Roopinder: 0:18

Hello everyone and welcome to the Masters of Technology Happy Hour, where once a week, I have a drink with someone I meet in the course of business, but someone I'd like to get to know better as a person.

Narrator: 0:29

On this episode. Episode 7, rupinder is going to be interviewing Scott Harris and they jump right into their conversation.

Roopinder: 0:37

You are part of the history of CAD, what you've done. You seem to be a humble guy, but let me do the bragging for you. You've done quite a bit for CAD in your time. Scott, really I'm going to make you blush here, but really you're behind the most important, popular professional MCAD software ever and the only professional MCAD software of this century. You've been behind both of those things. So I do want to tell that story, because that story does not get out much. I mean it's not well known. I mean people that know you. Certainly CAD insiders know this right. But if you ask people like who invented SolidWorks or Onshape, you know who. That is right.

Scott: 1:21

Sure, but to be honest with you, it is a team effort. There's a lot that goes on. You know the sum is way bigger than the individual parts. It really is a team effort and I think you know people still talk about you.

Roopinder: 1:36

I was at the other day. They mentioned they've been in CAD for a long time. So I said what was your first CAD program? They go SolidWorks. I said what made you get onto SolidWorks? Was it just because everybody else was using it and he goes? No, I got it. It was the UI. It's like all of a sudden I found out those people got me and I think a lot of that is with SolidWorks. There were mechanical engineers actually making the product. This was a bunch of coders, developers, trying to figure out how you make parts. No, no, this is how engineers thought, and it was engineer to engineer software.

Scott: 2:10

It's interesting because where I cut my teeth on CAD is on development. I started out as an engineer at Pratt Whitney and that's where I kind of learned about CAD. I was an aerodynamicist. I was designing nacelles and working on turbines and compressors and all the inlets and duct work within a jet engine. Everything was done almost by hand. We used to lay out big pieces of graph paper on the wall and use these big plastic chip splines to draw curves. We used to map out hard points and then measure off the points and then use that data entry into Navier-Stokes equations, because Navier-Stokes is everything we were solving dearest flow problems. And I was introduced to CAD by a guy I carpooled with. I'd seen CAD before, I'd seen 2D CAD and I thought it was terrible boring. I didn't care. He was part of the solid modeling group at Pratt Whitney, which was really interesting because that's back in the days when a company like Pratt developed their own CAD tools on top of various solid modeling kernels. So he showed me this stuff and all of a sudden I saw this as a design. It clicked 3D, solid modeling. I get it. It was a language that I really could understand and I got very excited by it and I started developing tools at Pratt to do flow analysis using the CAD to generate data for the Navier-Stokes so I could model out an inlet or a nozzle in the cell and then write a macro to dump out the points that were necessary to discretize data for the Navier-Stokes equations. And that's how I got into it. That was my entry point into the world.

Scott: 3:48

I went to a company called Computerv ision, which you may remember. I ended up running the geometric modeling group there. One of the biggest problems they had, besides being on antiquated hardware, antiquated software was a program written by mathematicians for mathematicians. They did a lot of interesting features, but what they didn't understand very well is engineering workflows and how mechanical engineers really fought. So it was very, very difficult to use. It was a crazy design language and I spent 10 years there. That's where I met Hirschtick.

Roopinder: 4:21

To your credit, you understood what engineers needed to do there. That's where I met Hirschtick. To your credit, you understood what engineers needed to do and here was a bunch of column eggheads telling you what you ought to be doing with their software. I've always said software designed by software people is easy to use if you know how to use it. But to everyone else, including engineers engineers some of the smartest people on earth right. Engineers can't use their software.

Scott: 4:44

To give you know I can't bag on CV too much because they sort of came from a 2D. You know they were drafting products and they were making these forays into 3D and into solid modeling. It was very incremental and they had a lot of baggage and so what they couldn't do was reinvent themselves. In fact Sam Geisberg actually worked at Computerv ision. In fact I kind of took his job. He left and started PTC and I came in as a young guy. I eventually was in the same group as Sam. He went on and started PTC. PTC was written by mathematicians but they had much better support for workflows. It was still pretty rough, you know. To change a feature you had to do a database edit. To keep track of your features you had to write down on a piece of paper the feature ID and the database ID or whatever and the value. So there was a lot more to workflow than just the editing part of it.

Roopinder: 5:41

Knowing you, were you that guy? That's like you really ought to do it this way. And then they got rid of you because you were. Well, it was interesting.

Scott: 5:48

Computerv ision was imploding. I was there for 10 years. It went from 12,000 people to 1,200. So everybody was gone. You know, basically this is when I first started talking with Hirschtick. We were both very frustrated by the direction of the company. Jon can explain. You can ask him when you're in for him. I was working a lot with customers and trying to get their models built and it was a very interesting exercise to go through. I spent years doing that with people and then developing to try to fill their needs. But there was this giant wave coming right. There was all these PCs. Everybody understood Microsoft UIs and the foundation class libraries like File New, File Open, Save. Those are just standard things but they weren't standard in CAD back in the day, back in 1994, they weren't standard. The idea was there are these modeling kernels out there like Parasol and ACES. There was a bunch of other ones. I looked at probably five or six different modeling kernels at the time to evaluate which ones were the best for SOLIDWORKS.

Roopinder: 6:57

This was 94. So you're just one year away from where SOLIDWORKS actually came out right.

Scott: 7:04

That's right. Yeah, it's amazing that SOLIDWORKS actually came out. Right, that's right. Yeah, it's amazing that SOLIDWORKS 95 came out. We started we were at the beginning of 94, right, and SOLIDWORKS came out at the end of 95. It was very fast pace, but it was an interesting ride. The trick was figuring out what functions to put in right. There's a thousand things you have to do and the hard part is figuring out what you're not going to do first. Right, because you want to do everything right.

Roopinder: 7:28

Is that because you couldn't fit the whole program that Computer vision could do? You couldn't fit that all into a PC.

Scott: 7:35

Not at all. It was just time to be able to develop all the different features right, every feature. Even though we're using a modeling kernel and a constraint solver and all the Microsoft Foundation class stuff and structured storage you know, all the database stuff Even though we're using all those libraries, we still had to write.

Roopinder: 7:54

How many of you were there at that point doing the actual development?

Scott: 7:57

Oh gosh, I don't remember. Jon would remember it started out with Jon, Bob Zuffanti, Jack Gray, Constantine, Mike Payne and Tommy Lee. So you know there was a group of people, but we were continually adding to the group. I actually have a CD in my shop. We sent the first gold CD out for revenue. It was like the first beta and I had the guy. It was Ken Van Swernijen, I believe. I had him from Echo. I had him install it and then send it back to me and everybody in the company could sign a CD. So that was when we first shipped the product. That's not too many people.

Roopinder: 8:39

Oh, I wish you were in your workshop you could show that to me. You actually have it. It's all signed.

Scott: 8:43

Yeah, you want to hold on. I can show it to you.

Roopinder: 8:45

Yeah, yeah, hang on a sec If you don't mind.

Scott: 8:47

Here, I'll take you into my workshop.

Roopinder: 8:48

Okay, this is a workshop in your garage.

Scott: 8:52

Yeah, behind my garage it's changed a bit since the last time we were there. I've converted it more and digital fabrication.

Roopinder: 9:03

So you're getting into metalworking now. Is your goal to make your own bike now?

Scott: 9:08

All right, so here on my wall, let's see, let me pull it out of its case here.

Roopinder: 9:16

There it is.

Scott: 9:18

In all its glory, everybody that was in the company at the time.

Roopinder: 9:22

Yeah, that's part of CAD history. Right, there should be a CAD museum in your garage. It could be as famous as Hewlett Packard with their garage in Silicon Valley.

Scott: 9:31

That's okay, I don't need to, that's not my ambition.

Roopinder: 9:36

It should be somebody's, though Maybe somebody in the media should be doing that. All right, a lot of things were happening. Windows was coming around, so you made a conscious decision to do this on windows, because everyone else was on Vax, right they were on workstations.

Scott: 9:52

You know there was Apollo, A dage, Silicon Graphics, DEC, Sun Microsystems had one. Everybody was coming out with these big, expensive unix workstations. Then all of of a sudden the Pentium chip I don't know if you remember the Pentium chip comes out and Windows NT. I think it was Windows NT, Jon will have it, he remembers this stuff better than I do. Windows NT just came out. So all of a sudden all the pieces were there. S oftware that would be sold through a channel for $4,000. That would run, at the time, PC and had all the familiar Microsoft user interface actually to run solid modeling and was powerful enough . It was a bit of a gamble but it was inevitable. Moore's law was in effect.

Roopinder: 10:45

Yeah, you were going up against giants. Computerv ision was kind of going downhill, you said, but PTC was going uphill. Yeah, and ProE and Autodesk...

Scott: 10:53

AutoCAD, because AutoCAD was still very, very popular. So the idea was could we capture the AutoCAD users that wanted to go do 3D. And on the way we picked up some fairly low-hanging fruit, frustrated Pro/ENGINEER users, and at the time we did not have parity on the feature set. We couldn't do a lot of what ProE could do, but eventually it could.

Roopinder: 11:27

Yeah, eventually it could. Let's say it's 80% of what PTC could do.

Scott: 11:31

The first couple of spins it wasn't even that, it was 30%. But the trick here was finding the right feature set which would satisfy people's needs. You can't do everybody. There's a million things you have to do. You figure out what is the feature set that's going to solve beachhead market, and if you could solve their problems and focus on them at the expense of some other ones, at least you're solving somebody's problem and you have value. If you don't solve anybody's problem, then it's a losing proposition. If you solve somebody's problem, at least you are valuable to them and then it's just a matter of expanding your feature set awareness.

Roopinder: 12:15

Looking back on it, it's easy to say you were on the right path. You were onto something. You were riding a wave of popularity that Windows is bringing on. Like you said, there was a certain disenchantment with ProE. They were high priced, so you were in the right place at the right time. But, scott, what did it feel like at the moment? Because you guys were out on a limb, you were as a small group, you probably didn't have all the money. You wanted to expand this thing, right? Did you have doubts? What was the feeling at the time when you were doing this?

Scott: 12:44

I would say there was no doubt. We did not have doubt. We were sufficiently paranoid in that, you know, we're always looking over our shoulder. We didn't want to be arrogant about it. There was nothing assumed and of course, anytime you're a startup, it's frightening, it's exciting. There's a lot of different, strong kind of emotions and feelings to sort of process. It was extremely tense, it was not at all relaxed. We argued a lot, which I think was actually healthy. I think you know, if founders and leadership never disagrees, then there's too many people like they're redundant or something. You want people that bring different points of view to the table and you make better decisions that way the group, even if it did disagree, stuck together.

Roopinder: 13:32

Absolutely, we stuck together.

Scott: 13:33

Absolutely. We stuck together. It was a pretty tight group. To this day I keep up with a lot of them.

Scott: 13:40

I keep up with a lot of people.

Scott: 13:42

It's funny because your paths cross, for example, a person that I met back when we were first getting started. We were evaluating Parasolid and John Stevenson. You remember, John?,

Roopinder: 13:52

I

Scott: 13:54

Yes, I keep up with John all the time. In fact, he taught a class with me two years ago and I see him pretty often,

Scott: 14:00

somebody I've known for 35 years. It's nice that you keep up with people in different ways.

Scott: 14:07

Even some people from the press. I keep up with. These people come out of woodwork.

Roopinder: 14:10

You're amazing and I've got to say for the audience we've become friends, we've ridden our bikes together. I always say this about cycling you ride with somebody for a hundred miles. They're your friend for life. And we did that. We did that with PMC.

Scott: 14:22

And the G2R2.

Roopinder: 14:24

Oh, yes, I may have to tell that story.

Scott: 14:28

You may not remember it though.

Roopinder: 14:30

Yeah, tell your version of that story. You'll remember some of that.

Scott: 14:33

Yeah, tell your version of that story. I remember some of that. So we were riding the Deerfield Dirt Road on Western Mass in Southern Vermont. You were riding very strong. You're on a borrowed bike. I think you were on probably one of my bikes. It's over 100 miles. It was like 105 miles and it had an insane amount of climbing. It was like 12,000 feet of climbing.

Roopinder: 14:56

I remember it was beautiful, it was scenic, the New England covered bridges and stuff. Gorgeous.

Scott: 15:01

It was largely dirt road but there was insane climbing. You crashed early on. We were probably 20 or 30 miles in and you took a digger. We stopped. I did a primary check on you. You clearly took a hard fall but you seemed pretty good. You were responsive. Your pupils weren't dilated. You sat right up. We sat for maybe five or 10 minutes. I did a primary medical kind of thing from my wilderness first aid classes.

Roopinder: 15:37

Right, because you were actually on a group that finds and saves people, correct?

Scott: 15:43

Yeah, the Central Mass Search and Rescue Team. You proceeded to get up, get back on your bike and just tore my legs off. You were like a fiend out there riding. We finished that ride and you were just crushing it. I thought you weren't affected at all by this. And then when we were driving home, we were talking about like hey, you know, I got to practice my first date on you and you said what are you talking about? You didn't even remember, you had no recollection of it and that was a little scary. I hope you went and got yourself checked out after that.

Roopinder: 16:22

I'm going to say I did.

Scott: 16:24

Okay, that was a little scary when you don't remember, like 10 minutes Pretty intense.

Roopinder: 16:31

I'll take the opportunity now to thank you for your attention.

Scott: 16:33

You weren't showing any signs of a problem and the the way you rode, you certainly weren't riding like there was any issue oh, the body must have taken over.

Roopinder: 16:41

But my favorite ride with you, Scott, my favorite one, you came to greet me, my tail end of my ride across America. For the last I would say like what 30, 40 miles you rode with me, that's right. And you had your whole team with you. SolidW orks right. And you came and greeted me and I thought it's like having a color guard.

Scott: 17:01

That's right, we escorted you in.

Roopinder: 17:03

Escorted me in to Boston.

Scott: 17:05

Made sure you made it all the way, make sure you didn't stop like 10 miles before the finish line of the across.

Roopinder: 17:11

Yeah, that was just awesome, and we rode PMC together, that's right. I say together, but I think you were way ahead of me. I don't remember Probably not, you were an exceptional bike rider. All right, fast forward from SolidWorks.

Scott: 17:26

That's right. Yeah, I retired in 2008 from SolidWorks.

Roopinder: 17:30

Well then, what happened?

Roopinder: 17:34

Because you guys came out to start Onshape.

Scott: 17:35

Yeah, it's funny how that happens. Yeah, I was really having fun in retirement and it was a really good balance of everything. We all kept up. We decided we would still keep up and talk about issues. Even though we went our separate ways, we still had a lot in common. At some point we felt that the job wasn't finished right, that there was new disruptive technologies that would change the game.

Roopinder: 18:06

Like the cloud right?

Scott: 18:07

Like the cloud, yeah, and a different way to architect products to be much more robust, more collaborative, more secure, faster, more accessible.

Scott: 18:19

Database instead of files Faster because you're working on these giant instances almost infinite memory, tons of giant CPUs, fail-safe types of operations. You can be modeling in SolidWorks where if there's a problem you get a core dump right, blue screen of death or whatever. With Onshape, parasols may crash or something may go wrong and as a user you may not even notice because the system is designed to not crash. If something goes wrong, something else takes over and your session continues, sometimes without even the user even noticing. So you don't lose data because it's a database and the way the database is architected. There's no install. All of a sudden, a new technology comes along that allows you to sort of reset things a little bit from a technology and from a workflow standpoint.

Roopinder: 19:17

Scott, I'm reading between the lines. Perhaps you said it was because a lot of technology came forward and you were able to take advantage of it, but how much of it is due to SolidWorks becoming difficult to deal with? Like you heard these reports, it's almost like the history tree keeps breaking or people are having problems with SOLIDWORKS. Had it gotten too cumbersome?

Scott: 19:39

Yeah, but I think so part of it is just problems inherent with the SolidWorks we've designed. You know at this point about what, 32 years ago. So you know there are problems inherent with file systems. There are problems inherent with the different types of operations and bringing all legacy data forward. So those problems have existed all along. It's not like SolidWorks got worse and worse, it's just that expectations we could reset people's expectations to be better than what their file system could provide. Data management, PDM, was always an issue. You always had to check out a file and then lock it and then, if there was parallel efforts, then how do you merge them together? You know there's all these issues all existed, have always existed them together. These issues all existed, have always existed. And so, with the eventuality of the cloud and database operations, the way we were doing things in Onshape- there's a lot of problems we had that we didn't know we had.

Roopinder: 20:43

My favorite one is a rather simple one that was fixed as soon as CAD got on the cloud. I don't have to save anymore. Not a week would go by. I wouldn't lose data because of a crash or power failure or something and I hadn't saved my data. You heard it in every CAD room. O h shit. T he power had gone off, somebody had lost their file. We didn't have to save anymore.

Scott: 21:03

It's a huge one. We allow collaboration sort of like Google Docs, kind of Google Drive. They set a new way to collaborate with people on data.

Roopinder: 21:14

So, yeah, that's a good example, and I think we also mentioned GitHub. Like, the way developers use GitHub is the way you wanted CAD users to work.

Scott: 21:22

You can fork and you can branch your designs and manage your projects in a different way.

Roopinder: 21:29

Right Now, looking back on it, it sounds like this is the way CAD should have gone, but I think the team sensed that they could only do this if they started over again, almost like.

Scott: 21:40

SolidWorks, everybody tried to virtualize. We do cloud, we virtualize. That's a non-starter, because the other way people are doing it is you'd install an app locally but they'd sync the data, so they just shove files back and forth, which is it's not the same.

Roopinder: 21:55

I remember talking to you one time. I think I brought this up Are you worried about people putting their stuff on the cloud? Everybody is. I remember exactly what you said. You said I don't know about that. We had to fix a lot of hard problems. We did we had to fix a lot of hard problems.

Scott: 22:10

I'll show you another thing on my wall here the things that make the place bonner in my shop here.

Roopinder: 22:16

I think one of the big problems would be latency right. Oh, yeah, let's see, I see it the fillet, the radius there.

Scott: 22:27

That is the patent right there. For a distributed system CAD in the cloud.

Roopinder: 22:36

Ah, okay, that's a copy of the actual patent application or the acceptance.

Scott: 22:42

That's the accepted patent, yeah.

Roopinder: 22:44

Ah, okay, very nice, and you keep them in your garage.

Scott: 22:47

You have to walk through the garage. I don't keep it in my living room.

Roopinder: 22:51

So now I gotta say I'm not nearly as accomplished as you, but the things that I have managed to do, like, for example, marathons, I keep mine in my office. It's like they're not at home. How many patents do you have? How about two patents?

Scott: 23:05

So, Onshape, it's three here. I've got a couple more that are there's one at SolidWorks and one that came up with a company that I was consulting with.

Roopinder: 23:15

So you got Onshape. You managed to launch Onshape. I would say, make it a real contender in the MCAD space. And then what you retired again.

Scott: 23:26

Yeah, I retired again in 17. I stayed on for a year after I retired.

Roopinder: 23:33

Did you?

Scott: 23:34

I stayed on working one day a week. I was having a good time working on special projects that I thought needed doing.

Roopinder: 23:41

When you retired the second time, had Onshape already been sold to PTC?

Scott: 23:46

No, it hadn't sold yet. It had sold maybe a year after.

Roopinder: 23:50

Did you miss out on a payoff then?

Scott: 23:53

No, I still had all my stock. You know it's important when you start companies to set up your equity structure to be one that make tons of sense, and we were really careful about how we set it up, that it made sense in the eventuality of an exit.

Roopinder: 24:11

Would you say you were more sensible about it with Onshape than you were with SolidWorks?

Scott: 24:17

What do you mean?

Roopinder: 24:18

In your equity structure when setting up the equity structure.

Scott: 24:21

We did it the same way.

Roopinder: 24:22

You got the band back together.

Scott: 24:24

We got the band back together and then some right, Because we got Hirschtick and I from the founding team of SolidWorks, but we also got a lot of the lead developers that worked at SolidWorks. Two of the founders were CEOs of SolidWorks, Jon and Johnny Mac, so you got the Head of Development with Dave Corcoran we got Tommy Lee. W e got Mike Lauer. It was quite a crew.

Roopinder: 24:48

You got the band. You got the road crew, you got the bands.

Scott: 24:52

Bands and people they were in bands, the road crew and employees who believed in us in the early days of SolidWorks, helping with testing products, things like that.

Roopinder: 25:00

Yeah, you make a lot of really close colleagues did you feel like now you got the job done. You were happy. You wanted to sit back now, ride your bike, enjoy your wine solidworks was in great shape when I left.

Scott: 25:13

It was really a great time. Great team left with a lot of great friendships and good feelings towards people there. And you know what? The same with Onshape. I l eft Onshape and it was just a great time For me It was just time to retire and continue. I started a program at Olin College. It's a senior capstone program. It's an engineering capstone for entrepreneurs, but it's specifically designed for entrepreneurs and, by the way, McEleney and John Stevenson are really involved with it. They meet with the students.

Roopinder: 25:50

I got to pass on my condolences really for Vic. He was very much in the fabric of SolidWorks.

Scott: 25:57

He was one of the main reasons SolidWorks was successful. He built a channel. You can build great products but you have to build a great business. We could all build a great product, but without a great business being built behind it we wouldn't have been as successful. Vic was instrumental in building a great business, that channel. He built the sales. The team he put together was just. He was a bigger than life person. I don't know how well you knew him, but I knew him quite well. We really miss him. Every time I open a bottle of wine, I think of Vic. He had a very nice wine cellar.

Roopinder: 26:34

He was in the Bay Area.

Scott: 26:35

He had a beautiful house. We still keep up with Diane. Just a great person, great business leader.

Roopinder: 26:43

That team was phenomenal. Everybody was so happy. I've got to say a lot of the success of On hape was due to the people that believed in the old team it was also the fact that we sort of came prefabricated as a team.

Scott: 26:55

We knew how to work with each other. We knew each other's strengths and weaknesses. Having those relationships already existing reduces the bandwidth to getting things done.

Roopinder: 27:06

You already know each other. You know what makes things, what makes things not work, so you could skip over all the formalities, as it were. Let's do this.

Scott: 27:18

For example, Dave Corcoran and I actually worked at Pratt Whitney when I was there, he worked at Computerv ision. When I was there, he worked at Oliver until we'd known each other for a long time.

Roopinder: 27:24

It's almost like you couldn't separate if you wanted to.

Scott: 27:27

It's amazing how your paths cross with people you just never know.

Roopinder: 27:31

That is an amazing journey what you've accomplished. I've always had it in the back of my mind to tell that story and I think somebody has to do it. Somebody has to tell a modern version. Weisberg did a history of CAD, but that was a long time ago and it doesn't include the most recent chapters. So I know you're out of the business now, but I got to say this is the most exciting time to be in CAD because of AI. A lot about AI.

Scott: 27:59

I actually mentor a few companies now. Even though I'm retired, I still try to not just keep current but also drive some of the directions on the way these companies are thinking about their approach to the problems they're trying to solve.

Roopinder: 28:14

Are these startups developing AI?

Scott: 28:16

Right now, every startup calls themselves an AI startup. Right, Of course, AI, it's almost meaningless at this point because it's, of course you're AI, but that's kind of like a CAD company saying they are a C++ company. It's a tool that everybody's going to employ. Right now. It's so new and everybody's trying different approaches and angles. Some of them will work and some of them won't, and so it is a very exciting time. I think there's so many problems that can be solved and we're going to see huge changes in workflow capabilities and integration, more than PCs gave SolidWorks and more than even cloud is going to give Onshape. I think AI is going to be a huge game changer and the people who get it right are going to win, and you see a spectrum of approaches to these problems.

Roopinder: 29:13

I totally agree. There's so much that could be done. When I write about it, I say, finally, after all these years, computer aided is in CAD. That could be. I should say, if it's done right, if AI used properly, you could finally have CAD that helps you design things. And let's face it, all CAD has done so far is just make 3D models more exact than we have in our heads.

Scott: 29:36

There are some things that that enables being able to develop a design language which allows you to articulate designs. Drafting was one sort of language. Solid modeling and feature-based modeling was another kind of language. Things tend to get more abstract, but there's power in that abstraction. So I think that the tools that become natural to use be successful, the ones that will help define a language that's more descriptive has more implications than sheer geometry does. Features have more metadata and CAD doesn't necessarily use it Like you put in a hole. Sometimes whoever's consuming the CAD data will know it's a hole, and sometimes they won't. Sometimes they have to interpret it as a hole or go to a drawing and get a call out and see what the hole is. I think that as the languages become more descriptive and more abstract, the ability to support the design process becomes much more interesting.

Scott: 30:39

You're not just designing each component. You can design something in a much higher level and in a much more iterative manner. CAD has always played well to linear process, even though, yes, you can go back and do parametric changes and all that stuff, but it's all parametric changes that you need to design in. But life isn't linear and design is a messy, iterative process, and I think the more we can support those types of design, the iterations and more top-down kind of design, I think the better, and I think AI is a disruptive technology that will allow this. I haven't seen anybody do it right yet. I've seen a lot of people poking at it and I give them a lot of credit for coming up with different ideas for design, language and for process, but I've yet to see it. I have not seen one that I think is perfect.

Roopinder: 31:36

I see this glimmer in your eye like there could be yet another. Who knows? Or at least you're going to invest it. We've spent a lot of time talking about the history of CAD and, I think, the future of CAD. We need to talk about it. We need to have another discussion about that, but I've got to let you enjoy your evening.

Scott: 31:53

I actually do have a time and strength that we have somebody give a guest for dinner.

Roopinder: 31:57

I'll let you go, but before I do, Scott, it's been wonderful talking to you and catching up. Is there any of your companies that you're investing in that you want to give a shout out to?

Scott: 32:08

I'm going to. I'm going to hold off.

Scott: 32:25

Because the problem is there's actually many.

Roopinder: 32:27

All right, well, listen, have a great evening, have a great dinner, say hi to Jody for me.