith Bentley
Founder and CTO, Bentley Systems

Listen

Subscribe

Watch

Read

Announcer:

Today on Building the Open Metaverse.

Keith Bentley:

There isn't just one Metaverse, everybody's view of what their digital twin and what the Metaverse means is going to be assembled. So we had to think of something outside of what we were doing until and that we now call iTwin.js. It's our future for sure.

Announcer:

Welcome to Building the Open Metaverse, where technology experts discuss how the community is building the open Metaverse together, hosted by Patrick Cozzi from Cesium and Marc Petit from Epic Games.

Marc Petit:

Hello, everyone, and welcome to our show, Building the Open Metaverse, the podcast where technologists share their insight on how the community is building the open Metaverse together. I'm Marc Petit from Epic Games and my co-host is Patrick Cozzi from Cesium. Patrick, how are you today?

Patrick Cozzi:

Hi, Marc. Hi, everybody. I'm doing great. I have a lot of reasons to be happy. This is the first episode that we're recording where I'm at the new Cesium headquarters. We haven't moved in yet, but it will have a full studio for recording podcasts. Right now, I'm in our future boardroom and, for those watching on video, I have a printout of the Time Magazine metaverse article that Matthew Ball has kindly signed for us. I'm also excited because we have a very special guest with a great story to tell today.

Marc Petit:

Yeah. Our guest today is Keith Bentley, who's executive vice president, CTO and co-founder of Bentley Systems. Keith, welcome to the show.

Keith Bentley:

Thank you for having me. You have a very loyal audience from what I understand and I've seen some of your podcasts and some of your guests, so I hope I don't disappoint.

Marc Petit:

I doubt you will.

Patrick Cozzi:

So, Keith, as you know, we love to start off the podcast asking folks about their journey to the Metaverse and yours is super special to me because, one, you've done it in the greater Philadelphia area, which is dear to both of our hearts and, two, you saw the opportunity for graphics and the PC before many other folks did. We'd love to hear your journey.

Keith Bentley:

Well, my personal journey and Bentley Systems journey started even before the personal computer. Back in 1984 we were founded, and our first product ran on a, I guess at the time it was called a mini computer from a company called DEC, Digital Equipment Corporation. Actually, our software ran on one of their OEMs from a company called Intergraph, probably your audience may have been aware of.  Anyway, the way our first product worked is you connected the VAX to a terminal, called a like Tektronix 4014 terminal, and you sent graphics over ASCII using escape sequences. That's the way our first graphics program worked, you used an RS232 port to a terminal. 

Right around that time, the IBM PC/AT came out and it became pretty clear that the concepts of running software locally on a local computer had some advantages over the mini computer. A mini computer cost a million dollars usually. The terminals themselves cost maybe between five and $10,000 and a personal computer was 15, maybe $20,000. So price wise it was a huge advantage but even more so was the advantage that you could have some autonomy. You were in control of your own destiny, your computer was yours. That's why it's a personal computer. 

Then chapter two of Bentley System story is that the CAD transitioned from being computer aided drafting to computer aided design and the purpose of a computer session would be not only to create pieces of paper in the main–still people generated paper–but to create an electronic model of some asset. Computer aided design gave way over time to something called BIM. BIM I think stands for, the B stands for building but IM stands for information modeling. The concept is that instead of just doing some modeling where something looks like its physical properties, you also model the properties about why it exists. 

Patrick Cozzi:

Wow. I love the parallel to how game changing the PC was and, potentially, how game changing the metaverses will be today. We certainly feel the same way. I love the origin story and I love that you were doing graphics in the greater Philadelphia area in the 80s. It was probably a dozen people, including academics, at that time so I think it's been very cool.

Keith Bentley:

Definitely true. We started just down the street from where you currently are in Center City, Philadelphia. Your office today probably is twice the size of the one that we had there. Bentley Systems began, I started it, then I started hiring my brothers. At one point I had all four of my brothers working for Bentley Systems so it's kind of weird story. Bentley Systems is not a typical story but one of the aspects that people tend to focus on is the fact that we were built by five brothers.

Marc Petit:

If I remember well there was, at the beginning of the PC era, a lot of CAD companies and only a few of them are still like 30 odd–

Keith Bentley:

Right, it's true.

Marc Petit:

Why don't you explain the longevity of Bentley over 40 years. What was the secret there?

Keith Bentley:

I don't know if there's any secret, but you're right. When we got started, the personal computer was new. Everybody saw that. I think the only person who really wasn't convinced that the personal computer was going to amount to anything was the president of DEC, Ken Olson. He was a really smart guy and gets a lot of credit for a lot of really good stuff but the only thing he's ever remembered for is his quote that, "Why does anybody want a computer at home?" Anyway, there were a lot of people who saw that the personal computer was going to be a game changer. Literally, there were 30 companies called CAD companies back then and yes, Bentley, I think we have a few virtues that have made our journey one that's been sustainable. 

One thing was that we were profitable before we got started. Every month of our existence we've made money and that's fairly rare. We didn't have to do lots of chasing of targets that were unattainable. We lived within our means for a while. Also, one of the things that stood us in good stead is I mentioned a company Intergraph. They were the CAD vendor of the day. They were a billion dollar company. They had a bunch of users that were initially our users because we sold them software that's compatible. 

Patrick Cozzi:

Usually on the podcast, Keith, we always ask for advice. I'm glad you're already giving us some good business and recruiting advice.

Keith Bentley:

People sometimes come to me and say, what's it like working with your brothers? I always say, well, I advise against it but it's worked out pretty well for me. We've managed to work together for a long time. There's only two of us left at Bentley but that's a different story for our company that’s somewhat unique.

Marc Petit:

Bentley is an incredibly important company. Some of the major pieces of infrastructure, some of the things we rely upon in our daily lives around roads and large scale infrastructures are being modeled, created and operated using Bentley softwares. It's not a household brand but I think you guys have had an impact on a number of critical components of our daily lives. Thank you for that.

Keith Bentley:

Don't thank me, I never brag about Bentley Systems. I only ever brag about what our users do with our products, which is pretty incredible. And you're right, some of the largest projects in the world are physical infrastructure. 

Marc Petit:

Well, let's jump into this. You said phase two of chapter two of the Bentley story was BIM and chapter three is going to be digital twin and the Metaverse. You launched iTwin, a digital twin platform a few years back already. Can you speak to your motivation and your vision behind the product line and how you see the future of Bentley there?

Keith Bentley:

The future of Bentley is around iTwins and digital twins. That's Keith's perspective on it, there can be lots of permutations on that theme, but I can tell you my commitment and my reason why I feel so strongly about it was influenced by this guy I met about five years ago who showed me this product. Actually, he didn't show it to me, it was shown to me by a programmer in France who worked for Bentley Systems. He said, look at this really great stuff and he showed me Cesium. I looked at what Patrick and his team had done with Cesium and I was just floored like, wow, that's in a browser. Wow. Just imagine all of the models that have been created with our tools. Suppose you were actually able to use it inside a browser combined with lots of other data sources.

One of the great strengths of Cesium is it does take data from many different formats and combines it together. I said, that's what our users need. That's what the world needs, not just our users. The world needs to be able to repurpose information in lots of creative ways. I was so amazed not only about what it did but by the fact that Cesium was open source. It was brought to me, to my attention, by someone who worked for Bentley Systems and was using it for something I didn't even understand. I didn't know why Cesium was relevant. So I started thinking about, well, the future's going to be a world where data gets shared way more so than it does today.

Part of the problem today is BIM is a really great concept but you need a BIM tool to read a BIM file, largely. There's ways you can exchange it in IRC and so forth, but usually the data is modeled by something that you have to learn, a tool you have to be really good at to be able to do wonderful things, to understand what's where. You need a license to it. You need training. You need the right version. All those things combined just means that what people do today is they exchange information in PDF. What a terrible outcome. 

If we're modeling all this stuff and we have all kinds of wonderful–BIM stands for information modeling but if you've exported to a PDF, you've just lost all the value in it. So, what Cesium brought to me was, look, information should be exchanged not in some dumbed down format but in a format where it could be live. I started thinking about, well, how could that possibly happen with our current generation of BIM products, Bentley's or anyone else's at the time?

Marc Petit:

For those of you who don't have a camera looking at Patrick blushing, it’s actually real fun for me.

Keith Bentley:

I've given him a lot of credit and he deserves a lot of credit. I literally wouldn't be here talking about the metaverse if I hadn't seen Cesium…five years ago? How long ago was that, Patrick? I don't know.

Patrick Cozzi:

It was about five years ago.

Marc Petit:

That's an incredible story, Keith, thank you for sharing this. We’ll keep talking about interoperability then.

Keith Bentley:

Okay, so interoperability, I don't think you can have a digital twin or a metaverse of something that you buy from a single vendor. It just seems so remote to me that anybody's going to constrain the types of information that they can model in their digital twin to being only available from a vendor or even a group of cooperating vendors. It's going to be assembled from parts that people will build. Now, there's going to be people that are experts in doing this so not every digital twin user is going to build their own digital twin, but I believe there will be companies that are really good at pulling these pieces together. How can that possibly work? Well, in the world of open source, there are just so many plugable tools that all build on the same technology stack that I think it is conceivable that you could assemble a digital twin from parts, not if you have to go school for every part. But things like node.JS, I love it. It really does make the concept of open parts plugable work really well together. JavaScript, game engines, these are all things that now are available and not price prohibitive, not so hard to envision, oh, well I would have to employ people that are experts in each one of them. I can be an expert at something that I really spend very little time with because I'm familiar with the techniques it uses. Openness is an absolute prerequisite for the success in infrastructure. I can't speak for Metaverse outside of the one–

Marc Petit:

That's such a radical departure from what we've heard from CAD companies over the past 40 years. So I want to congratulate you, but acknowledge that we have a lot of people coming to this podcast. People from the CAD world usually are not that blunt about it.

Keith Bentley:

I can tell you there is no zealot like a convert and I'm a convert because I was one of those people who said, free software? I pay programmers, how can I make free software? That was me. I said that. Now I say, well look, yes, we're in this to make money. Don't get me wrong, I'm not in this for charity, but I believe the sum of the parts is going to be so much greater when it can be plugged together. 
People talk about vendor lock-in. That's one of the criticisms that people have against Bentley, against our competitors who will go nameless. Vendor lock-in's real. If I were out there, I'd say, hey, if I use your tool I don't want to have to have a subscription to your service to use my own data. That's insane. It's absolutely insane. Even if I promise you, oh, we're great guys, we will never rip you off. You might get mad at me for other reasons and you need to have some leverage. Hey Bentley, we don't like your policy. We're going to leave. If your data is locked up in our format, how scared are we or our competitors of that happening? But when we give you the tools to use your data without anything from us, now you have the ability to leave us. Hey, that's not a good thing for us but what I think is we can prove that there's a lot of reasons why you should pay us to provide you solutions. You can get tools but we give you solutions and I think we're going to win. I think if our platform takes off, our iTwin platform based on open standards like glTF and all the other reasons why people will use lots of things that aren't from Bentley, we will be a part of an ecosystem that's going to be way bigger.

It’s not the first time I said this, so I've made that case to our users in front of our users. I've made that case to our product managers, none of whom are particularly happy about the prospect that someone can use our platform without paying us but I tell them the ecosystem is the thing. It's yet to be proven that I'm right but I'm still pretty confident. Like I said, I'm more excited now than I was at the beginning of the CAD era. So vendor lock-in, if any vendor tells you that the only way you're going to be able to use your data is with a subscription to their cloud, you should slam the door in their face. That's what I say.

Marc Petit:

That's a fair point. I mean, it's a pretty obvious thing, even though we know the practices tend to–

Keith Bentley:

You brought up an important point. A lot of the infrastructure in the world that is modeled in our software is owned by government agencies, so they have a lot of concerns. Governments today are pretty concerned about, oh, if I'm an organization in China, do I want to have software from a company in the United States or do I want to have German software running on an American desk? Those kinds of combinations, things…. These owners of infrastructure worry a lot about the 40 year, 50 year life cycle of the data that they're going to, digital twins is going to become something pretty valuable to their operations. Suppose they have their disaster recovery built on top of their digital twin solution. Are they going to listen to a vendor like Bentley tell them that, okay, here's how it's going to cost you for the next 40 years? We wouldn't be able to give them an answer to that if they wanted to. Anyway, they need flexibility, they want flexibility and they need openness and it's possible. Patrick proved to me that it can work well.

Marc Petit:

Absolutely. So Patrick and I are part of an adventure called the Metaverse Standards Forum and the idea is to learn from the USD open source library, developed by Pixar and taken further by Nvidia and propose the standardization of how you could compose reducing, which I think is exactly what you are describing, as parts of digital twins. I think there is a lot of convergence between those ideas and now is the right time to lay that foundation that we can assemble virtual worlds or digital twins from multiple sources, arguably in multiple formats, into a single representation. It's a very, very timely conversation and timely topic.

Keith Bentley:

When you talk about the ways that people will assemble information from multiple sources, sometimes we like to point out that a digital twin of an infrastructure asset is going to have three main components.

One is going to be data that is observed. In Reality, people fly a drone, get a reality mesh, get a point cloud. That's a real important part of understanding what you have out there. Now, drones and LiDAR is only of what you can see. Then there's below the ground and there's ways that you can sense that. So that's the operational data, what is there. Then there's sensors, you connect ways to read current values of how many cars are passing over this road, how much strain is there in this bridge. Connect a sensor, put a battery on it, put a transmitter to it. Now you can have this thing that has a basis in fact because you've observed it, connected with what's actually happening. The current values but you can also look at the past values.

That's all really cool. Sometimes people think of that as a digital twin, but then think about, well, what happens when you're going to propose a widening of that bridge or something like that. There's an engineering project that goes on. That's where things get a little bit more complicated in our world than some worlds because what you're actually doing is modeling the future and the future is non-linear. There might be five permutations on the way the changes are going to be made. You might want to model the construction sequence and how do you get from here to there? Those future states of the digital twin are just as important. Sometimes the future states of the digital twins are things you don't want to have happen, you model disaster scenarios. So the engineering model plus the reality model plus the real time data, they're all really important parts.

The complication of how are you going to get all this stuff to make sense to someone for a specific purpose is a very hard problem that I don't think any one vendor is going to be able to solve. That's why I think open standards for interchange, like hey, we don't do GIS, but we recognize that a lot of people have a lot of real important stuff that they've saved in their GIS system or their GIS system has a current state of a lot of stuff. We need to be able to, through things like Cesium, get the current representation of that, draw it in ways that our users can understand it, send it off to things like Unreal to be modeled in real time and do a real nice job of simulating the way things actually look in the real world. 

Anyway, engineering plus operational data plus reality data is a hard problem..if 20 years from now we'll look back and say, my goodness, can you believe all they did? What we’re doing today. It'll be really, really important.

Patrick Cozzi:

Keith, I love the open ecosystem philosophy and the idea that look, we're going to grow the pie so big that we can each be really good at a part of it. Interop I think was a really great strategy and I think that's what led to the success of the internet as you step back and you kind of had silos of AOL or Prodigy and then adobe and internet came with all the interoperability. So your philosophy is great and Marc previously mentioned-

Keith Bentley:

Well, I don't know if it's great or not, but it's our philosophy anyway. Time will tell.

Patrick Cozzi:

Well, this Metaverse standards forum that Marc and I have been involved in just launched a few months ago and it's up to like 1,700 or 1,800 companies. The idea is, hey, this metaverse thing is real and can be big and we need to work together to facilitate and enable that interoperability.

Keith Bentley:

Well, I'm cheering for you guys because I really think the more success the standards formats have, the more data there is for all of us and those of us who are working on trying to make it all work together, it'll help us help our users and help the world. I have an infrastructure view of the world and I'm worried that we need to solve big problems, climate change, power generation, all those things. If somebody doesn't solve them, forget making money, we're going to not exist.

Marc Petit:

But talking about scale, you led to solve some incredibly difficult problems around, for example, your context capture technology which can scan large scale places. How far do you think we can take that capture process and automate it? Are you looking at machine learning technology we've seen that a number of times. What's your prediction on that?

Keith Bentley:

I'm always inspired by examples of machine learning. Cars driving themselves and we have some examples of just trying to reverse engineer paper that has images of something that was modeled either in 3D or 2D and try to regenerate. It works incredibly well. Algorithms are not the right answer, it's inferences that's going to be the long term solution to those kinds of things. So I do believe context capture plus a lot of machine learning–there was a demo I watched just yesterday about trying to get a digital representation of a road surface and detect cracks. People do that today. They go out and they look at it and they mark and okay, there's a crack there, but you can do it in real time. You could fly a drone, run a machine learning algorithm on the images from the drone, generate a 3D mesh and then determine that's where we need to send a crew tomorrow.

It's all possible today, or at least not all, but a lot of it's possible today. I really think we're at the very beginning of machine learning applied to engineering type problems. I tell our people, I think that's going to be Bentley's future for sure. I don't think we're going to be able to do it all or, even in the crack detection world, you need experts. You need people that are trained at making the models right. So the intellectual property, who owns that? I don't know. I think there's going to be some business problems around trying to make sure that the technology, people who add value get paid for it but that's a nice problem to have.

Marc Petit:

What's your vision about making sure, because you mentioned there's talk we have a component like a reality captured component and the parametric model, how do you keep those in sync? Make sure that your nuclear plan, the actual nuclear plan actually matches the digital model that you're going to use to do predictions?

Keith Bentley:

That's the essence of our iTwin platform. If there's one thing I think it does that we've never done before, and I don't think anyone else has addressed, is change management. Think about the way software engineers work. We use GIT. GIT's purpose in life is to keep track of who did what, what they changed, you can roll the clock back. As I mentioned, the engineering projects, they're always dealing with a future state but it's not one future state. There's one team working on one part. It moves at one pace. They branch and merge just like we do in software. So the essence of our iTwin platform and something we call iModels is change tracking, change merging, and it basically is GIT for infrastructure models. That's the essence of how to keep track of who changed what and what's the current state of something.

But there's not just one. I just want to point out there's not one state of the digital twin when you're talking about an engineering model. For example, sometimes people model what will happen after some catastrophic event and they want to save that. They don't want anybody to use that in real time. That's not the current model but, if that disaster happens, they want to be able to bring it up instantly. Look at what has all the planning, how should it look and who should do what. Planning for what might happen is part of the engineering problem, storing that information in a way that you can get to it quickly, save it in the cloud. That's not an easy problem, but that's what we've worked on.

Marc Petit:

It's first time I hear this analogy with the software development process and I think it's very relevant.

Keith Bentley:

Another thing that's very analogous in software is we will have different versions of our products and sometimes we have to go back and make a change to an old version. Engineering projects, sometimes there's the plant currently being operated and someone's got to go back out in the field and replace a leaking pump or something like that. You’ve got to get the state of the model as of when it was actually built. That's not the state of the model where it's currently proposed. The plant is undergoing some transition but you need to go patch, that's what we call it in software. People do that in the real world. How do they do that today? They take sheets of PDFs out on iPads and how do you know you got the right one? It's a mess. It's really is.

Furthermore, one of the wonderful things about GIT for software developers is I have 5,000 source files and I changed three, what did I do? Finding change and how does that happen in the world of engineering today. I kid you not, people hold PDF files up to the light and try to find what's looks different. They don't literally do that but it's all about trying to track change after the fact rather than keeping track of change and storing change. That's what iModels do. We store change. We don't store the current state, we store deltas. That's a brand new thing and a new concept for Bentley Systems. It isn't the way the world in the main works today.

Patrick Cozzi:

Taking that technology and that understanding from one domain, from the software domain, and then applying it to the infrastructure I think is fantastic.

Marc Petit:

And you said you've built iModel on top of an open source?

Keith Bentley:

Yes. So my favorite software product, number two is Cesium, but number one, it must be said that really opened my eyes, is SQLite. SQLite, everything about iModels is built on the world's most open database. It's actually the world's most used database. There’s more data stored in SQLite than all other data formats, all other relational data formats. It's a relational database in a file. It's written by mostly two guys. They put the whole thing in the public domain and it works. It's a library. It's a C, not even C plus plus, it's a C library that does full SQL, full transactions and it works amazingly. If you don't know anything about SQLite, you don't have to understand how it works but I guarantee you every one of your phones is running at least five copies of SQLite, including your mail application. That's the way all data is stored on mobile devices today. We use SQLite as the persistence format for iModels and we track changes through a tool in SQLite for keeping track of which rows and columns were changed in a transaction. That's the way change tracking is done.

It's a really cool technology. I love SQLite. I'm a geek's geek and I tell people, if you want to learn how to program computers, if you're just getting started in software and you'd like an example of something that works really well, look at the SQLite source code. It's one of the best documents, the best tested, most reliable source code. It's not a large code but you would think that, oh, how many lines of code is there in a SQLite or in a SQL engine as a query planner and a transition? It's not that much. You can learn a lot about how to write good software by studying that. I do. I look at it a lot. Patrick, I love Cesium, but I love SQLite even more.

Patrick Cozzi:

I feel really good about Cesium coming in number two.

Keith Bentley:

Okay, I didn't want to insult you.

Patrick Cozzi:

There's one other type of software that I wanted to ask you about, Keith, and that's game engines. So on this call you mentioned painter’s algorithm. You're probably one of the few people I know who has implemented the painter's algorithm.

Keith Bentley:

I'm proud of it. It's just the way we had to do it.

Patrick Cozzi:

Through and through you and your brothers are rendering programmers, right? Graphics programmers who rolled your own graphics engines early on and then you've at least, to some capacity, adopted game engines for non games. I think these are a really, really exciting area to talk about.

Keith Bentley:

Yes. No matter what we would do, Bentley Systems, our mission is not being the experts at rendering. If you look at what game engines now do, it's just absolutely amazing to me. I remember looking at them some years ago and A, being impressed but thinking, okay, well we have all kinds of different problems and sometimes in CAD we don't want it to look real. We have views where we show edges and it's not like realism is our primary goal but all the time our users want something that looks real. The more real it is, the better they can use to give presentations for example.

We'll never be the rendering kings. We don't even want to be. In fact, in our new open environment, I'm hoping people will use lots of different visualization techniques, Cesium for web form factors, game engines not only for physical devices but it's going to look so much better, perform better. It's really, really wonderful technology. Frankly, they make a lot of money selling games so they don't necessarily need to price it out of reach for people to use for non-game applications.

Marc Petit:

Some people do that.

Keith Bentley:

We're relatively new at that. Our use of game engines, of the game engine technology stack, is nascent but I see lots of opportunity for the combination of open standards plus a real quality rendering system that can run both in the cloud and on a mobile device. I think that VR is going to be used way more when it works and is priced the way that people can use it on everyday projects. 

We haven't done as much with game engines as I know that we will, but I really feel it's the right answer for a lot of problems that I don't want to have the Bentley Systems people working on. It's one of those areas where if we combine efforts, both us and other people that will use that technology can create a solution that neither one of us would be able to write on our own.

Patrick Cozzi:

We had the same observation with geospatial and applying that to game engines. When we built the Cesium for Unreal plugin, overnight we made 30 years progress when I saw this because they used the rendering engine and everything there–and I did consider myself a hardcore rendering person.

Keith Bentley:

You teach it, right?

Patrick Cozzi:

I'm like, wait a minute. I can't keep up with this.

Keith Bentley:

And next year's game engines will be better than this year. It's a march that innovation is always going to happen. I feel like that's the kind of thing we can combine with and not compete with.

Marc Petit:

There are even some open source alternatives now.

Keith Bentley:

The technology stack that you would use to assemble a solution today is–imagine 10 years ago talking about using a game engine. How hard would that have been? The game engine vendors didn't think of it being used outside of the game form factor so they didn't put any work into making it possible for people to use it for other things.

Marc Petit:

I think JavaScript is a solid computing platform.

Keith Bentley:

Yeah.

Marc Petit:

10 years ago would there have been a-

Keith Bentley:

You would've given up on that, right? You would've said, eh, I got to write it in C or one of these compiled languages. I'm a huge JavaScript bigot now Patrick convinced me of that too. I like Typescript. I don't know if, Patrick, you're a Typescript guy.

Patrick Cozzi:

Not yet.

Patrick Cozzi:

Keith, we wanted to start wrapping things up with two questions. So first there's probably a short list of folks that if asked who I would want to model my career after, I would name you as one of them. You have just incredible technical depth. I think you've probably written more code than me recently, which I'm quite jealous of, and you've built an amazing company and business. So, for myself but also for our audience, what career advice would you give for folks that want to be in the field?

Keith Bentley:

I'm a programmer Patrick. I love programming. The only thing in my life that I'm any good at is writing software. I was lucky enough to get into this business at a time when there were huge opportunities and I combined with some really smart people and my brothers. I wouldn't give advice to anybody other than that some people say, oh, pick a career you're good at and you'll never work a day in your life. I say I've worked every day in my life but I've enjoyed them all. What else can you look for in a career but to have done something you like to do? I've happened to also pick something where it's been financially successful for me. I've been pretty happy with the outcome from that but I'd be equally happy if I was able to do all this all over again and we didn't make a billion dollar company. That wasn't the goal so don't ask me for advice. I say get lucky.

Patrick Cozzi:

I think that's part of the strategy.

Keith Bentley:

I did get some advice when we were starting and that was always hire the smartest people you can. Maybe that doesn't apply in every business, but in software there's a category of programmers who are...they see in code. You see a problem, you see a simple solution for it. There's other people that might be really smart at other things but they see a problem and they try to make it hard. Software just doesn't come naturally to them. So, in software, I think hiring the smartest people is always good advice. Making things simple is always good advice. Those are two things that I would say I know of.

Marc Petit:

I tend to agree with this. The productivity you get from those smart people, it takes sometimes entire teams to replace them.

Keith Bentley:

Yeah.

Marc Petit:

The only thing you fear is that, afterwards.

Keith Bentley:

There is that, yeah. You worry too much about dependence on one individual. There was a period of time in Bentley Systems’ past where we thought, well you have N dollars, hire more people. You don't have to get the top. That just never worked out. Smart people like to work with smart people, work for smart people and you really have to make a team that enjoys what they do and are really good at it and they're not cheap.

Marc Petit:

Our last question, although I think you've covered it already, is there a person, institution, or organization that you would like to give a shout out to today that is not Patrick Cozzi?

Keith Bentley:

No, no. I won't blow Patrick's ego up anymore. To me, Bentley Systems is where it is because we had really, really good users. People did and do incredible things with our software. They've inspired us. When I go home at night, I think my goodness, I have a list of problems that one of our users might have. I think to myself, man, they trusted us. They thought that our tools were going to solve this and if it's not working, I really want to help them. Then I'd look at what they do with it. I just think, my gosh, it's really cool that I could play a little part in that.

Marc Petit:

Amazing shout out. Thank you so much. Keith, you created one of these big important CAD companies 40 years ago with your brothers. What incredible achievements that you've been doing. Now you have this new fun religion of open source. Diving deep into the Metaverse and digital twins was amazing for you to share this with us. It's been a real pleasure to have you on the show. Thank you. Thank you very much.

Keith Bentley:

Thank you very much for having me. I believe in both of your missions, so I'm cheering for you guys. Like I said, I've listened to several of your podcasts and you've had really good guests, so hopefully you'll be able to find something useful in what I said.

Marc Petit:

I think it's great that we get to expose people like you. As we said, the Metaverse has a lot of hard problems to solve. Some of those are being solved by you guys in Bentley. So once again, Keith, thank you very much for being here. Patrick, next guest we'll try to make you blush a little bit less. I'm not sure we can do this.

Keith Bentley:

Sorry, Patrick. All right. Thank you very much for having me.

Marc Petit:

Yeah. Thank you very much to our audience. We hear lots of good things about the podcast. Keep on hitting us on social. Let us know what you think. Let us know who you want to hear from and we'll be back with another episode soon. Thank you, Patrick. Thank you Keith, thank you everybody.

Patrick Cozzi:

Thanks, Keith. Thanks, Marc. Thanks everybody.