The critical role of software in operations
Software is everywhere in our modern world and that includes some of the most risk-averse and conservative industries. Then challenge is making it work for a business rather than making the business work on the software. To delve into the role of software in factory operations, we sat down with Mark Hindsbo – someone who grew up with and built a career around software.
Nick Finberg
00:09
Welcome to the Future Ready Podcast from Siemens. I’m Nick Finberg and we’re covering Software-Defined Everything today. More specifically, I have our co-host, Mark Hindsbo, the head of operations software at Siemens Digital Industries. He’s here to talk about his experience in the transition to software-defined from an operations software perspective. Mark, you’ve got such an interesting background. And if we had more time, I just want to talk to you about your time at at CERN and the large Hadron Collider. But maybe to focus our discussion a little bit, would you mind talking about how you got into computers and software in the first place?
Mark Hindsbo
00:44
No, for me it was actually a magic moment, I would say, of discovering computers. As a young kid. So I grew up in the countryside of Denmark, you know, amongst farmers. And computers were in general far away and they typically were certainly not in the home. But one of my friends who had a I guess a pretty curious and forward-looking dad had bought a Commodore VIC-20 So I guess I I date myself with that. But but he uh he let us play with it, you know, as kids. And and I just remember uh keying in that first basic program and it actually doing what you told it to do. uh you have stuff appearing, uh you figure out that you made a mistake, right? Typically when a computer does not do what you want it to do, it’s your fault, not its fault. Uh it’s they’re very stupid.
Nick Finberg
01:36
They just follow your commands.
Mark Hindsbo
01:38
Exactly. Uh no, so that that really was a you know uh it hit me at the right point in time, I guess, of curiosity, of of everything else. I then spent a couple years picking strawberries. to afford my own computers. I get up at 4 a. m. in the morning and uh uh it was all you know with the uh intent of getting to own my my own computer which was a Commodore 64. And since then I’ve been uh in and around computers and software. I uh I’ve programmed uh since then I still program uh for fun No one allows me to do production code anymore. Those days are over. Wehave sm I have smarter people on my team that that do that. But yeah, I am I still every now go down in my basement. I teach a class in AI at Carnegie Mellon, which is one of the local universities. So I get to write a little bit of Python code together with my students. Uh so yeah.
Nick Finberg
02:32
Very fun.
Mark Hindsbo
02:33
It’s been a very it’s very much been a lifelong journey.
Nick Finberg
02:36
Well, um I guess kind of a continuation, what brought you to Siemens then from this background?
Mark Hindsbo
02:42
So I I think a couple of things melted together for me in my life. I am an engineer by uh by education. And and it’s been a real pleasure in my career as well to I mean I grew up with software. I I’ve worked for instance at Microsoft, so I’ve I’ve seen how pure software you know shapes the world. And it gave us, you know, Microsoft, we talked about Google, you know, Netflix, your internet bank. But I characterize them as more or less pure software experiences. You know, I I do really believe that a lot of the things that are both challenges and opportunities in this world require the physical and the virtual world to really come together. So whether that be the robots that are gonna help take care of me when I might need it a couple years from now uh whether that uh will be uh sustainable products or the energy challenges we have, uh personalized medicine, you name it. those are things where software can really enrich the real world and and in the opposite. And ultimately that’s what brought me to Siemens. I think Siemens has a tremendous opportunity to help our customers, the ones that build self-driving cars and next generation chemical plants and uh you know solar uh energy of the future to give them the software tools so they effectively can manage the the real and uh digital world. So that that brought me to see I feel that we’re in a point in time where we’ve gone from a call itmore or less pure software revolution. Well, let me take a step back. We had a mechanical revolution first, right? We had an industrial revolution uh that was a hundred percent mechanical, uh that we had a software you know revolution if you want that was almost a hundred percent digital. I think we’re now in the era of combining those two
Nick Finberg
04:41
Okay, perfect. Well, I in in our last episode we talked with Rainer Brehm about uh the concept of automating automation and how that kind of fits into the story at Siemens and software defined and he was really making the point of it is the connection between hardware defined and software defined. Um I’m curious of your take on that and maybe more specifically how that idea might integrate with operations software, which I know you’re leading up.
Mark Hindsbo
05:06
Yeah, no, and I think uh you know when when Rainer talked about automating automation or you know hopefully when you hear me talk, Rainer I really share that mission and belief of what needs to happen. But but let me take a step back because even when we talk about hardware defined or software defined That’s really a technology-led story. And I think we should we should we should maybe take a step back and say, you know, what does it do in outcome for our customers and their customers’ customers, so to speak? And and we we are in an extremely dynamic world for for good a lot of time, but also some bad, to be honest. But our customers want to create more individualized products. We had a revolution of mass production, which was fantastic. And what’s the old saying you can get uh your 4T in any color you want as long as it’s black. Right? So so those days are gone. I want it in my color, right? Internally, externally. I need my medicine Specifically tailored to my DNA. We have supply chains that are constantly being refactored. I might have a supply in one part of the world today, another part of the world tomorrow. We might have a piece of PFAS regulation that says this plastic is not optimal, please substitute it with something else. That requires our customers to be a lot more adaptable. both in their product design processes but also in their production processes. So the if you want linear assembly line that was pioneered by by Ford, right? where we go from step A to step B to step C now needs to be a much more dynamic one where we might go step one five seven for one product, we might go step you know one, three, and nine or another product. The assembly process needs to auto-adapt itself to whatever’s coming in and coming out. And that requires this software-defined hardware or hardware-defined software, whichever way you want to talk about it. to really come together so that our uh you know our customers can have those facilities. So that’s what you know what it means to me.It’s about the The supply chains being more adaptable, it’s about the product being more personalized, it’s about the product being more sustainable and energy efficient. And I only believe that we can reach those capabilities by by now automating automation, in Rainer’s words. Right? Because for me the automation, I think Rainer would say the same, right? The automation is this tremendous benefits we’ve come out of of taking this linear you know uh assembly line concept and and massively automating it.
Nick Finberg
07:58
Yeah.
Mark Hindsbo
07:58
Now it’s time to be able to have that be more dynamic and break it up and allow it to be nonlinear.
Nick Finberg
08:03
Yeah, he had mentioned there’s a a fear or maybe hesitation with some of our customers.
Mark Hindsbo
08:16
Yeah. Well, we’re all engineers and the old saying is if it’s not broken, rif while why fix it? So so so so so I get that. uh on on the flip side uh uh every engineer is cura curious and the first thing we do is take stuff apart and break it. And and I I I do generally think we have a real opportunity in to reassemble uh things. And in a lot of cases, we are at the point now of diminishing return for sort of the classic thing. The next level will come from us having more adaptable and and and autonomous uh capabilities in in uh in our factories.
Nick Finberg
08:58
Okay. Well to that point, I I’ve heard you talk about operations software in the past and there’s this this one word you use to define it. Define the future and then it would be opportunity. Would you mind explaining what you mean by that?
Mark Hindsbo
09:10
Well, I think let me just take maybe a step back and then we’ll get to uh to opportunity because maybe for Some of our listeners we have to define what operation software is. When we talk about it at least in a Siemens, because there’s many operation software, right? For uh depending on what uh part of the industry or vertical you’re in. But when we talk about operation software, you know, for my part in Siemens, it’s the software that either allows you to design a factory or plant and then subsequently operate that factory or plant. to to produce whatever product that that you might be producing. So that’s sort of the scope that we’re we’re we’re we’re talking about. And if you think about Why do I say opportunity is is back to this other notion that we are at this place where these worlds are melting together that will allow us some great new capabilities. So if we, for instance, think about the design side Of designing a factory. We’ve had classic, call them cat-based tools that allow us to design factories for some years now. But we really need a uh digital twin output from them that allows us to in the virtual world massively simulate our our production capability and what we can do to automate it. We’ve done that in some respects for product design. Even today you could argue that we build physical prototypes before we have to. But Think about the auto yo automotive industry. They used to crash test cars physically, prototype cars. That’s a million dollars roughly per car that’s crashed and you can only afford to do a couple. Today you can for a million dollars you can virtually crash test 10,000 cars and you can have a much bigger design space for trying to optimize it. And then we take a production car off the assembly line and we physically crash it to make sure that our digital uh uh prototype uh told the right story.
Nick Finberg
11:19
Yeah validate that that simulation.
Mark Hindsbo
11:22
We do not generally do that for our factories today. But you could produce tens of thousands of variations of your factory and the humans and the robots and the machines and how they interact and and how you get the optimal throughput as part of your planning process And that was uh by the way part of what we announced with NVIDIA at CES, that we’re now uh taking the capabilities we have to simulate your production process. We’re putting them into the NVIDIA metaverse so you have an even richer ex uh 3D visualization experience of those. But uh but at scale now you can do those tens of thousands. And we’ve done that for our own factories and some of our forward-leaning customers have done it. And the productivity gains that they get are not multiple orders of magnitude, but certainly close to order of magnitude in their throughput and productivity of the factories that they’re building going forward. But then we need to take that digital twin and also put it into the production itself. so that you can play with it in real time before you make a decision to choose to now route this product through another part of the line. you know, let it let the simulation run and figure out what’s the optimal uh uh uh way of configuring in real time and adapting to a changing world We need to put AI on top of that so the AI can play, but have that deterministic model if you want. So when the AI hallucinates. uh that the deterministic model catches it and say, oh no, that won’t work, right?
Nick Finberg
13:02
Yeah.
Mark Hindsbo
13:03
Go try something else. So you get this interplay between uh call it stochastic algorithms like AI’s and then you know deterministic algorithms and and and you get a win-win. So in the operation phase, I think a lot of that will will then, as we unlock especially the data layer on the factory as well. So you have information about everything that’s going on. uh across your operations and and production facilities, we can really unlock a lot of new value and we can allow actually both humans and AI to play together and and leverage who’s best at at what. So that’s what I talk about when I talk about opportunity. We have a fantastic foundation, right? We have classic factory automation in place. We have uh SCADA systems, DCS systems, classic MES systems. That allow you to automate everything. And without that, you could not automate automation. If the AI cannot, or the human being that’s play, that’s directing the AI. Do not have an automation layer beneath them. They can only talk about things. They can’t act. Right? So it’s that interplay between the new capabilities that are coming on board with the foundation that we already built. that that spells opportunity. You know, not to be a monologue. But if if we want to kind of go back to where we were with physics, wasn’t it Newton who said if I seen further than anyone else, it’s because I stood on the backs of giants.
Nick Finberg
14:41
Yes.
Mark Hindsbo
14:41
I think some sometimes when we talk about new stuff, and I catch myself in this as well. We get fascinated by AI and new opportunities. They would not be possible if we did not already stand on the foundation that we built.
Nick Finberg
14:55
Yeah, that is so universal that it could be hard to remember how often it is the case. We’re here because of the many small steps taken over long periods of time and investments and innovation by those who came before us. I definitely want to keep chatting with you, Mark, but I think let’s leave it here for now. We’ll be back soon with even more on software defined everything and the role of operations software. If you haven’t already, we have some great episodes with Rainer Brehm talking about the automation side of all this, or you can check out even more on our website. We hope to have you back again soon for another episode of the Future Ready Podcast.
