Episode Transcript
[00:00:00] Speaker A: Most digital transformation and AI projects in manufacturing don't fail because of the technology.
They fail because they were never built for the reality of the plant floor in the first place. Clean models meet dirty data, algorithms meet operators, and suddenly everything that worked in theory doesn't.
So what does it actually take to make AI work in a real factory? That's what we're getting into today with Josh Pino, a professional engineer who's been working at the edge of industrial AI in some of the toughest environments out there.
This is Control Alt Manufacturing.
Hello, hello, hello, everybody. Welcome back to the Control Alt Manufacturing podcast, Resetting and Rethinking Manufacturing, where we are helping to explore some of the people, technologies and strategies that are driving the digital transformation of manufacturing. I am Feng one. Gary Cohen, my other host. Over here, Stephanie Neal. Hey, Stephanie, how you doing?
[00:01:03] Speaker B: I'm thing two.
[00:01:04] Speaker A: I guess thing one thing too. It's a whole Dr. Seuss thing.
[00:01:10] Speaker B: Yes, I'm happy to be here. And I don't know when this recording is going out, but it is the first day of spring, so I am so delighted to be here on this beautiful first day of spring. It's nice out. What are we doing recording a podcast?
[00:01:26] Speaker A: I was just gonna say you're delighted to be here, but you're actually lying to us. You're not delighted to be inside right now. You want to be. We both live in cold weather places. She's in the B.O.
i'm in the Chicago area. It's like 70 degrees and sunny today, and both of us, like little kids, are like, I want to go out and play.
[00:01:42] Speaker B: Yeah, exactly.
You know what?
[00:01:45] Speaker A: I did.
[00:01:45] Speaker B: But I did sneak out earlier. And so this is a Get to know Gary question.
[00:01:50] Speaker A: Oh, good.
[00:01:51] Speaker B: Before we start to talk about the important stuff, we got to get to know Gary. So you're going out for coffee because you need a little pick me up in the middle of the day. You're getting ready to record a podcast and you see that there's a Dunkin and a Starbucks right next to each other. Which one do you go to see?
[00:02:08] Speaker A: I am. I am going to blow up your entire question just from the beginning. I don't drink coffee.
[00:02:13] Speaker B: I knew it.
[00:02:14] Speaker A: Yeah. So if I were going to go into one or the other, I.
I don't even know. There's. There's more. I mean, you live in Boston. There's more Starbucks than there are Dunkin's around here.
[00:02:25] Speaker B: No, no, no, no, no.
[00:02:27] Speaker A: But I know that the answer you want is Dunkin, so I'll say I'd go in and get a donut, but
[00:02:33] Speaker B: oh, get a dunkin.
[00:02:34] Speaker A: Oh, there you go. Yeah, I actually do have a dunkin within like a four minute walk of my house. So I could actually do it. But yeah, I take my caffeine.
[00:02:43] Speaker B: How do you survive when you.
[00:02:45] Speaker A: I take my caffeine in Coca Cola form.
[00:02:47] Speaker C: Oh yeah.
[00:02:49] Speaker A: So usually when I'm doing one of these, I've got like a Coke sitting over here to wake me up a little bit. But yeah, I just, I am a really very stubborn human being and, and I never liked the taste of coffee when I was younger and all my friends, you know, in college and where, wherever else were like, you'll get over that, you'll start to like it. I was like, why am I gonna keep drinking something I don't like until I start to like it? And so I just never, never. It's not like it's good for you. And so I never picked it up. Now I know it's.
[00:03:18] Speaker B: You know what though? I, I started early because so my parents, my family is originally from Rhode island and so in Rhode island they have this thing called coffee milk. Now it's not necessarily like filled with caffe is filled with sugar. And so like the kids are raised on this like coffee. Like it's like a syrup that you put into your milk, like a chocolate milk, but it's coffee milk. So we get acquainted with the taste of coffee at a young age over here in the.
[00:03:47] Speaker A: Yeah, I will say when I was in college, I remember people who'd pull all nighters and we'd go to IHOP or wherever college students go. I think it was an IHOP and they would have, you know, endless coffee and at some point somebody convinced me, you should drink this, it'll keep you up and you'll be great.
And I did. And as a non coffee drinker, I just got, I got so like jittery and shaky and couldn't concentrate and was like, yeah, this is not for me.
[00:04:13] Speaker B: So Coca Cola it is.
[00:04:15] Speaker A: Coca Cola it is. Yes, that's right. We.
That one. I started young, so.
So there you go. You got to know Gary and I refused to answer your question properly.
I told you I was stubborn.
[00:04:28] Speaker B: Yeah.
So that's what we learned about you today.
[00:04:31] Speaker A: That's exactly what we learned. You'll learn more and more. Yeah, I know. I wanted to ask you a food question, but now I'm going to save that till next time.
So let's talk digital transformation and AI. It's what we do here.
So, you know, week to week, we're talking smarter factories, connected systems, this sort of avalanche of data that manufacturers have to work with right now. And what we're going to be talking about with Josh today, obviously now you've got AI layered on top of all of that.
And so you would think if you weren't in manufacturing, that digital transformation would be working great. Right now, it all makes sense. We've got more data and tools to analyze it, and computing power is as great as it's ever been, and AI is making things more accessible.
But, you know, if you talk to people who actually work inside of plants, that story gets a little bit more complicated. It's not that all of the projects don't deliver value, many of them do, but.
But a lot of them are still stalling out, or they don't scale, or.
The classic engineering problem is they solve a problem and they create two more problems.
And I think a big part of that, Stephanie, comes down to what I talked about in the intro, is that disconnect between how systems are designed and then what the actual plant floor environment looks like out there, which is far from a clean, controlled system.
[00:05:58] Speaker B: Yeah, no, I agree, but I think it's even more than that. It's just the amount of data that's being generated, and people are drowning in that data, and they don't know how to make sense of it. It's just a lot of noise.
And I have to tell you that I pulled this off of Josh's LinkedIn page, but I just love this. So we're gonna talk more to him about this, but his role is to serve as the adult in the room for innovation, ensuring technology investments solve business problems rather than create technical debt. I love that. Like, you need somebody to come in and just be like, all right, we're not gonna do it this way, we're gonna do it this way, and here's why. So I think that there's just a lot of confusion out there because of what you talked about, the different systems, the amount of information that's being generated. There's so many different paths to take.
You need someone to guide you down the right one well, and I think
[00:06:59] Speaker A: there's also kind of a human factor in that, which is a lot of times on the plant floor, you've got people who have been doing this job for 20 or 30 years. They know their process inside and out.
You know, they. They've got instincts that are honed in the industry that they're in. And then a new system comes in. You know whether that's AI or something else and it's not automatically trusted and it doesn't really, you know, line up with what they see out in the real world. And so there' this. There can be a tendency, even though the technology is sound, that somebody on the plant floor can be like, I don't want to use that, or can push back on it. So I think it's true. You do need that adult in the room, that person who can tell you why this should work and how it should work. That's the beauty of system integrators. You know, we talked about this with, with Adrienne Meyer in a previous podcast who is the head of CSIA of like, sometimes you just need a human to come in and actually unpack what works and what doesn't work and how all of these systems fit together and kind of bridge that gap between the theory that's out there and the reality of a plant floor.
[00:08:12] Speaker B: Absolutely. And so let's bring that human in that. Yeah.
[00:08:15] Speaker A: That's why we've got Josh Bino here. He's going to be talking to us about, you know, he is someone who has been working at that intersection between advanced technology and real world manufacturing.
He is a. Josh Pino is a fractional technical director with a leading voice in industrial AI. He's got a background in heavy manufacturing and engineering leadership, specifically in the 1500 degrees Celsius world of glass, which I'm sure puts some things to the test. Josh specializes in bridging that gap, as you said, between pristine academic theory and messy factory floor reality. Josh, we are super happy to have you here with us.
[00:08:51] Speaker C: Yeah, thanks for having me.
[00:08:52] Speaker B: We should also say that Josh is helping shape the next generation of automation engineers as a part time professor at the University of Toledo.
So thank you, thank you for that.
[00:09:03] Speaker C: Yeah, it's rewarding. It's a bit rewarding.
So I enjoy it.
[00:09:09] Speaker A: Every time we talk to somebody we're talking about the skills gap and everybody's reaching retirement age and there aren't enough young engineers coming up behind them. So we're really leaving it to you to create this next generation of engineers and solve this problem for the entire manufacturing industry. Industry. So thanks.
[00:09:24] Speaker C: Yeah, yeah, I think you're right.
There is for sure a skills gap and there's a big wave of skills gap coming as well that needs to be addressed.
[00:09:37] Speaker B: So, so let's get into it a little bit more and we'll talk a little bit about that next generation skills gap later on in the podcast. But give us A little background. Josh, you spent your career in manufacturing. Like, tell us a little bit about that. And what led you to focus on applying AI in environments like glass production?
[00:09:59] Speaker C: Sure.
I mean, so yeah, I spent a career in manufacturing, primarily automation and control and did, did a bit of work in R and D, did a bit of work in factory floors.
So I've been in sort of all aspects of the manufacturing. But the biggest gap that I've seen recently, which is how I've gone down the path that I have, is I. The. There's a divide between the technical, the technical world and the plant floor world. And there's lots of startups and academic endeavors and academic research that's doing some really great things with this stuff. But the reality is until you have, you know, it gets deployed in the, on the factory floor, you're not going to realize and see, you know, what it can deliver. So, yeah, I sort of position myself to be that technological gap or that, or that filler between the technology and the reality of the plant floor.
[00:11:06] Speaker B: So are you.
[00:11:07] Speaker C: Yeah.
[00:11:08] Speaker B: Are you specializing in like, glass and that. In that type of like, segment or.
[00:11:15] Speaker C: Not necessarily. I have spent a good portion of my, I mean, I've been in manufacturing for over 25 years now, but I've spent a good portion of that in the glass industry. So that's an area that I know quite well and it's an area that actually needs a lot of development, a lot of innovation and a lot of catching up in the, in the digital manufacturing world.
[00:11:41] Speaker B: So. So you go in there and you're sort of, you know, clearing things out, like get, creating some clarity. Right. So kind of looking at, maybe doing an audit, looking at what they have, like helping them make vendor decisions and those, those are the types of things that you're going to walk into a facility and do.
[00:12:04] Speaker C: Yeah, yeah, I do actually. I have like. So I kind of do. There's three focus areas that I kind of am focused on right now. And 1. Is that what you just said, being a technical leader or as you guys put earlier, the adult in the room in manufacturing.
And so, yeah, I would go in and evaluate where they're at, where they maybe need to be and help them create the roadmap to get to where they want to go, help them vet some of the vendors, help them understand what the technology is capable of.
You know, their, their software vendors are telling them that they can solve all their problems and you know, for, for a couple million dollars, it won't have any more problems. But that's Obviously not true. So it's helping them sort out what, you know, what works, what doesn't work.
That's one pillar. The other thing is kind of on the other side of that too. I also am working with some software startups to help them make sure that their tools can deliver the, what's needed in the platform as well. So I have a fractional, you know, I do the fractional technical direction directorship or the fractional technical executive that I like to call it. So I.
And so I'll do that for software companies that want to deliver into manufacturing and they don't know, they don't know manufacturing, they don't understand the realities of manufacturing. So helping them to guide, or help guide them to make their tools to be suitable for the harsh, you know, the harsh environments that they're going into that they probably don't understand very well. And then the third pillar is kind of in the academic world too. I also work with a consortium in Northwest Ohio here, the Northwest Ohio Innovation Consortium, and also with the University of Toledo and Bowling Green State University here, also in Northwest Ohio, to try to bring their academic R and D into the commercial reality of manufacturing. So serve as that bridge. So in all cases it's some kind of bridge between a technical aspect and manufacturing. So that's where I found myself.
[00:14:21] Speaker B: So that bridge is interesting and it requires a human in the loop. And you have said that digital transformation fails when companies treat the factory like a math equation. What do you mean by that?
[00:14:35] Speaker C: Well, what I see a lot is people, you know, companies want to deploy. Excuse me, what I see a lot is companies want to deploy AI or some of digital intelligence without even going onto the plant floor and talking to the people that are the experts in the industry.
And that's where it tends to fall apart because it's a people problem too. And you have to get the people that understand the manufacturing process involved the whole way as you do it. The first question that people should be asking when they're thinking about digital transformation or deployment of AI is what do, what are the problems? What are the actual problems on the plant floor? What are the actual problems that the operators are having? Go talk to them, ask them specifically what can we do to actually make your job easier and solve those problems first? And then you're going to get buy in every time from the, from the, from the operations team and the experts.
[00:15:44] Speaker A: I want to drill a little deeper into that sort of the people problem that you were just talking about. Kind of ask about some of the cultural issues on the plant floor.
If there's an easy answer to this, and I'm not sure there is, what is actually happening on the plant floor? What are you seeing? Whether it's culturally or operationally, when digital transformation initiatives break down, what are some of the main causes of that?
[00:16:11] Speaker C: Maybe trying to solve problems that aren't problems yet. I see that sometimes trying to solve problems that aren't quite there. Or maybe the people in the IT departments think they're problems and they're not plant four problems. But I think primarily it has to do a lot with the. A lot of it has to do with the skills gap that we have. And I was just reading your guys. Well, not your guys's, but you guys are affiliated with the plant engineering, the salary survey and the age, the average age of people in manufacturing. And you look at that pie chart that's on one of the early slides and shows I think it's like 12% are under 30 and most people are 40 and above. And like at least half it was, more than half were above 50.
So our manufacturing is. And that reflects what I'm seeing also. So our manufacturing world is losing skills and losing talent and all of the expertise lives with that older generation. This is what, you know, has been commonly called as the silver tsunami. In 10 years from now, we're going to be losing half of our manufacturing workforce.
And what we should be really focused on is trying to get the knowledge and skills from the people that have been doing this for 30 years and digitizing that rather than just trying to solve, you know, the everyday problems. The biggest problem is the one that's in front of us is we are going to lose skills. And we need to find ways using technology to capture that data from the experts, capture that skills and that knowledge, and then use that to digitalize our factories.
[00:18:00] Speaker A: That makes perfect sense. I want to ask a little bit about AI. As far as I assume that's one of these technologies going to help us.
But why does traditional AI tend to create either fear or resistance among operators, among people who are actually working in manufacturing?
[00:18:19] Speaker C: Well, I mean, quite honestly, I think you read the headlines every day and you hear about more and more layoffs across the tech world. And companies are saying that it's related to AI taking those jobs. I'm not so sure that it's that or just a correction in the market, but so people fear that it's going to take their job and it's sometimes often being sold as something that's going to take their job. So I think That's a bit of reluctance to it as well. Where it shouldn't be viewed that way. It should be viewed as an enhancement. It allows them to do their job more effectively rather than replace them.
It's an iron man suit, if you will, for the operators.
[00:19:04] Speaker B: So just like as a follow up to that, you advocate for machine teaching. How is that different from how most companies approach AI? Josh?
[00:19:16] Speaker C: Yeah, so there's a school of thought and I can break down industrial AI into three pillars, and I'm sure that there are three pillars that people have heard, I've talked about them. But the traditional AI is the more the pattern matching type stuff and then there's generational, generative AI. But when we talk about machine teaching, we're talking about the autonomous AI for the factory floor. The stuff that people say, well, you, you can't, you can't put AI onto the factory floor. Well, not entirely, but machine teaching, what that really means is teaching machines to think the way the humans do about a process.
So it's machine, you know, it's, it's a, it's a sort of a subset of machine learning. But instead of just letting the algorithms learn from the data, you're actually teaching it how to think. This goes back to the DeepMind project, the Google DeepMind project, where they taught it to, you know, where they taught AI and AI agents to be able to play the game go, and then they taught to play the game chess. So basically they taught it the rules of the game, let it go out and figure out what are the best strategies. Or you can apply that to manufacturing as well by teaching the machines, the AI machines to teaching the AI models what are the limits of what the process can do and how does the process behave. And then let it learn on a simulated process. You develop a simulation of the process that you're trying to automate and then you let it go and learn and teach itself how to do that in over hundreds of thousands of simulations. If that's what it takes. It's about a better way to say it is.
It's almost like bringing in an apprentice instead of the human apprentice, you're bringing in the AI apprentice and you're letting it work alongside an expert, a guy that's been doing it for 20, 30 years.
And it's learning how that guy does his job, how he understands the process. And it's. So it's the apprentice that you're bringing in. Except we don't have that many more apprentices. The industry's short on apprentices, right? Now very short on apprentices. So can we bring the machines in? Can we bring the AI in to learn how to do it that way?
I was, I had a successful project doing this actually a couple years ago working with, actually working with a company called a mesa.
They weren't called that at the time. They went through a name change recently. But there was a process in glass manufacturing that is, requires a lot of expertise. The people that do this particular process, this is the, the feeding of glass from the, from the glass feeder into the glass bottle forming machine.
The glass comes out in a, in a, in a chunk of glass called a gob of glass. And it's, it requires a lot of expertise to get this process dialed in. It requires a lot of manual operations and a lot of checking. And anyway, we took this process of machine teaching and we applied it to that process. So we had a lot of data from, from our machines, a lot of data from our feeders. We put it in, we built a simulation of this process and then we let it go and figure out, you know, what's the best way to get the results that we're looking for, what's the best way to achieve the optimal KPI. So we give it a KPI that we're after or a set of KPIs that we're after and let it figure out all of the different ways. It's like the best analogy that I gave when I was talking about this with the team was you've got about 30 different knobs you can turn to get this dialed in properly. A human can't manage 30 different knobs all at once trying to do that. So you let the machine figure out, okay, what's the best option for all of these different settings to get the optimal setting. So anyway, we let it do it. We let it learn how to do it. And it was quite successful the first time. In fact, it actually made some suggestions that the operator probably wouldn't have thought to make.
And we actually had to, we had to, we had to intervene a little bit because the operator was like, no, no, you can never do that. That'll never work. And we're like, let's, let's go, let's let it try, see what happens. And then we did and it was actually successful. Wow. I didn't realize you could, I didn't realize you could do that. So it's good in that regard.
[00:23:45] Speaker B: So in that example though, Josh, like, what made that successful? Was it the KPIs that you put into place? Like how, like what made the AI successful in that actual deployment.
[00:23:58] Speaker C: It was the knowledge of the expert that we brought in alongside of it. Right, so we're teach. The AI is learning, but it's learning all of the strategies that the, our expert. I mean, we had a guy that had been doing this for over 20 years, 30 years, maybe he's been the expert in this particular part of the process. He knew this process better than probably most people in the world. So we had him alongside, helping guide the algorithm, the AI algorithm to make the right decisions. And you put that in place and you, you give it that set of guidelines and you give it a practice environment to go learn in and then it'll go and study and learn and it'll provide. So then in the end, what, what, how we initially deliver this is as a decision engine. So it's helping guide that operator to say, hey, this is probably the ideal settings right now. And they say, yeah, that's good, I like that, I like that what you come up with. Or they say, no, I don't think that's right.
That's also, you know, a good approach to deploy this, rather than just saying, all right, well, you're done making decisions about this. We're just going to put this in and let it control it. That's, that's scary to a lot of people.
I don't know any manufacturers that are quite ready to just let it take over an important part of the process like that. So, yeah, you have to start as like maybe a decision guidance and let it say, okay, well, these are, you know, based on what's happening right now, here's the best set of settings, or best group of settings, if you will, that will give you the optimal outcome. And, you know, put these in or, you know, give it a checkbox. Yeah, okay, go ahead and set those settings in. And you're, you still have the human in the loop, still have the human making decision, especially for critical process decisions that need to happen.
[00:25:44] Speaker A: Yeah, yeah, that absolutely makes sense. I mean, yeah, we all know whether you're using it for glass forming or if you're writing a letter to your boss or whatever, if you're using a large language model or a more, you know, a more complicated AI system, it's really only as good as the information that's going into it, the information it's being trained on.
So. Yeah, absolutely. And I think that is an interesting. Because there's, as you said, there's a lot of fear about AI coming in and taking our jobs, but AI, at least in the Environment right now still needs humans to guide and train it, especially in environments like that.
[00:26:19] Speaker C: Yeah, and again, I mean, that's an area that should be focused because there's a lot of knowledge. And what, and what things like LLMs are good at is actually bringing a lot of knowledge from a lot of different places and, you know, summarizing it in one place. I mean, that's what essentially the ChatGPT and all those are doing. They're just taking a whole lot of information that they're trained on from a lot of different sources. They'll even go out and fetch information.
I've built, I've built a couple of different systems that can actually just do that specific for an industry. We did it internally for a company to collect all of their, you know, SOPs and stuff like that to build. And then I'm currently working with a consortium to build one specific for the glass industry as well.
So it's just taking all the information and all the knowledge that is out there about, about the glass industry. And this one for, this one in particular about the glass industry, you could do it for anything. And it can, you can ask it questions and it becomes the expert.
It becomes the expert or the subject matter expert in that. And that's, that's one of the things that alarms are good for. Maybe not quite ready for the plant floor except, you know, for controlling things. But you can, you can bring a lot of information into a good place. And as I was talking about, you know, we've got people leaving. How do we capture the knowledge that they have and then put it in a place where we can summarize and get that back out to, for, for younger engineers or, you know, even for people that have been around that just want an easy answer to the question.
[00:27:58] Speaker A: You said you've created AI model, created some of these models for a glass environment. I imagine it's one of, if not the harsher environment, it's a very harsh environment.
Why do clean data models and this sort of textbook AI approaches often fail in an environment like that where you're dealing with, I think, 1500 degrees Celsius furnace?
[00:28:21] Speaker C: Yeah, well, I mean, I see this a lot talking to younger engineers, younger people or people in academia that because factories aren't clean environments, the stream of data isn't as clean. There's missing data, there's lots of information that is just not there.
And you think from an outsider, from somebody studying the industry from the outside, you might think, oh, wow, they must keep track of that variable or they don't have that variable. I was just recently on a, on a call where somebody was asking about a particular variable in a process and we were talking to the manufacturer like, well, we just don't measure that. Like it's just not there.
So there's that. And I think, you know, again, I think a lot of the real knowledge about a process is tied up in the heads of the experts, the heads of the operators. Operators. The heads of the operators. Operation people.
I say I've either heard this or have talked about this before, but I think the most important asset in any factory is not any of the machines. But the most important asset in any factory is those experts that have been there for 20 plus years. That's where the real knowledge of the process is. And those people are leaving. And if we don't find a way to take what they know and apply it then and we're going to lose, we're going to lose our ability to manufacture anything. Yeah.
[00:30:08] Speaker A: Let's talk a little bit about fractional leadership. You are a fractional technical director.
It's one of my favorite phrases from some of the information I've got on you. You describe yourself as a fractional bodyguard.
Can you explain that to us a little bit? And how does that help manufacturers avoid costly digital transformation mistakes?
[00:30:28] Speaker C: Sure. So fractional really just means part time. I mean if you want to sum it up. So I serve as a technical leader or a technical director or technology officer, whatever on a part time basis. And that helps.
The reason that that's helpful is because, you know, you know, mid sized manufacturers or software startups don't necessarily have the budget to hire full time executives in the technical, in their technical world. Mid sized manufacturers maybe don't have it or they don't even have a need to do it full time. They just want.
Think of it as an architect building a house. Right. You might hire an architect to build the roadmap or to build the blueprints, to build the foundation and then maybe the architect is sort of coming to check on things that, that it's, that it's getting built up at a time. You don't need them there all the time building that.
So from a fractional leadership perspective, it's, it's just doing the key pieces that they, that the manufacturers need or that a software company that's the other piece is. I, you know, I mentioned the software companies. It's the, providing that technical leadership that a software company going into the manufacturing world might need on a part time basis.
It saves money, but it also the other advantage is that because of the part time nature of it, I am doing a whole broad spectrum of things in technology. So it keeps me sharp, it keeps me in touch with what's happening. I'm not just confined to the four walls.
Whatever company I'm working for, I'm seeing what's happening. So in a given week, for example, I might be sitting in, sitting in a group of academic researchers trying to understand what are they working on, what's in the pipeline of what's coming next.
And then the next day I'm probably sitting with a software startup leaders talking about the strategy for how to make their software fit the plant floor. Right. And then maybe another day I might be walking through a 100 year old factory. In fact, as just last week I was walked through one of the oldest glass glass factories in North America. So all of that in one week gives me a wealth of knowledge that I can spread across all the clients that I have.
[00:33:01] Speaker B: Yeah, that's interesting. And then another day that week you might be standing in front of a classroom, right?
[00:33:07] Speaker C: Yeah, of course. Teaching the next generation of.
[00:33:10] Speaker B: Let's talk about the next generation. What are you teaching your students today that goes beyond traditional controls or PLC programming?
[00:33:21] Speaker C: I like to take it from the practical side. So I've myself have worked in controls for, for a very long time.
So I teach, I teach what it's called mechatronics, but it's ultimately industrial automation type stuff.
And the approach that I take is not academic at all. As a matter of fact, the coding of PLCs is probably very close to being done also the way that it's being done for all software today. Right. We're very close to being able to have the LLM build PLC code. So the practical approach is to how to understand how to troubleshoot, how to solve the problems in factories, how to all of the considerations that go into the design of a system, not just the programming piece of it. Right. The programming is a small piece of the overall picture anyway.
But then you know, that's even becoming as much smaller piece. So yeah, it's the practical aspects, what to think about how to think like a plant engineer, how to think like a controls engineer and troubleshooting. And that's a big part of the labs that the students work on actually is a lot of troubleshooting.
I could even give them the answer, give them the code and then let them figure out how to troubleshoot it. Like the basic stuff that controls engineers deal with, like just getting connected to a PLC can take A long time, because network connections and all the versions and the firmware versions, all that, these are all things that the students struggle with, too. The same things that we do. And that's, but that's, that's, that's the reality of it. And that's the things that should be learning and that's the things that are going to make them valuable to, to the manufacturers that they go work for.
[00:35:09] Speaker B: Yeah, just exposing them to the everyday issues that you've experienced. You have that experience and you can share that. So that's really valuable.
So, so let's sort of end this conversation with the big picture. Right. We, you know, we've kind of talked about, you know, from the, the students to the leadership and everything in between. Let's talk about the global competitiveness. You've worked with international manufacturers.
Where's North America falling behind and what should companies be investing in right now?
[00:35:40] Speaker C: I think North America's falling behind in the fact that we have always treated automation and digitalization as a luxury rather than a necessity.
And that's put us a little bit behind. I think when you look at, I was just reading a statistic, that number of robots, and obviously there's a lot more to manufacturing the robots, and it's a good, it's a good litmus test for how far levels of automation. Right.
Korea has the highest number of robots per 10,000 people. It's almost 1100 factory robots per 10,000 people.
Germany has a high level.
Some other Asian countries as well, like China, I think. I don't remember the full list.
They have had this silver tsunami that we are in the, in the face of right now, much sooner. So they've had to automate and they've had to digitize and they've had to get all of this information extracted from their experts sooner than we have. And we're, we're now sort of catching up to that in North America. So it's always been a luxury and we're now facing the point where it's almost a necessity and we're going to have to adapt because we're not as many young people are coming into manufacturing and more people are losing or losing more people in manufacturing. So, yeah, that's the difference, I think fundamentally is that it's never been viewed as something that we have to do. It's just something that's nice to do. And now those, you know, the other parts of the world are catching up or not catching up. We're trying to catch up with the rest of the world, but the Rest of the world has done it because they had to. And I think internationally, I'm seeing a lot of companies from Europe, Germany, companies from Asia come to the US to say, hey, we've done this before.
We know how to solve some of these problems. So they're coming and doing that too. I'm a, you know, one of the roles that I fill as a technical director for a company in Korea, from South Korea called UDM Tech, they have already perfected the control, logic, troubleshooting, root cause analysis. So they have a software that, that can translate PLCs from all different manufacturers into a common platform and provide insights and logic, the logic analysis of a PLC and then provide that out back out. So think about this.
You've got your, your, your plant goes down for some reason at 2 o' clock in the morning. This is always the scenario people are talking about, right? You maybe have a maintenance guy or an electrician there. He doesn't understand the control system. He doesn't understand how to do code troubleshooting. So your controls engineer is going to be getting a phone call at 2:00 in the morning. By the way, I've been that guy that's gotten that phone call at 2 o' clock in the morning before. Right. But if you can translate all of that and have a system and a system that can understand what's happening and then provide the insights and tell the guy, hey, this is exactly where your problem is at. This is how to solve it.
You know, you save the trouble of that. You save having to hire a controls engineer. If you don't even have a controls engineer, you save from having to hire one and come in to do that. So the rates at 2 o' clock in the morning are much higher than the rates at 8 o' clock in the morning. Right?
[00:39:21] Speaker B: Exactly.
[00:39:23] Speaker A: Absolutely.
Really interesting stuff. Josh, thanks so much for, for coming on and sharing your time with us and your expertise. Yeah, it's, I didn't, I don't think I expected we'd talk so much about the skills gap. We kind of jumped into it right at the beginning and it's, it's. Yeah, obviously you're coming at it from a different perspective because you're not only working in the industry, but you're teaching some of the young people who are maybe coming into the industry. So. Yeah, very, very interesting insights.
[00:39:47] Speaker C: Yeah. Yeah, I think that's the biggest thing. I mean, it's probably clear from the discussion, but I think that's the biggest problem to solve right now. And I think that's the biggest place that the AI can help us.
[00:39:57] Speaker A: Yeah.
[00:39:58] Speaker C: Yeah.
[00:39:58] Speaker A: Well, thank you so much for being with us. Great conversation, and hopefully we'll run into you again sometime. Yep.
[00:40:04] Speaker C: Thanks a lot.
[00:40:05] Speaker A: All right, thanks, Josh.
[00:40:07] Speaker B: Yeah, I really appreciate Josh doing his part to help fix these problems. I mean, really interesting discussion. And I know you always write down something that you heard that piqued your interest or that you thought was interesting, but I think the whole concept of, like, AI as the apprentice or the assistant, I heard somebody earlier today refer to it as AI as the sous chef. Right. So that's a really good way to
[00:40:33] Speaker A: look at it, actually, is like, you still need the main person, but it can help you. Yeah, yeah.
[00:40:41] Speaker B: So, yeah, lots to digest and think about. But, I mean, I.
I really think that there's this whole concept of, you know, the fractional cto, a fractional professional, or fractional bodyguard. Right, right.
That's. That's really interesting, too. And it can help some smaller manufacturers that just don't have the resources as the. The larger companies to put towards the. The developments that they need. So it's, you know, solving a real problem.
[00:41:14] Speaker A: Right. And to your point earlier about the. Your. Your sous chef analogy, somebody a long time ago told me, you know, this fear that AI is coming to take our jobs, and in some sense it is, and it will, and that'll. But I remember them saying to me, look at it this way. Right now, AI is not really equipped to do anything. It needs to be trained. It needs to be. So if you prepare yourself and become an expert in how to use AI, you suddenly are more essential than anybody else out there. You know, your job is now secure for the time being. So it's the people who basically bury their head in the sand and say, I don't want to have anything to do with this new technology, that you may have a problem. But if you really embrace it and learn how to use it, it needs, as Josh said, that person who's been doing whatever the task is. Glassworm, for 20 or 30 years to train it, and then it can be useful.
[00:42:10] Speaker B: Yeah. And, you know, we could go on with analogies all day, because then I just thought about, all right, we're training AI like we're training a dog. We have to be the alpha. We're the alpha AI. Right. And we've got to train. We've got to train the artificial stuff to do what we want it to do.
[00:42:25] Speaker A: Yeah. Which made me think of it. Noam, we're going to stop the analogy train there. But yeah, Great discussion with Josh. Always a fun time.
I also like that he brought up the plant engineering salary survey. He's been looking into our stuff, which is a great reminder that if you want more terrific information on, honestly, most aspects of engineering, from mechatronics like he brought up to basic automation, control engineering to plant engineering to design, we've got all of those sites under our umbrella, WTWH Media. Please check out our suite of engineering websites. Great stuff out there.
And as always, thank you for joining us. We appreciate having your ears and eyes on this podcast.
Yeah. Always happy to have you here and to talk to intelligent people.
[00:43:10] Speaker B: Yeah. And please reach out to us if you have some ideas for the podcast. We want to hear from you.
We are all ears and we're going to go like, we're going to explore all the topics that are important. So let us know what's important to
[00:43:24] Speaker A: you and check out our personal LinkedIn pages if you want to reach Stephanie or I. And there's a control alt manufacturing LinkedIn page. And I've got to say it legally, contractually, give us a, like, give us a follow.
[00:43:35] Speaker C: It helps.
[00:43:37] Speaker B: Absolutely. And thanks again for joining us. And shall I close with my signature? Closing.
[00:43:43] Speaker A: It's why I'm being quiet.
[00:43:45] Speaker B: Have a wicked good day.
[00:43:47] Speaker A: Thanks, everybody.
