Latest / Automation Ladies / From AI to OT SCADA CON: The Future of Industrial Innovation with Bryan Thyken

Transcript

From AI to OT SCADA CON: The Future of Industrial Innovation with Bryan Thyken · Automation Ladies

• 1:07
  SPEAKER_01: And uh yeah, he he's the one I think who I've ever first heard that from.
• 1:13
  SPEAKER_01: And uh yeah, let's talk about that and uh where you came from, how you got where you are.
• 1:19
  SPEAKER_03: Yeah, no, absolutely.
• 1:20
  SPEAKER_03: Uh thank you for having me on the the show.
• 1:23
  SPEAKER_03: And yeah, my name is Brian Thykin.
• 1:25
  SPEAKER_03: I'm the head of revenue for Sorba AI.
• 1:28
  SPEAKER_03: Um and yeah, my career kind of started off coming out of the military.
• 1:32
  SPEAKER_03: I was uh an electronics technician um in the Navy.
• 1:36
  SPEAKER_03: I I did mostly topside work, meaning I worked on uh radars, satellite communication, terrestrial communication with uh onboard ships.
• 1:46
  SPEAKER_03: I I joined the Navy after an unsuccessful first year of college when uh a roommate of mine put in uh an anonymous little mailer saying uh I wanted to join and become a Navy SEAL, and I got a call from the recruiter.
• 2:01
  SPEAKER_03: And looking at my bills and looking at uh my grades, I decided to jump on my skateboard and go down to the recruiting office and sign up for not a normal four-year stint, but a six-year stint in the military, much to my my parents' uh dismay.
• 2:17
  SPEAKER_03: So um after eight years in the military, um, uh two deployments and a lot of time out in the middle of the ocean, um I got out, uh, was going back to school while working at Siemens, where I started my career as a controls engineer, as they called it, uh, engineering specialist, and worked there uh while getting my degree, actually with a Bachelor of Science in Psychology, uh, with the intention on actually going into academia and be getting my PhD in psychology and industrial psychology uh specifically.
• 2:51
  SPEAKER_03: Speaking to one of my professors, kind of found out how much they made.
• 2:54
  SPEAKER_03: And at the same time, I was looking at going into sales um after a few years working at Siemens and uh ended up choosing the sales route because there's uh it was a few more dollars involved.
• 3:05
  SPEAKER_03: And and my my buddy promised me too that all I was gonna do was uh drink martinis and play golf.
• 3:10
  SPEAKER_03: And so I figured, why not?
• 3:12
  SPEAKER_03: You know, let's let's get after it.
• 3:14
  SPEAKER_03: And it turns out that's not the case at all.
• 3:16
  SPEAKER_03: And so after about nine years at Siemens, I believe it was, I left and went to a large systems integrator at Maverick Technologies, which I believe you and I were both there at the same time, roughly.
• 3:25
  SPEAKER_03: And Paul uh Gilleski is very much a mentor of mine.
• 3:29
  SPEAKER_03: Um, learned a ton from him, had a lot of incredible opportunity to um, you know, have him coach me.
• 3:37
  SPEAKER_03: And even to this day, uh, you know, it's always great to get a text from Paul or shoot him a message and ask his opinion on different things.
• 3:43
  SPEAKER_03: He's uh, you know, in terms of the business of running a systems integration shop and and the industry as a whole, he's a very brilliant individual, you know, and a leader in that.
• 3:53
  SPEAKER_03: After Maverick was acquired by Rockwell Automation, I uh left for the startup world to kind of go cut my teeth on a few other things, still really hate staying in the industrial automation space and went to a couple of different startups doing different things from you know hardware to um some software technologies, and then kind of journeyed out of the startup world into a big company, an ABB, and then back into the startup world where I ran back into Sorba, uh, who I discovered while I was at one of those startups early on.
• 4:26
  SPEAKER_01: So let's talk about Sorba and what Sorba is.
• 4:30
  SPEAKER_03: Yeah.
• 4:30
  SPEAKER_03: So Sorba is an end-to-end AI ML platform to enable subject matter experts to build models and deploy them in real time and even perform closed loop control.
• 4:42
  SPEAKER_03: Everything starts off at the data collection layer.
• 4:44
  SPEAKER_03: So we have over 65 plus different data connectors.
• 4:47
  SPEAKER_03: Um, so you can help create that unified namespace.
• 4:50
  SPEAKER_03: Go into an auto ML piece where you can build AI and ML models, everything from anomaly detection, classification, digital twins, advanced process control models, and forecasting models, um, all in no code.
• 5:03
  SPEAKER_03: So you as an SME can build those models out and then deploy them into real time and actually run them in closed loop.
• 5:10
  SPEAKER_03: And then the models actually retrain on themselves.
• 5:12
  SPEAKER_03: So the kind of brilliance behind it came from two co-founders.
• 5:17
  SPEAKER_03: Um, those two co-founders are both controls engineers, uh uh Aldo, uh Ferrante, and Yani Perez.
• 5:23
  SPEAKER_03: They they really wanted to make it where the subject matter experts who knew the processes, knew how the equipment ran, knew the ins and outs of everything, could build models, use on-prem data, not have anything connected to the cloud, or have those dependencies, and be able to rapidly iterate and solve problems.
• 5:42
  SPEAKER_03: And that's what they've built over the last 10, going on 11 years now with the company.
• 5:46
  SPEAKER_03: They've they've really driven every feature we release is is, you know, has to run on-prem first and foremost, um, be able to run within their OT infrastructure.
• 5:56
  SPEAKER_03: And it all is about enabling the subject matter experts to really build those models and and leverage their expertise without having to know Python or data science or any of those skill sets that are one hard to find, but two also lack the context of the industrial marketplace.
• 6:13
  SPEAKER_01: Well, what about uh hardware?
• 6:14
  SPEAKER_01: Does sort of provide the hardware?
• 6:16
  SPEAKER_01: What do you need for hardware?
• 6:18
  SPEAKER_03: So, from a hardware perspective, we we are just a pure software company.
• 6:22
  SPEAKER_03: So um, you know, while many people, I think, I think there's a lot of, I don't know if I want to call myths because it'd only be a myth if if other companies didn't require it, but we we run on a CPU.
• 6:35
  SPEAKER_03: So our standard install is actually a 32 gig, eight core, 500 gig hard drive um machine.
• 6:43
  SPEAKER_03: So you can run on pretty much your standard desktop.
• 6:46
  SPEAKER_03: You can put that, uh, run an AI deep learning model in said hardware, uh, run it all on-prem, retrain it on-prem, do everything on-prem.
• 6:57
  SPEAKER_03: Um, and that's usually provided by our systems integrator partners, by the end users themselves.
• 7:04
  SPEAKER_03: We we just provide our software um very much following in the inductive automation manner where we have a free two-hour license.
• 7:11
  SPEAKER_03: You can install it and play around with the tools um and and see what you can build.
• 7:16
  SPEAKER_03: And then uh, when you're ready, go and put a full license in and build away.
• 7:21
  SPEAKER_01: So, what can you say to other controls engineers about AI?
• 7:25
  SPEAKER_03: So I have been in the AI space for I'd say about six years now, maybe seven.
• 7:34
  SPEAKER_03: First, most companies require the cloud to do AI.
• 7:36
  SPEAKER_03: And I think that that is kind of antithetical to a lot of industrial automation stuff in general.
• 7:42
  SPEAKER_03: I mean, you know, when you make cookies or make oil or make beer or whatever your product is, you know, the values in the recipes, the values in all that information in there, that you know, how you make it and that information and having that information up on the cloud, you're immediately exposed.
• 8:01
  SPEAKER_03: I think you know, AI, because you start becoming cloud dependent, you know, is the first downside to a lot of the applications.
• 8:09
  SPEAKER_03: The second piece around AI is it's really heavy on having so many other resources involved who don't necessarily know the plant.
• 8:18
  SPEAKER_03: And having worked in other tech companies, you have AI, you know, data scientists who are brilliant and and you know know a ton of stuff, but they've never been in a plant.
• 8:30
  SPEAKER_03: They've never seen a PLC cabinet, and they've never opened that up and seen the rat's nest.
• 8:35
  SPEAKER_03: And again, I've seen photos from you know your LinkedIn, uh uh, you know, all sorts of people who are out there to show these just, you know, the dirt, the grime, the dust, you know, being out there.
• 8:47
  SPEAKER_03: They don't understand, you know, that you're running on a slick uh that you know has been in there for 20 plus years, running just fine.
• 8:55
  SPEAKER_03: You know, they have very much an IT focus on things.
• 8:59
  SPEAKER_03: And so it's really hard to marry the two things.
• 9:02
  SPEAKER_03: I actually there's that saying, you know, the ITOT convergence.
• 9:07
  SPEAKER_03: And I actually very much disagree with that statement.
• 9:10
  SPEAKER_03: I think OT realized IT has some cool tools and they grabbed them, but the IT people really don't fully understand the complexity, the challenges that occur when you're trying to make something.
• 9:23
  SPEAKER_03: And, you know, a lot of these companies also, you know, AI comes in as this like salvation to, you know, making things better.
• 9:30
  SPEAKER_03: When the reality is none of these businesses would be around if they weren't making products profitably.
• 9:35
  SPEAKER_03: Um AI, in my opinion, is something you're helping to augment a workforce that is struggling to get people into it in general.
• 9:47
  SPEAKER_03: Um, I mean, I remember a stat, and I'm sure the stat is definitely out of date from when I was told it by Paul as part of our sales pitch when we were at Maverick.
• 9:56
  SPEAKER_03: The average age of an industrial automation engineer was 55.
• 9:59
  SPEAKER_03: That was, I don't know how long ago was I was in Maverick, but I mean 10 years ago.
• 10:03
  SPEAKER_03: That is, yeah, that was, you know, it's I doubt it's gotten better.
• 10:08
  SPEAKER_03: Um, and so you you're using tools like Sorba to help, you're using AI to help augment the workforce.
• 10:18
  SPEAKER_03: It's it's definitely not gonna replace right now.
• 10:20
  SPEAKER_03: There's not enough people even to fill in the gaps that are currently there.
• 10:24
  SPEAKER_03: So to me, when I look at companies that go out there and and kind of laud all this incredible savings that AI has brought, first off, I, you know, it's it's um these companies are probably doing a pretty exceptional job of the tools they have and they're making profits.
• 10:40
  SPEAKER_03: What you're trying to do is I feel like AI is about creating consistency in a product.
• 10:45
  SPEAKER_03: You know, PID is gonna have swings, it's a reactive thing.
• 10:49
  SPEAKER_03: Whereas machine learning in AI can be a predictive thing.
• 10:52
  SPEAKER_03: It's gonna say, hey, I know when these conditions are met, I need to be at this set point.
• 10:55
  SPEAKER_03: So rather than PID hunting for it, I'm just gonna go to it.
• 10:59
  SPEAKER_03: And I think that that's kind of when I talk to people about what AI can do, it's about creating that consistent behavior in your machines, predicting what it can be.
• 11:08
  SPEAKER_03: So you instead of having a little bit of swing and having potential loss on that swing, you're just staying at a nice steady yield because you know, you can only put um, you know, 10 gallons of water through a certain size pipe, or you can use whatever uh colloquialism you want there.
• 11:24
  SPEAKER_03: Um, all you're trying to do is make sure you're consistently sending that through.
• 11:27
  SPEAKER_03: And that's where AI and machine learning help.
• 11:29
  SPEAKER_03: And it also then takes out, you know, it can augment the operator there by doing a closed loop control and ensuring that is staying at that rather than having an operator, you know, shift to shift, have differences, you know, how they run a machine, etc.
• 11:45
  SPEAKER_03: So very long-winded way of saying, you know, like I look at AI as much more as an augmentation in the industrial space of really supporting the controls engineers, the process engineers in a plant where they're already overtaxed with too much work uh in trying to keep a plant up and profitable.
• 12:05
  SPEAKER_01: And very similar question.
• 12:07
  SPEAKER_01: What do you have to say to manufacturers who don't want to use AI?
• 12:12
  SPEAKER_03: I honestly think that there is, you know, one of the biggest challenges I think with holds people back is you go into a project, they've heard these stories.
• 12:27
  SPEAKER_03: Um, but you know, usually the objections are, well, I gotta put my data in the cloud, that's your risk.
• 12:34
  SPEAKER_03: Or, you know, the projects take a long time and I don't know if we get to value.
• 12:38
  SPEAKER_03: Um I personally, you know, again, if you're profitable and making products and you have your staffing great, I think it's important just to stay, you know, find find the way that makes you feel most comfortable.
• 12:50
  SPEAKER_03: There are technologies out there, though, that can help meet those different pain points and help you get to value quickly.
• 12:58
  SPEAKER_03: Um, one of the things that I think is very unique about what we do is the data science process is very long.
• 13:06
  SPEAKER_03: So you like, let's say you want to like optimize the boiler and you want to use machine learning for that.
• 13:12
  SPEAKER_03: You know, you have to do a bunch of data ops, it could take days and weeks to go through your data.
• 13:16
  SPEAKER_03: Then you go into the model building itself and you're going through and sequentially picking, you know, what algorithm you want to use and testing out for the best fit.
• 13:25
  SPEAKER_03: Then you deploy that model and you put it out of production, and that takes time.
• 13:31
  SPEAKER_03: And then you find out it didn't actually solve problems.
• 13:33
  SPEAKER_03: You go back to this like three, six month process, and you spent money and more importantly, time, which you don't have.
• 13:40
  SPEAKER_03: And that becomes very like scary for people to get engaged in that process.
• 13:45
  SPEAKER_03: And what we've done, and I was over visiting with one of our customers overseas uh a few weeks back.
• 13:51
  SPEAKER_03: We sat in a room and with our software, we were able to iterate within minutes on you know the different things.
• 13:56
  SPEAKER_03: So we were able to like test out an algorithm and a model that we want to put in production on, like, say a fill line.
• 14:02
  SPEAKER_03: And we were able to go, okay, that one didn't work, build a new one, build a new one.
• 14:07
  SPEAKER_03: And in a matter of minutes, not days, weeks, months, get to kind of value.
• 14:13
  SPEAKER_03: I think that that when people say, I don't want to do AI, I think it's a lot of one, they they see ChatGPT and like, you know, my data's going out here, it's gonna be all in the exposed.
• 14:23
  SPEAKER_03: They see the time.
• 14:24
  SPEAKER_03: There's just a lot of those things which they don't have.
• 14:27
  SPEAKER_03: I think technologies are coming out.
• 14:29
  SPEAKER_03: I know tech obviously we're store, but I know there's technologies out that can solve that and help that process be a lot less painful uh and scary, to be honest.
• 14:39
  SPEAKER_03: Um, you know, and again, we're reusing in some cases a lot of techniques that have been around for a long time.
• 14:45
  SPEAKER_03: I mean, I feel like AI is also a little bit of a rebrand of, you know, looking at old techniques, but just with new IT technologies that we're able to kind of absorb over uh from that side of the house and and use in the OT environment.
• 15:01
  SPEAKER_01: What advice do you have for young people that are that are trying to get into uh AI and machine learning?
• 15:07
  SPEAKER_01: What should they be teaching themselves?
• 15:09
  SPEAKER_03: So uh I have a bunch of friends who are data scientists.
• 15:13
  SPEAKER_03: And to be honest, like I, you know, you want to talk about something that AI is gonna get rid of is data science.
• 15:20
  SPEAKER_03: I mean, it's so don't be a data scientist.
• 15:26
  SPEAKER_03: It's it's my hot take right there.
• 15:28
  SPEAKER_03: Love it or hate it, that's my hot take.
• 15:31
  SPEAKER_03: Um, you know, hate mail, please send it to it.
• 15:38
  SPEAKER_01: That's that's the truth.
• 15:40
  SPEAKER_01: Like that's the truth.
• 15:41
  SPEAKER_01: So it's okay.
• 15:43
  SPEAKER_03: Well I I I truly believe, you know, I I've been in the industry for 20 almost 25 years now.
• 15:52
  SPEAKER_03: Uh, if you count in my military time as well.
• 15:54
  SPEAKER_03: A little more than that, actually, if you count that in.
• 15:56
  SPEAKER_03: Um, I I think really when you get to the core of it, and you know, uh some other people who are out there that I think are you know brilliant in this space, uh Rick Bellata, who is a co-founder of PTC, uh, you know, he talks about all the time.
• 16:12
  SPEAKER_03: And um, I always forget the guy's name from Dirty Jobs.
• 16:14
  SPEAKER_03: Like the core work at the end of the day that you need to know those fundamentals of how a process works will never be replaced, in my opinion.
• 16:23
  SPEAKER_03: Again, hot take if you want, will ever be replaced by AI.
• 16:26
  SPEAKER_03: It can tell you how a process might theoretically work, but you physically going out and troubleshooting with a multimeter in a panel, understanding, you know, how that relay logic is working with something else.
• 16:39
  SPEAKER_03: I would tell every junior engineer, I tell my kids, um, know the fundamentals of everything you're doing because at the end of the day, you know, things like Chat GPT are wonderful, but they're also a crutch.
• 16:52
  SPEAKER_03: Uh, you don't learn your basics, you know, you need to know arithmetic before you move to algebra type of thing.
• 16:57
  SPEAKER_03: Focus on that and and don't take the shortcut around the kind of core learnings you need to have, because at the end of the day, how that electron flows from you know the power source into that relay to then turn on a pump, to do all that stuff, is always going to be part of it.
• 17:16
  SPEAKER_03: Um, you know, uh, even when our robot overlords come to take us over, they still have synchros and servos that are working from all those core principles.
• 17:25
  SPEAKER_03: So, one of those core principles because I feel the software-defined stuff, as we're doing with Sorba, like data science is automated.
• 17:33
  SPEAKER_03: It's math.
• 17:34
  SPEAKER_03: Computers are great at math.
• 17:36
  SPEAKER_03: That is a piece that's gonna go away.
• 17:38
  SPEAKER_03: You're gonna have tools that are gonna help you with that, but you're never gonna have tools come along that are gonna help you understand, you know, when that relay logic doesn't make sense anymore, and you've got to go in there physically and start playing around with things to get that machine or line that, you know, makes popsicles or beer or oil or whatever back up and running.
• 17:57
  SPEAKER_03: So that's what I tell a lot of people is like stay with the fundamentals, you know, learn all the dirty parts of the job.
• 18:04
  SPEAKER_03: That hands-on piece is super important.
• 18:06
  SPEAKER_03: And I and I unfortunately I see a lot of younger engineers who I don't feel that they've become very software-centric in everything they do.
• 18:17
  SPEAKER_03: And because of that, I think we'll see the cascading effect 10 years from now, probably or or longer, where there's gonna be a big gap of knowledge from the guy who, you know, could, you know, theoretically put his hand on, you know, the motor and tell you what's wrong with it, that whole adage, you know, you you pay me for my 30 years of experience, not for the 15 years it took me or 15 minutes it took to fix it.
• 18:40
  SPEAKER_03: Um, I think there's gonna be a frightening gap that occurs because of the software-defined world that everybody's growing up with right now, that physical world, the the actual knowledge of how the things work is gonna you know start to evaporate.
• 18:52
  SPEAKER_03: It's all gonna be very dependent on what what a prompt tells you.
• 18:57
  SPEAKER_01: Right on.
• 18:58
  SPEAKER_01: My question is in AI, what are the skills that keep people from losing their AI job?
• 19:06
  SPEAKER_03: Ooh, the skills that keep you from losing your like losing your job to AI?
• 19:11
  SPEAKER_02: Yeah.
• 19:11
  SPEAKER_03: That that's a great question.
• 19:13
  SPEAKER_03: So I again, I it goes back to that kind of core understanding of a process.
• 19:18
  SPEAKER_03: So um, you know, a lot of customers will ask us specifically, will you help, will you build the models for us?
• 19:25
  SPEAKER_03: And my response is the same, you know, we we'll help you, we'll teach you to fish, we're not gonna fish for you.
• 19:30
  SPEAKER_03: And the reason is it's like you can go into 20 different cookie making plants, or or we'll use something that's even more like just you know, generic.
• 19:42
  SPEAKER_03: Making beer is about the most, you know, you can go to every brewery and they're all gonna have the same processes, right?
• 19:46
  SPEAKER_03: Every brewery is gonna look nearly exactly the same.
• 19:48
  SPEAKER_03: And in fact, the I.O.
• 19:50
  SPEAKER_03: count in almost every brewery without outside the packaging lines and stuff are nearly the same from a beginning to end.
• 19:55
  SPEAKER_03: But yet, how you know an operator runs that, et cetera, understands the process, understands where, you know, this flow meter, how it interacts with some other process or their specific recipes, that to me is how you stay relevant, is understanding truly the physical aspects of that.
• 20:13
  SPEAKER_03: Because while you can ask, you know, uh an AI agent or anything to tell me how something works, um it's not the same plant to plant, you know, even just in the most traditional automation way, you can have two exact same refineries built side by side on literally the same acreage.
• 20:35
  SPEAKER_03: They're both going to behave a little bit different because, you know, a pipe was plumbed just slightly different.
• 20:41
  SPEAKER_03: Um, you know, a pump was replaced and replaced with a different, you know, brand of pump that maybe has a different power curve.
• 20:48
  SPEAKER_03: Like nothing, that knowledge that you have on the operations of it is something that AI can't take in.
• 20:57
  SPEAKER_03: Now, yes, you can build optimization models, you can use tools like ours to help augment that operational piece, but it's still never gonna take away, you know, you, Ally G going in there and knowing exactly like, you know, on Tuesdays when Steve is operating this piece of equipment at 4 p.m.
• 21:16
  SPEAKER_03: when he likes to go, you know, get another cup of coffee, then this is gonna happen.
• 21:20
  SPEAKER_03: I mean, you know, those are those are behaviors in a plant that unfortunately, you know, AI looks at macro.
• 21:26
  SPEAKER_03: These become micro details that help you keep things up and operational.
• 21:29
  SPEAKER_03: And and you know, ultimately, yes, you can feed them into AI models and you'll keep learning and be a little bit better, but it's never gonna replace people in that manner.
• 21:37
  SPEAKER_03: Again, another hot take.
• 21:39
  SPEAKER_03: I feel like I I even have internally to our team where it'll be like, no, no, AI will do that in the future.
• 21:43
  SPEAKER_03: I don't think so.
• 21:44
  SPEAKER_03: So uh personal is I mean, I just saw a video of some lawyers who got like just dressed down in in court because they obviously used some GPT to build their briefs, and the the judges were like.
• 22:00
  SPEAKER_03: Like you, this isn't even real case law.
• 22:03
  SPEAKER_03: It's made up the the chat engine made it up.
• 22:08
  SPEAKER_03: Um, and and so I I you know there's I I feel like that's why we'll we'll always be in the loop.
• 22:15
  SPEAKER_03: Uh there's that expertise especially in making stuff, making it's just such a tangible thing.
• 22:21
  SPEAKER_03: I mean, they've been pulling grease out of the ground, the oil and gas industry for 100 plus years.
• 22:26
  SPEAKER_03: I mean, at this point, you know, you're still using, I mean, uh, you're down in Houston now, right?
• 22:30
  SPEAKER_03: Have you gone down to the natural history museum there and seen the fourth floor?
• 22:35
  SPEAKER_03: I mean, that it's it's the most rudimentary stuff.
• 22:38
  SPEAKER_03: And they're like, AI is gonna fit, like they've been pulling Greece out of the ground for a long time.
• 22:43
  SPEAKER_03: It's it's a plunger, it's literally the most basic physics stuff, and they're still using that same technology forever.
• 22:54
  SPEAKER_03: So that's why I just I you know I don't worry too much about the uh the robot overlords.
• 23:01
  SPEAKER_01: Okay, how about how about digital twins?
• 23:03
  SPEAKER_01: Can you tell me like what's a real digital twin versus a bullshit one?
• 23:07
  SPEAKER_01: Because there are BS ones.
• 23:10
  SPEAKER_03: Yes, I agree with you.
• 23:11
  SPEAKER_03: I and I think maybe there are a lot of BS ones.
• 23:15
  SPEAKER_03: I think also, too, uh the definition is so broad that's it.
• 23:20
  SPEAKER_01: Maybe that's the real the real it.
• 23:22
  SPEAKER_03: Well, no, it's both.
• 23:24
  SPEAKER_03: But I think it's real, but so there's like the I think some people look at a digital twin as like I have this um replication of my plant that I can visually see, and I'll have like indicators on it.
• 23:38
  SPEAKER_03: Uh a good buddy of mine at a different company uh called Antea, they have this beautiful digital twin that they build for reliability modeling.
• 23:47
  SPEAKER_03: And um uh Floyd Baker and his team have this incredible, like it's beautiful, but I look at that from my perspective, and you're trying to do real-time automation and trying to, you know, what if scenarios, etc.
• 24:00
  SPEAKER_03: That's kind of the other side of the coin of digital twins.
• 24:03
  SPEAKER_03: That's I'm looking at time series data.
• 24:05
  SPEAKER_03: I just need a chart.
• 24:06
  SPEAKER_03: I need to know if I drop the speed down on this or I add, you know, extra amount more of this chemical into this process, what is my outcome gonna be?
• 24:15
  SPEAKER_03: And am I gonna get more yield, yes yield, and kind of see that prediction uh or interaction between it and just a time series data, not you know, a visual representation of your plant.
• 24:28
  SPEAKER_03: So to for Sorba, the digital twin piece is really around that that second part.
• 24:33
  SPEAKER_03: It's how do we take the data we have now, relive data from our historian or live data or from our historian, build out a model that replicates the current operation of what we consider like to be most performant?
• 24:47
  SPEAKER_03: And then let's let's you know turn the dial down on this and see what happens.
• 24:51
  SPEAKER_03: Or, you know, we want to test out what, you know, we have a 15-horse motor in there now.
• 24:56
  SPEAKER_03: What does 20-horse get us?
• 24:57
  SPEAKER_03: You know, how does that change the behavior going through?
• 25:01
  SPEAKER_03: To me, there's practical applications for both.
• 25:05
  SPEAKER_03: Um, you know, if you're an operator, maybe the visualization piece is better, but from a process engineering and optimization standpoint, um, the time series data that that you know, what is scenario piece on digital twins is far more valuable.
• 25:19
  SPEAKER_03: Because at the end of the day, the operator just needs to know, like, hey, digital twin is telling me, or the APC or the anomaly detection is telling me blank, how do I need to respond appropriately?
• 25:29
  SPEAKER_03: So we, you know, increase yield, reduce scrap, you know, whatever, whatever your KPI is that you want to make sure that that data is informing to your operator on the line.
• 25:38
  SPEAKER_01: Do you have any surprise information?
• 25:41
  SPEAKER_01: Tell us about you again.
• 25:43
  SPEAKER_03: About me again?
• 25:44
  SPEAKER_03: Uh well, I I surprise information.
• 25:47
  SPEAKER_03: Um, you know, there's there's so many things.
• 25:53
  SPEAKER_03: Uh, you know, uh Yeah, I mean, but before earlier, I think, you know, back when I joined the Navy, it was uh, you know, much to my my parents' chagrin, but my roommate, you know, him and I would sit around watching Jerry Springer, and I was doing very poorly at school.
• 26:07
  SPEAKER_03: You know, that's always a good little nugget of fun.
• 26:10
  SPEAKER_03: I think, you know, um a lot of the uh uh a lot of the you know things that I I think about when it comes to like the future of all this is, you know, one, we need to get more people into this industry just as a whole.
• 26:30
  SPEAKER_03: You know, I this I don't know this is really surprise type stuff, but this is just more you know the the what keeps me up at night.
• 26:38
  SPEAKER_03: Um I I think there's a a lot of fear mongering around stuff.
• 26:42
  SPEAKER_03: And I also think that people need to get back into the dirtier jobs.
• 26:45
  SPEAKER_03: Uh, you know, remembered his name, Mike Rowe.
• 26:48
  SPEAKER_03: You know, that there is a truly aspect.
• 26:51
  SPEAKER_03: I mean, I told my two older kids, you know, um, and I've told my youngest, my youngest one as well, there's a little age gap in them, but um, go into the trades.
• 26:59
  SPEAKER_03: I mean, I I it's the one thing I remember when I became got out of the military and I went to work for Siemens.
• 27:05
  SPEAKER_03: I remember sitting there, I worked to a lot of electricians and pipe fitters uh all the time.
• 27:09
  SPEAKER_03: I mean, I was like side by side with them, would have lunch with them every day.
• 27:13
  SPEAKER_03: And I mean, one, I was shocked.
• 27:15
  SPEAKER_03: I I had no idea my dad would never ever have my dad was an accountant and a CFO and that type of stuff.
• 27:21
  SPEAKER_03: And never in a million years would he suggest going into the trades.
• 27:24
  SPEAKER_03: And I've told all of my kids, go into the trades.
• 27:27
  SPEAKER_03: Like it is just so needed.
• 27:30
  SPEAKER_03: And and you learn so much, and they make great money.
• 27:35
  SPEAKER_03: Um, you know, so go be an electrician, go be a pipe fitter, go do that stuff.
• 27:40
  SPEAKER_03: Because even if later on down the road you want to get into software, you want to get into being a controls engineer, you want to do that, you really have, I think, far more core skills around it.
• 27:50
  SPEAKER_03: So um, yeah, totally didn't answer your question there.
• 27:53
  SPEAKER_03: But that's kind of like, you know, my my hot take.
• 27:55
  SPEAKER_03: I just look around the industry right now, and too many people are kind of going into the soft world of software and you know, this like I want to be in an air conditioned office and have a ping-pong table and snacks.
• 28:06
  SPEAKER_03: And real the reality is we need more people out there willing to get dirty, you know, um, going out there and troubleshooting panels and getting all those PLC fives out of there or, you know, old Siemens gear and stuff, and and getting, you know, you know, getting that infrastructure built out so that we can continue forward uh in this space and that AI becomes applicable there.
• 28:32
  SPEAKER_01: So I'm a millennial and I grew up being told that if you don't go to college, you're wasting your basically your abilities.
• 28:44
  SPEAKER_01: And I know that that's what all of my classmates were told because that's why we have the giant skills gap that we have today.
• 28:53
  SPEAKER_01: Do you think we're finally getting over that?
• 28:55
  SPEAKER_01: Or do you, I mean, do you see any movement in it's okay?
• 29:00
  SPEAKER_01: I mean, obviously you as a dad are doing it, which is really awesome.
• 29:05
  SPEAKER_01: Um, but outside of you as a dad, like are people coming to terms with the fact that we do need the trades and that we need to stop saying that the trades is not good enough for you.
• 29:21
  SPEAKER_01: I mean, that's where that's what they were saying.
• 29:22
  SPEAKER_01: They were basically saying if you have good grades, you're too good for the trades.
• 29:27
  SPEAKER_01: And I think that's insane.
• 29:28
  SPEAKER_01: What do you think?
• 29:29
  SPEAKER_03: I think 100% agree.
• 29:30
  SPEAKER_03: I I think well, so I don't know if there's a shift happening.
• 29:34
  SPEAKER_03: So I'm a Gen Xer.
• 29:36
  SPEAKER_03: Um I I, you know, college was just kind of, you know, my parents' college was not even an option.
• 29:45
  SPEAKER_03: And of course, I joined the military and they weren't happy about that.
• 29:48
  SPEAKER_03: So to me, I think one, college has gotten really skewed towards almost hyper-specific behavior or hyper-specific training.
• 29:57
  SPEAKER_03: Oh, almost as a trade in and of itself.
• 30:00
  SPEAKER_03: There is actually this really uh uh interesting company.
• 30:03
  SPEAKER_03: You know, you travel a ton, I travel a ton, you're on flights, you have conversations with random people that you're sitting next to in chairs, um, flying back home and this time I were having a conversation.
• 30:15
  SPEAKER_03: And he made his comment about like how most entrepreneurs uh or uh or uh CEOs have actually liberal arts degrees, you know, history and you know, like soft skills or soft uh sciences, you know, like psychology, sociology, et cetera.
• 30:31
  SPEAKER_03: And that level of critical thinking that you learn in a liberal arts degree is being lost in all of these like hyper focused, you know, like even in computer science and even in engineering, where you don't develop a lot of the other skills that are required to be in, you know, a professional setting.
• 30:51
  SPEAKER_03: And like they have this whole thesis around a business where if you go to like major engineering firms and you ask them, like, you know, how do you hire a project manager, which requires both soft skills and tactical skills?
• 31:02
  SPEAKER_03: You know, how do you oh, we go to Purdue and we find somebody with a 4.0.
• 31:06
  SPEAKER_03: And you're like, cool.
• 31:07
  SPEAKER_03: Like I know a lot of people with 4.0s from Purdue can barely hold a conversation, let alone eye contact.
• 31:14
  SPEAKER_03: Um, you know, how is that applicable to a project management role?
• 31:21
  SPEAKER_03: And so I think one college has been distorted in a way where now people are getting almost too specific, almost like a trade degree in a way.
• 31:30
  SPEAKER_03: They're they're losing a lot of that general critical thinking that goes on in a university, um, to, you know, just think outside of, you know, one plus one equals two, but just really like, you know, here's history, here's how we're extrapolating things and ideas and thoughts, and even just writing a paper.
• 31:48
  SPEAKER_03: Um, so that's that's one part I think is, you know, different about college now that I don't know if people see some of the value going, but I don't know if trades have really been a replacement for that.
• 32:00
  SPEAKER_03: Um, I don't see it rapidly changing.
• 32:03
  SPEAKER_03: I've I need to go catch up with a few of my buddies in the trades here, um and and kind of find out if they've seen more or less people come in.
• 32:11
  SPEAKER_03: But I I do know this.
• 32:13
  SPEAKER_03: I was um going after data centers about um, gosh, that was before it became even more crazy.
• 32:21
  SPEAKER_03: Um and there was deals where like electric electrical teams uh they bid on a data center hall, and you would have to get a quote from these electrical contractors that were absurd and you know, bonuses every day, you know, higher per diem rates, you know, uh 30 hours and then overtime.
• 32:39
  SPEAKER_03: And these crews would swap because somebody else would just come in and there was so few electricians to win another deal, you just sweeten the deal, and this whole crew would just shift because that's how it worked.
• 32:52
  SPEAKER_03: There was just not enough people.
• 32:53
  SPEAKER_03: And these they were coming, you know, there'd be a project in Des Moines, and you'd have people coming in from California and Chicago because there just wasn't enough electricians to handle a$2 billion data center being built.
• 33:06
  SPEAKER_03: So I I would I'm gonna go on a limb.
• 33:10
  SPEAKER_03: I don't know the data, but I'm gonna go on a limb.
• 33:11
  SPEAKER_03: No, it hasn't changed.
• 33:13
  SPEAKER_03: It doesn't seem to be rapid enough to get more people in the pipeline.
• 33:16
  SPEAKER_03: And I I know I saw an ad the other day of uh IBEW like basically begging people to come and apply.
• 33:24
  SPEAKER_03: Um, and electricians make good money if you haven't looked.
• 33:26
  SPEAKER_03: Yeah, you know, after you get out of your apprenticeship, you're making like 45 on the check and you have benefits and all this.
• 33:32
  SPEAKER_03: It's crazy.
• 33:33
  SPEAKER_03: Like, you know, not to sound like I'm uh, you know, pumping going into the trades, you know, too hard, but it is a great gig.
• 33:40
  SPEAKER_03: I I mean there are some days I think about it.
• 33:42
  SPEAKER_03: Like, I love I put conduit in my own house just for fun because I saw my buddies doing it.
• 33:46
  SPEAKER_03: And I was like, I can bend pipe.
• 33:48
  SPEAKER_03: This is fun.
• 33:48
  SPEAKER_03: So, you know, it's um I think the trades are are woefully underserved.
• 33:52
  SPEAKER_03: And I don't think there's a lot of promotion, honestly, from the states and the federal government to really get people going that direction.
• 34:00
  SPEAKER_03: I think that there needs to be more involvement there.
• 34:02
  SPEAKER_03: And even I think it was the CEO of Ford was even talking about how they need to push more and more people into the trades because they're they're struggling to find people to help manufacture vehicles in in the U.S.
• 34:14
  SPEAKER_03: So I think there needs to be a bigger concerted effort around that.
• 34:17
  SPEAKER_01: If you're in the trade, what advice do you give them?
• 34:20
  SPEAKER_03: If they're in the trades right now, like what would you give them or if they want to go there?
• 34:24
  SPEAKER_01: They're already in the trades.
• 34:25
  SPEAKER_03: Oh, I mean, I had uh a buddy who uh unfortunately I worked with one of the startups who unfortunately passed away.
• 34:33
  SPEAKER_03: Um sorry I mean, just be it, yeah, be it be an entrepreneur.
• 34:38
  SPEAKER_03: I mean, go after it.
• 34:39
  SPEAKER_03: I mean, this guy, uh first off, he was one of the kindest souls.
• 34:43
  SPEAKER_03: Yeah, he was one of the kindest souls I've ever met in my entire life.
• 34:47
  SPEAKER_03: He was so just so generous with his time.
• 34:49
  SPEAKER_03: But he started, he was a pipe fitter, uh, no college education, started his own pipe fitting company, sold the thing, I think, to MCor for like$25 million.
• 35:01
  SPEAKER_03: Like, just go do it.
• 35:03
  SPEAKER_03: Like, you have a skill that is all like you write Python, you can see that your job is at risk.
• 35:10
  SPEAKER_03: Um, I think too, I I'm gonna misquote, I'm gonna, I'm gonna misquote Rick here, but you know, if if you work in bits and bikes, you should be in trouble.
• 35:19
  SPEAKER_03: But if you move electrons, you're probably fine.
• 35:22
  SPEAKER_03: Um if you're if you're if you know how to to run wire from point A to point B, you're gonna have a job for as long as I'm gonna be alive, at least, and probably as long as my kids are gonna be alive.
• 35:34
  SPEAKER_03: But if you're in, you know, so go be an entrepreneur, go start your own company, go get other people into that field.
• 35:41
  SPEAKER_03: There's plenty of money to be made there.
• 35:43
  SPEAKER_03: You will you will never be without a job.
• 35:46
  SPEAKER_03: It it is it is a place where I think that you can thrive.
• 35:49
  SPEAKER_03: And I mean, as a person who works way too many hours, there are days that I wish that somebody told me it was time to punch the clock and go home and pull out a fishing pole because you had nothing else to do because your shift ended and they make a decent.
• 36:04
  SPEAKER_03: I mean, I remember the day my buddy, I was an engineer at Siemens, I was a PM actually at the time, and my buddy told me he just turned out as a journeyman and he told me how much he made.
• 36:12
  SPEAKER_03: I was like, you work like 20 hours less a week than I do, and you make more money than me.
• 36:17
  SPEAKER_03: This is ridiculous.
• 36:19
  SPEAKER_03: So it's I highly recommend doing it.
• 36:22
  SPEAKER_03: I I uh if I told 18-year-old me, I'd probably do that than join the military.
• 36:26
  SPEAKER_03: Although the military was a blast in its own way.
• 36:28
  SPEAKER_01: So that's great.
• 36:29
  SPEAKER_01: If people want to find you, where can they find you?
• 36:32
  SPEAKER_01: Can they find you on LinkedIn?
• 36:35
  unknown: Yep.
• 36:35
  SPEAKER_03: Go to LinkedIn.
• 36:36
  SPEAKER_03: Um, I'm on there.
• 36:38
  SPEAKER_03: Uh Brian with the Y, Psychen, T H Y K-N.
• 36:43
  SPEAKER_03: Um, or you know, go to Sorba.ai and um, you know, hit the contact us button and um it'll hit me or one of my salespeople.
• 36:51
  SPEAKER_03: Um, or you know, shoot me an email uh at bziken at sorba.ai.
• 36:57
  SPEAKER_03: And uh yeah, love to chat about all these things and whatever else.
• 37:02
  SPEAKER_01: That's great.
• 37:03
  SPEAKER_01: Thank you very much for being a guest today.
• 37:05
  SPEAKER_01: And uh, we've had a really good episode.
• 37:08
  SPEAKER_01: I really like this episode, actually.
• 37:10
  SPEAKER_03: Well, good.
• 37:10
  SPEAKER_03: I'm super excited to be down at OT Skaticon.
• 37:13
  SPEAKER_03: Uh it's my first time going.
• 37:15
  SPEAKER_03: I am super stoked.
• 37:16
  SPEAKER_03: I'm uh pulling together all my slides right now to talk about predictive maintenance and and how you can solve a lot of very interesting problems right now.
• 37:26
  SPEAKER_03: And and of course, you know, how it's uh SME focused.
• 37:29
  SPEAKER_03: So yeah, I hope to see people there.
• 37:31
  SPEAKER_03: Um, I will be there speaking hopefully eloquently and with some information that is valuable to the people.
• 37:38
  SPEAKER_03: Um, but uh yeah, super excited about that.
• 37:40
  SPEAKER_03: I was actually just booking my hotel rooms for that trip, making sure I had everything dialed in for being down there.
• 37:47
  SPEAKER_01: That's great.
• 37:48
  SPEAKER_01: Yeah, I remember trying to learn about predictive maintenance myself.
• 37:52
  SPEAKER_01: And I remember there was we had a downed chiller, and it was really easy to tell that the chiller was down, but we just had to have that data.
• 38:02
  SPEAKER_01: But it was it was data that wasn't impossible to get.
• 38:05
  SPEAKER_01: I remember I remember that being like a really important like part of my career where I could see really obviously that like there was some data that just we needed to know, otherwise we were gonna lose production.
• 38:19
  SPEAKER_01: And I thought, you know, wow, like is it really like why is there nothing else built for this?
• 38:26
  SPEAKER_01: Because it wasn't.
• 38:27
  SPEAKER_01: It was just me saying, okay, if the temperature goes this low for this long, that means we lost our heating to our pond.
• 38:37
  SPEAKER_01: And and and or whatever it was that we were losing.
• 38:40
  SPEAKER_01: And, you know, nobody cared, but everybody cared about the the outcome.
• 38:45
  SPEAKER_01: So I'm excited to see kind of what happens with you know sorba and what happens with all of the AIs that are uh that are happening right now, because you know, it's changing the world, I think.
• 38:57
  SPEAKER_03: You're spot on.
• 38:58
  SPEAKER_03: And I think the bet the most important part about it too, though, is is there's so much data out there right now.
• 39:04
  SPEAKER_03: There's so much data.
• 39:05
  SPEAKER_03: I mean, insane amount of data.
• 39:07
  SPEAKER_03: I was actually told like uh, you know, like MQTT, one of the biggest users of MQTT data is actually Facebook.
• 39:13
  SPEAKER_03: You know, so you think about that like in terms of Really?
• 39:16
  SPEAKER_03: Yeah, so like I didn't know that.
• 39:18
  SPEAKER_03: And I didn't somebody please fact-checked me on that.
• 39:20
  SPEAKER_03: But I remember somebody telling me, and I was like, wait, wait, why like for their messaging because it's a very efficient protocol, right?
• 39:26
  SPEAKER_03: So there's just bajillions and bajillions of bits of data floating around out there.
• 39:30
  SPEAKER_03: But to your point about like solving a chiller maintenance issue, you still need to know you as an expert on that chiller, how it's configured.
• 39:36
  SPEAKER_03: Is it a duplex, tripex pump?
• 39:38
  SPEAKER_03: Is you know, you using a cooling tower, using a you know, how that whole thing is configured determines how it fails and how you create context and how you predict whether a failure is gonna happen.
• 39:48
  SPEAKER_03: So I think how you build AI models out and how you leverage your knowledge is the key piece.
• 39:54
  SPEAKER_03: That's that's one of the biggest parts of what I'm gonna talk about is there's no such thing as one size fits all model.
• 40:00
  SPEAKER_03: And if when companies sell you that idea of like we have these bajillion pre-trained models, it's not how your plant runs.
• 40:06
  SPEAKER_03: That's just how a a motor or multiple motors were used to train that data run.
• 40:11
  SPEAKER_03: And uh I think that there's a very important piece to know that like you as this to me, inside that plant have a very specific knowledge set that needs to be embedded into those predictive maintenance models to truly be effective at uh creating that value.
• 40:26
  SPEAKER_00: Well, thank you very much for being on the case.
• 40:29
  SPEAKER_03: No, thank you.
• 40:29
  SPEAKER_03: I super appreciate it.
• 40:31
  SPEAKER_03: It's been a blast.