ChatGPT+Manufacturing Series Episode 2. Key Insights for Gen-X and Baby Boomer Leaders to Consider
In this Industry 4.0 Club fireside chat, co‑founders Michael Ungar and Martin Cloke interview former MXI Technologies CEO Brad Forsyth to explore what ChatGPT and similar large‑language‑model tools mean for manufacturing leaders, especially those from Gen‑X and Baby‑Boomer cohorts. The conversation covers the surprisingly low barrier to start using AI, the hype surrounding it, and concrete ways it can add value: personalized operator training and on‑demand SOPs, natural‑language interfaces to ERP and analytics, rapid data‑lake querying for root‑cause analysis, and automated meeting‑note summarisation. They discuss how AI can act as an “expert assistant” that listens, interprets context, and iteratively improves training material. The panel stresses cultural and change‑management risks – the danger of creating new waste, over‑reliance on the tool, and loss of human problem‑solving – and recommends a cautious, incremental adoption approach (start small, fail fast, measure outcomes). Practical starter use‑cases such as using ChatGPT for paperwork, meeting transcription, and task assignment are highlighted. The hosts conclude with a call to experiment responsibly, leverage AI to unlock human potential rather than replace it, and continue the series with future guests.
Chapters
Good afternoon, and welcome to the Industry 4.0 Club's fireside chat. Our purpose is to further the deployment of Industry 4.0. Um, it's really quite simple. We do this by holding meaningful conversations on LinkedIn Live a couple times a week. And we also now— the Industry 4.0 Club is present at many, uh, upcoming shows to, to, to share our thoughts on Industry 4.0 and to help others share their perspectives. We're going to be present, um, in a couple of weeks on April 26th in Greenville, South Carolina. at Manufacturing X.O Southeast. We'll be at East Tech in Springfield, Massachusetts, May 16th through the 18th, and we're going to be at IME East in New York City from June 16th— I'm sorry, June 13th to June 18th. My name is Mike Ungar, and I'm one of the Industry 4.0 Club co-founders. We're glad you decided to join us today for our fireside chat. with fellow Industry 4.0 Club co-founder Martin Cloke and our guest today, Brad Forsyth. Brad's a former CEO of MXI Technologies, which is an aviation maintenance technology company. So if you're a baby boomer or a Gen Xer like me, in this second episode, we're going to talk about what we need to know about ChatGPT. And it's interesting to think about that for those of us that have been Working a long time in industry. As we get started on this conversation, if you've got comments you'd like to share or questions for any of us during the broadcast, please put them in the comment section on LinkedIn Live, and we'll be glad to get to it during the show.
Yeah, yeah, I know. So Mike, yeah, I'm super excited about this show and happy to hear, you know, how well received the last one was. You know, one of the reasons why I thought to get Brad to jump on and kind of ties into the topic of our show, and not saying that we're all old here, and I would say that I'm the young guy, but like at some point when I first got interested in ChatGPT and was sort of looking to those around me in my network who are also interested, the first one that popped up was Brad. And then Brad and I went back and forth and just started chatting about, you know, what it was, and, you know, where, you know, the opportunity that we see with this technology. And what struck me in discussing with Brad was 2 things. First off, the fact that, you know, you do not necessarily need to be, you know, 2, 3 years out of school to wrap your head around the technology. It's a brand new technology that, you know, is easily understandable on the surface. At the core, it's quite complex. And the other one is to be able to connect, you know, new capabilities with actual practical applications is something that you get with life experience in business and as an engineer. And in some ways, that combination of things just really made it exciting to me as like a Gen Xer. And then as we started chatting, it was sort of strange that Brad and I sort of connected on this. And I thought, given Brad's experience working in, you know, a parallel field to manufacturing or aviation maintenance, and the fact that you have gotten so drawn into this, I thought this would be a perfect, you know, addition to this series of conversations. And I guess maybe first, like, what piqued your interest in— because Brad, of all the people that I know, you have jumped in pretty wholeheartedly into this kind of technology. So what— why did you jump in?
Well, before I started, you know, started my own company and ran a number of companies, my background is actually engineering, and I did a master's degree in design optimizations. And way back, way back in the day, Um, and at that time, I, I was exposed to machine learning at an early stage, and I— and it wasn't, you know, it wasn't even close to being ready to be used, but it piqued my interest. So I've always sort of been following it, but I just— I've always been a technology enthusiast, and I've really— what excites me about is not so much about the technology, what you can do with it. And, you know, I've been following things like blockchain and crypto and all that, and I said, okay, this stuff, there's some interesting things there, but it's a long ways off, you know, to get an application that's really you know, this idea of a ledger that you can be immutable and all that other great stuff, it sounds good, but it's like, okay, where's an application that makes sense, particularly manufacturing or maintenance or, you know, quality management or any of these issues that I— that are important? Um, and then I came across this and went, okay, this is— the barriers to entry to get new technology in place is usually things like capital costs. It's like, um, the whole infrastructure to support it. It's the, you know, getting people up to speed on it. This stuff is all— it's, it's— we're there. Like, you know, right now you can jump onto Bing on your Edge browser and start using ChatGPT. I mean, there's— the barriers are all— it's amazing to me how the barriers are just— are not there.
The barrier to use it. The barrier to use it.
Sorry?
The barrier to use it, but not the barrier, I think, to get value from it. I think that's kind of the thing where In the past, the primary constraint was, you know, I don't know how to program Ruby on Rails. And I couldn't even like wrap my head around figuring out how to do that.
Yeah.
But now there's no barrier. But then the barrier is to not get distracted and not waste your time. And I think, you know, even the word enthusiast, it like, even using that word around this stuff, it's like, for me, it's like, if you see it as an enthusiast, you're missing the opportunity. Because at some point, like, I feel like I'm wired as an opportunist versus an enthusiast. And it's just like that first thing, the barrier to entry. And I think that's one of the topics here that we're going to get into later in our conversation. But the fact is, the barrier to entry to begin to wrap your head around this is low. And this is for folks who are, you know, executives in manufacturing who are, you know, further along in their career. And I know, Mike, you've even dabbled. I don't know, would you categorize what you've done as dabble, or is it more than a dabble?
No, I've just dabbled. It's, you know, one of the things that, you know, dabbling with it, I think we'll get more into this thing as we talk about it. But, you know, about sort of the opportunities and then the risks with it as we— but I've just dabbled. I haven't really gotten into the— to say, well, a real practical application yet, to be honest with you.
Yeah. And I guess like, Maybe I'll give you a hard time to not dabble, like I just gave you a hard time for hearing your printer in the background, but I'll park that for this conversation here. Okay, so you got drawn into it, and I think maybe other thing, and for those of you who are following along and want to post questions, so Brad has gone deep into what's happening here, and this is not necessarily the conversation to go in deep, but if you—
Yeah.
You gave me like an hour tutorial describing all of like the sort of the science that went into it. It was just It was just fantastic. And I think, you know, it's important to understand a little bit of that as an engineer, but you don't necessarily need to go that deep to understand this. Okay, so then maybe just touch on how deep have you gone, and then just to set the frame of the conversation. So we're, you know, I think I would say, Mike, you know, we're speaking to somebody who is quite familiar with something that's evolving like day to day here.
Well, that's right. And, you know, Brad, for me, You know, I'm playing with it almost like it's a toy. So I guess I'm really curious to get your thoughts. You know, is it a toy? Is it just hype? Or, you know, what do manufacturers maybe need to take from this and think about applying it? Maybe you can talk about that a little bit.
Yeah, exactly. I mean, there's absolute hype. I mean, there's no doubt, unfortunately, something like this comes along and the hype meter, you know, is way over. But having said that, there's a— you step away from the hype and you start looking at what's there. We have something immediately that's very extremely valuable, right? And you know what, what they've been able to do was, you know, if you delve into it, what I was showing Martin was the fact that the actual math behind this thing is not that complicated. What is just mind-blowing is the scale that they did it with, right?
Yeah.
What they did is they took the internet and they started training very simple math but on a huge scale. And the thing that's so neat about what they did was they basically really started to extract out of all this material meaning from what words have meaning. And the example I gave Martin, which I thought was kind of funny, is this is that if I say a phrase, you know, on, you know, during Bob's run in the morning, he was thinking about how he could run the bank so that it wouldn't be a bank run, you know, the, you know, Silicon Valley bank run. So that the word run is used there 3 times, but it means very different things. One run means the action, you know, um, and you could have other similar words like jog, sprint that all have similar meanings but not quite the same. Then you have run the bank, which means operate, manage, different meanings. And then you have the final one, bank run, which means, you know, everybody takes their deposits out. So what this learning thing did is it started to create— it started to attach to a given word all of its different meanings and realize that in context, the meaning could be quite different. And that's a very powerful thing because it allows them for it to do all kinds of things in terms of generate text. But it's doing more than that. It's actually extracting meaning out of what you're asking it and giving you back sort of like super search type capabilities of saying, oh, here's what you— I know what you meant. And so now I can give you back meaningful results. And that's where I went, wow, think about all the things that we do now that, you know, in an operations, like, you know, You know, set up a manufacturing operation where you have standard operating procedures, you have QA manuals, you have operating manuals, troubleshooting guides. You know, all these things are all documents. And we, you know, we try to interact with them either physically, paper, or online or whatever, but they're still— they're all these documents. Instead of doing that, what happened if you had a really good experienced operator standing in your ear? they could just ask questions like, how do you, you know, how do you operate this? Oh, what's our standard operating procedure here again? Oh, what's, you know, just think about the ability to quickly ask a question and get— because the meaning's there, it gives you back meaningful, immediate results.
It's just—
Sorry to interrupt you. As you describe that, it takes me back to, you know, when I started my career as an industrial engineer. And so I'm writing a work method, right? And so I'm— I've got a— I've got a college degree and I'm writing it for folks that don't have a high school degree. And so I've got to go to— I remember going to— oh, I even forget the term. But where it would give me a— would grade my— what grade was my writing at? In other words, for 7th grade, 8th grade, 9th grade, right?
Right.
And so I would— I would get back an 11th grade. Okay, too, too, too high. Lower it down. And I would go through and I would change the words. And what you're describing now is a tool that will understand its audience by the way they ask the question and absolutely provide feedback in a way that they can understand. Even I can imagine having moved down here to the south of the US in the early '80s. Even with dialect, even with the choice of words that might be unique to this area. It's pretty fascinating.
It is. I saw a really cool example of, of all things. Someone was trying to understand some tax code, right? And they said, explain this to me as a senior accountant, and then explain this to me as a layperson. And the answers they gave back were both excellent but totally different. Like, one is, you know, very detailed from a tax accountant standpoint. The other one's like, no, Joe Blow, this is how you, you know, this is what it means for you, right? And that's so it can take context, it can take the level, and that's what's— because it's getting the meaning and it knows that there's ways of using different words back to you that have the similar meaning, but, you know, can be simpler or more complex depending on the nuance of the situation.
But there's a key thing that Mike said that I want to delve into, because in some ways, when, you know, the first example that you gave us was to be able to interpret a sentence. And then, you know, I'm sure people have played with it to write a poem.
Yeah.
Or, you know, like write a '90s rap song, you know, wishing my kid happy birthday, right? And it's like, okay, cool, that's a gimmick. And in some ways, like, so what Mike said was, you know, he was writing for, you know, somebody with a high school education, college education, trying to adapt that. So that's like one step. So you write something But you said an important thing here, which was to be able to detect the reading level. So one of the things that's sort of like happening now is this idea that you have an optimization cycle. So if you have a training manual that is where somebody's interacting with that training manual, you can gauge whether or not there's a certain level of understanding. You can gauge whether or not it is effective. And as part of sort of this technology, not only can you create it the first time, you can optimize it. And that's one of the sort of the things that are sort of, you know, what's happening in the last few weeks is this idea around using AI agents to effectively create an objective function. So the objective function is to train an operator so that they can perform at a certain level. So, well, how do you need to train them? You need to provide them with the training material. But that's not only it. You need to make sure that they are able to take the information in And that the evidence of training should be, you know, visible through their performance. So this idea that you actually have an objective function, and not only can ChatGPT or these kinds of technologies create it the first time, they can also interpret. And I think one thing that's underplayed right now, because we're so excited about ChatGPT's ability to speak, and in some ways, like, the last thing we need is just more people speaking. Like, we spend all of our time Like as a society, we're like awesome at speaking. We're pretty crappy at listening and even crappier at remembering, right? So one of the key things that this is able to do is you can listen, digest, and that means you could listen to create a procedure. You can listen to see how people are interacting with it. You can listen to see what, you know, I think that ability to listen, interpret, and improve is really exciting. Like the talking part is just sort of the, the part that's gaining attention because it's the easiest to sort of start to play with.
Now, you bring a good point up. I think, you know, I was thinking about this iterative perspective that you were talking about with respect to training manual. And you can imagine the first time through, and then the individual tries to do it, someone's observing, and okay, we don't quite get it the first time. And so we tell the, we tell the tool, hey, this didn't happen. Can you, can you change how you describe it so that we can make sure that person doesn't make that mistake? And I can imagine the enhancement and how fast that can happen. Boom, boom, boom. And it can be personalized to the individual. It's just really cool, I guess, when I think about it.
Well, Brad, you just did ER— I think you were helping a company with ER implementation. An important component of that is training, right? Yeah, I'm guessing.
I'm guessing. Well, absolutely. Well, there's all kinds of aspects to putting in place an ERP or a main, you know, a manufacturing execution system, right? And this was a particularly interesting one because it was a— it's talk about future manufacturing. This company creates aluminum powder that you can use for 3D printing parts, aluminum metal parts instead of plastic. Just really amazing technology. And so very unique internal process, but we could put in place an ERP system. And what I'm thinking, well, you know, there's all kinds of things around training and use, but I really think, um, a huge potential for this going forward is, as a, you know, is a whole new interface to the damn ERP. Because ERPs are so bloody complex, right? They're just— who loves their ERP? I don't know. You know, I used to build one, um, maintenance for airlines, um, but I, you know, and they liked me, but I don't think they love the software, right? They liked it, but it was not, you know, it's just ERPs are complicated because they do complicated things. It's particularly when you want to pull out the analytics, right? But just think if you could sit there and ask it questions instead of having to figure out how to do some new dashboard or figure out how to generate a report. You just tell, you know, say, you know, what is the, you know, what is the production, you know, per whatever, right? You know, what is our waste? You know, what is our— and, you know, what can we do? What's the contribution to our waste elements? Like all these things you could start to ask and talk to it and Instead of having to go through a complex UI, right? And that's just—
Well, and Brad, I had to ask you to kind of comment further. So when you say ask the question, if I understand correctly, it can actually give you the— you know, we talk about one of the principles of Lean is visual management. It can actually give you the graph with a picture, right?
Well, that's where it's going very, very quickly, right? I mean, you know, it's go— it's— there's Interfaces that are— they're starting to build. For example, Wolfram|Alpha, you know, that the math, all those, the big math, where they basically have a new plugin for the GPT that allows you to actually— it'll— because I think we talked about a little bit earlier, you know, the one problem with— we'll get into it in the future. One problem is that it's not always that accurate with math and stuff. But if it knows to go ask the right question, it then interfaces to something like you know, Wolfram system, it'll then go to the system and get the real math, or go to the ERP and get the real data. And that's where it gets really, really interesting, because then it accurately could do things like, you know, invoke the UI and, you know, give you the graph, give you the tables you want. You know, it's all possible now, right?
It's interesting, as we're talking here, a friend of mine named Steve Feltovich on LinkedIn, I think He, old school Toyota manufacturing. I don't know if you know him, Mike. Great guy. But he would basically throw us under the bus and guys like, you guys are just talking about the wrong thing. Like, you know, we're gonna go and ask a whole bunch of people a whole bunch of questions. He'd like waste, waste, waste.
Yeah.
And as we're talking, like, do we really need another tool to ask a question we can answer in another way? And that's why I think, you know, if I were to go to the classic, OpEx consultant who is thinking about the core principles around driving improvement, which is what all these organizations are trying to do. It's like just having, creating systems where we need to ask more questions and we ask more questions of people. It's basically, you know, you're probably taking people away from the thing that they're trying to do that provides value to the customer. So, you know, in the example of the ERP implementation, you know, like one opportunity could be to ask new questions. The other opportunity could be to see if there is evidence of waste where somebody doesn't quite get it. And maybe the question can be, You know, are you having trouble with this, or is there any way that this could be better? So like, you know, if thinking about wiring a system so that it is making it easier for organizations to identify waste and elevate that waste, but while at the same time making sure that you do not, you know, position yourself where you become a big source of waste. Because one of the things with, you know, this technology in particular, and I'm sure you've all seen that graph, showing how long it took to get to a million users and is probably $100 million now. This stuff is intoxicating and it's intoxicating hundreds of times more than— I never played any of those games on my cell phone, but what, with the Candy Crusher or whatever? So this game is like that times 100. So we have to be very cautious with this kind of stuff because it can create a massive amount of waste if we don't implement it in the right way, which is why, kind of circling back to our topic, You know, why having that filter of a practical executive that's been burned and seen technology fads come and go, I think that filter is sort of, you know, is the one that's going to take this technology that's kind of going out like this and actually point it back at specific things that can provide value. And that's the biggest challenge today.
I think following this conversation around making something visual and data, You know, we talked today about data lakes and getting— putting together the data lake for the key things in your process. And one can imagine that, going back to your point, Brad, about being able to point it to certain elements to pull the right data, you know, we— what we always struggled with was getting to the root cause by asking why 5 times, right?
Yeah.
Trying to get to the root cause of the problem so we solve the problem. Well, if I've got the right data lake and I've got something like ChatGPT that can help me harvest that data very quickly, well, then I can engage the workforce on problem solving. Now, how do we fix this relatively quickly? To me, that's really exciting. And do you see the same thing or what are your thoughts on that?
Yeah, I mean, that's, that's what— there's so many, there's so many elements like that that are so exciting. And I think what Martin just said, though, just reminded me of something that I think is really a little bit of a tangent, but really important, which is, you know, when we first started putting in place ERPs and other systems, um, accounting systems, other systems, the first thing we did was we wrote software that looked like the paper, right? Because our mental model was, here's the paper, this is how we're going to write the software.
And that's—
and it was really bad, right? It was just extremely bad. And then we realized, wait a minute, it's a new paradigm, there's a better way of doing it. We don't have to have We can actually think about it quite differently. And we started to get much better systems. And I think that we're in that little bit of that paradigm right now is that we are used to, you know, systems that are based on, you know, very hard, very prescribed methods of interfacing to equipment, right? Like, you know, here's the control panel, here's the— we're still locked into some of that stuff, right? And even our data analysis, we're still a little bit locked in. But I think you're right, Mike, is that what's going to happen is, you know, it's going to evolve quite differently, right? It's going to evolve into these— and we don't want to make it sound like a person, you know, anthropomorphize it and say, well, it's going to be a person, but it's like an expert you can actually start to say to work with, you know. And I think the ability to have something like a system that you can actually work with and not direct, right? We're so used to directing the software. We give it the commands and we do this. We don't actually work with it, right? Which to me is really interesting is we're going to have a potential to start working with, right? To do the 5 Whys in the sense of it's not just responding to you. It actually can do more. It can actually extract meaning and give you new meaning about, hey, have you thought about this? And allow you to work with the system. And that is going to be, I think, an enormous shift in how we think about this thing. And it's not going to be— like I said, we've got to be careful. It's not some magic thing. It's just— but it's a different way of working with a computer.
Right.
And it's very powerful because it's going to be a transition from a controls controlled method to an interactive method.
I think the example of, you know, early development of ERP software is perfect. So you took the paper system and you recreated it with software, and I think that kind of highlights exactly what, you know, what it feels like right now. And but the new paradigm, you know, with regards to how people interact with technology, so the one we're used to now is we're used to interacting with these systems with a live interface that's primarily data, charts, graphs, analytics. In some ways, we've all been trained this idea of data being the new oil, analytics. We've been trained to think like that. That's how you actually are supposed to communicate to people. And I think the only reason that we are now consuming these analytics tools is because we haven't been able to automate the creation of text content to actually tell us what the meaning and what the insight is. And now— and there's background noise, guys. So the text content itself, now that with software we can create both text content and data content, you know, I would imagine that the nature of, you know, the way that performance insights are communicated is going to change, is going to have to change. And like, to use this kind of technology to enable like a, you know, better analytics and reports, I feel is the same thing that you did early on with your ERP software, which is to basically— and I think that's not the thing. We are, as humans, you know, we're not designed to interact with zeros and ones. We can understand it. We all went through math and all that kind of stuff. And for us on this call, we probably did more math than was needed. But the thing is, as humans, we interact with, you know, at best with concise communication between each other. And then if somebody— I think that's the key thing here where I think that's a paradigm shift that's happening that I don't think we'll have to struggle to try and interpret fancy charts and graphs to figure out what the meaning is. I think that there's a better way.
I somewhat, I sort of agree, I sort of disagree, 'cause a chart can be a thou— a picture can be a thousand words, right? So there's that always still, but imagine a system where you said, oh man, how do we increase our, How can we increase our production by 20%? And if the system came back and said, well, you could do this, this, and this, and here's a chart to prove I'm not, you know, of how it would work, I still think it's going to be there, right? You know, but it's—
Yeah, yeah.
It's a different— it's not like, oh, I'll generate the chart and then try to interpret it. It's like, no, I asked the question, it'd give me an answer that maybe is a chart. And I think that's what Mike was referring to, right? If you start to ask it like, you start to interact with it and say, you know, what would happen if we were to, you know, introduce this new machine into our production line, right? What do you think would happen? And it would, you know, it could maybe give you a new model of the production line throughput, right? I mean, those things are going to still be there, right? But just a different way of doing it. So it's not like I'll start generating all these charts and then I'll, you know—
I think, you know, and I'd like to respond to it and throw it to Mike here and then just, just so we keep on track for the conversation here. But I think what you're highlighting So in that, in the transition from ERP on paper to ERP with software, there was an overlap period where you also created the paper because people don't want to see the thing that makes them feel comfortable disappear.
Yeah.
Right? And then at some point you realize, like, you know, if you're using GPS in your car, you don't need charts and graphs. Just tell me where to go and tell me because it's like, there's like a, there's roadkill or there's a cop around the corner. That's what you, you need to slow down because there's a cop around the corner. I don't want a graph saying like, you know, the number, here's a number of tickets that got issued in the last hour. So I think the idea that there's overlap as we transition from one paradigm to the next, that overlap is important. And I think for us, analytics is not going to disappear, but the thing that's likely more effective to communicate to our teams en masse is not quite charts and graphs. But yeah, Mike, I'll let you jump in here.
Well, no, I think it's a great conversation. Let me flip the— if I can flip the script just a little bit, because we're getting really excited about the technology. What can go wrong here? So what are the mistakes that could be made, do you think? I mean, because it's, you know, like a lot of things we used to talk about, we don't want to— we didn't want to automate waste, you know, because that's not good. What could go wrong here?
Brad?
Yeah, yeah, no, I think, you know, with regard— I'll sort of maybe highlight in the manufacturing context. I think, you know, it's creating waste. It's creating, you know, at some point, If a manufacturing company wants to drive improvement, you know, they need 3 things before we bring in technology. Like, does the team recognize that improvement's necessary? Do they believe that improvement's possible? Are they willing to take actions to find and solve problems? So if these tools get in the way of those 3 things, full stop. So I think we need— if these tools can augment and improve your ability to make those 3 things, to create that culture of continuous improvement, not only help the culture, because the culture comes first, actually help them find problems and solve problems and all that good stuff, perfect. But I think the thing that, you know, organizations in general and in manufacturing in particular need to focus on is like, let's not introduce chaos into the system. You know, there's ways to manage change. There's ways to introduce new things collaboratively. You focus on people first, then process, then tools. And I think that's something that is true for this technology. It's true for all technology, but this one, given how powerful it is, it's almost— we need to pay attention. And the fact that this is not something that is going to be taking data from a PLC or a new kind of vision system or a robotic arm that's going to replace a process. This has the potential to interface. This is designed to interface with people. So from a cultural perspective, the potential negative impact of misdesigning these interfaces is huge 'cause you're like, the culture of an organization is at the core here. So I think that's the big challenge here, which is basically, I'm sure for those who spent their careers in manufacturing talking about change management and managing culture, like that's pretty much your job as a leader. But this is, be hyper-aware of those kinds of things 'cause You know, that's the biggest risk, just messing with team, messing with culture.
Absolutely. And like, and I think, yeah, and it's, I think it's easy to get to not be focused. I think you really want to think about what kind of results you're trying to achieve and really keep it up. Don't lose sight of that, right? I think it'd be kind of easy to lose sight of what results you're trying to, you know, the improvements you're trying to make for the organization and make sure you're getting them. And make sure you do it incrementally, right? And this— and there's, you know, one of the criticisms of this system right now is that it's currently— the current implementation of it is really around this meaning-type concept. But it still doesn't quite have what we consider common sense, which is basically, you know, we have built into our minds and our abilities the ability to model the world around us, right? We have a fundamental concept of gravity, right? And that stuff's going to fall, right? And stuff's going to break if you do certain things, which this model, the model that's currently implemented doesn't really quite have some of that. And that's some of the criticism. And so it will make mistakes. And so you want to make sure you're using it for the right kind of things, and that you have a review process. Because, you know, I, in another analogy, it's like, you know, I don't know about you, but when I get to talk to my doctor, they start telling me stuff, I still question it, I don't take it verbatim, right. And so this is an expert, this is a system is giving you its best guess, but it could be very wrong. And you still have to have checks and balances. And you have to do a little bit of a risk assessment of like, if this is wrong, how bad would it be, right? Those kinds of things still come about. So, you know, it's— that's why I would strongly recommend that you think about, okay, you know, what would be the— what are the outcomes we're trying to get out of this system? You know, whether it be training, whether it be whatever, you know, productivity increases, um, whatever. But make sure that you, you don't lose sight of what the goal is and then implement it in a way that you can measure those things and see if it's having an impact and adjust, right? And that's the thing that I'm excited about is, after, you know, Martin and I've been playing with this thing and you can do so many things so quickly, right? It's just, it's not like setting up an ERP can take 6 months to a year, um, and then it, you know, and then to get it up and running. You can do some of this stuff in days. It's almost scary how quickly you can do this, but you can do it incrementally and you should be doing it incrementally.
So I think almost practically speaking, like, so if, you know, a plant manager is listening to us right now, they go, you know, you need to do things in a careful way and think of the objective. And I think they're just thinking like, I don't have time for this. And I'm wondering, like, in some ways to accelerate adoption, given the the fact that it's not that complicated technologically. And to your point earlier, it doesn't have this common sense thread. And I think one of the things is that for, say, OpEx consultants in manufacturing, like the core principles that enable manufacturers to improve are extremely, like they make a ton of sense. You know, they touch on culture, they touch on process, they touch on a whole bunch of stuff That, you know, it's been very similar for decades. And it's almost that skillset, the common sense skillset of just an OpEx consultant, either big firm or small set, that's combined with the ability to, you know, fill the data lake full of all the content. And now I think, you know, as you say, you take the skill of an OpEx consultant, you assume that the leap to actually get them to be able to wrap their heads around this is not a huge leap. And then the core thing that drives the value of these systems is that data lake, which historically has been full of machine data. And now you use this technology to fill it full of a whole bunch of context, like data from operators, a bunch of whys. Like you get as many why is this happening from as many people as possible. You listen to it perfectly, and then you have this giant source that can be actually harnessed by this thing. So I think that role in particular Could be one way for manufacturers to sort of wrap their heads around next steps. But then now the OpEx consultants need to figure out how to wrap their heads around, you know, getting the value. But I would almost want to— that, I think that's the right challenge, you know, have a skeptical OpEx consultant, you know, finally, like, have that light bulb go off in their head. And I think that's almost one of the things as an industry here that would help accelerate things because, you know, I could just see somebody going like, oh, I hear what you're saying, Brad, I just don't know how to do it.
I, you know, I wanted to come back to something that Brad had talked about too, Martin, and it's— I'm reminded about, you know, when we— when I first started, uh, in manufacturing, one of the things we were trying to do was engage the workforce. You know, uh, we had this, this, you know, for the longest time it was, um, you know, check your brain at the door, so to speak, and just, just follow the task, right? And then we move to— then we move to saying, you know, we want you to help us learn what's going on and get involved in problem solving. I think there's a risk here too that we want to pay attention to, and it's that we rely on this tool and stop thinking. And I think, Brad, you kind of alluded to that. We have to think. We have to still encourage people to think. You know, it's going to be easy for an individual who's Who's, you know, we ran 24/7 and run on night shift and 12-hour shifts on the night shift, the last 2 hours of the nights. I mean, look at people's eyes, it's like close to zombieism. Okay. And so how do you keep, you know, it'd be easy to rely just on the tool and not think. And so I think we have to really be careful and make sure people are still thinking while using the tool to leverage it. The other thing that comes to mind as we talk about anything that's new, you know, with any Industry 4.0 technology, the adage of start small, fail fast, and then scale. You know, the idea that, you know, let's get in, start using it, trying for things, make some mistakes on a small scale where it doesn't cost you much, but then once you learn where you can leverage it, you got to go fast. I think that's another opportunity we have here to follow that principle and really apply the tool, apply ChatGPT to this kind of AI to help us improve our results.
Exactly. You know, I was— and the other thing I was thinking, you know, a good place to start, I know what I would do is, is if I was a manager, like, and Martin said, oh, they're too busy. Well, I can make you less busy. Just start using ChatGPT for your paperwork. It's, you know, start using yourself for every time you do paperwork, just, you know, do a quick try of ChatGPT to see if it can help you. Because I'll just give you an example. Some of my— I have a couple CEO friends now are doing this. They, you know, because they do all online meetings now, it's pretty easy to hit the transcribe button on the meeting. And basically, they have the full text of their meeting, right? And how many people are really good at taking minutes from meetings?
It's a pain in the ass, right?
So what happens is they take the transcript, they say, hey, ChatGPT, summarize this meeting for me. Boom, it gets down into one page summarized. And then he says, oh, and assign all the tasks. Right? Because Martin put his hand up to do something in the meeting, and it basically starts assigning tasks out of it. And he said, my friend said it just saves them so much time. And the meetings, people that miss the meetings now have a summary for it. They don't have to read through the transcripts. You know, it just— and the productivity gains there, it's like that. I guarantee you, if you start using it yourself, not in your operation, just to do your job as a manager, you're going to free up time to start thinking about—
Yeah.
And it's going to show you, oh, there's all kinds of light bulbs going to go on of, oh, I could use it for this, I could use it for that. So that's how I would start personally.
And I think maybe— and I know we need to wrap it up here, but I guess, you know, like, um, Mike, you're gonna post on behalf of the Industry 4 Club a very well-written blog and a social media post. And just, I don't know, are you— how much credit are you going to take for last week's post?
0.1?
0.
I mean, I get Yeah, I hit the button to say, uh, you know, download the video and/or the audio and hit the second button to get the transcript. And then, you know, third button for the blog. Yeah, there we go. So it's that, it's that, but then look at it really quick to say, does it, is it accurate? And then off. And again, that thinking process has to still be there, I think.
Yeah. And I think that like, this is, you're basically describing what, uh, uh, a seasoned OpEx consultant would, would, you know, hammer our heads saying the 8th waste. You know, like if you create a system that generates more of the 8th waste, the waste of human potential and ingenuity, that's probably the worst waste to waste. And I think the idea here is that that's— it needs to be framed around this idea of unlocking, you know, that stuff. So I— we're at the end here, so I'll sort of let you, you know, kick it back to Mike to help us wind this down because we could go all day.
Yeah, so it's been really great. And maybe we'll start with some final thoughts then as we get ready because we've talked about a lot of great topics today. Martin, what are your final thoughts on this whole conversation? Then we'll go to Brad. Did you not hear me? Brad, can you hear me or not?
I can hear you.
Yeah.
All right.
So Brad, I'll go. Yeah, I said, you know, my thought is that if you're a manager or anyone in this area, You know, I would, you know, don't be afraid to start trying it, using it. And I would use it in your daily work, and it's gonna, it's gonna reveal to you, um, its strengths and its weaknesses. Um, the only thing I would make sure you, that you, when you're using it, that you don't try to use it like Google search where you just put a very short, um, you know, what is this, um, because it has no context then, right? Remember I told you before, it's looking for meaning. So you want to make sure that you Give it some context. Um, you know, write a longer, a little bit longer question, or dictate a longer question, um, in terms of give it some context. You know, say, I, you know, I'm a manufacturer of, you know, of transmissions for, you know, these types of cars and blah blah blah. And then, and then put in a question or say, you know, what, what's, you know, what's the best operating procedures you know of for this, or whatever. Give it some context, but try it out. and then see where you can use it. And I think it's going to be— you're going to be able to better understand its strengths and weaknesses, and it's going to open your eyes to potential. And I think having done that myself now for a few months, it's just like, it's hard to keep the hype thing down. I'm trying to do that, but I'm super excited. It's like when I went— I'm old enough that I can remember when the internet came about. Right? We first, you know, and all of a sudden, instead of having to run to the library and, you know, to get manuals for my engineering crap, I could just get it online. It's like, wow, how did I ever do this? How did I ever do market research before? How did I ever do this without the internet? Right? Well, this is like that on steroids. It's just like, how did I ever— if I could have taken this magically, put it in my pocket and had this, man, I could have been such a better CEO.
That's fantastic. Martin, do you have a final thought on our conversation today?
Oh, we lost your sound, Martin.
You're on mute. You're on mute.
Yeah, sorry. I missed a little bit there because my headset kind of frizzed out here. But yeah, I think tremendous capability. And I think we said that again and again. And I think Brad mentioned you know, trying to, you know, sort of quiet down the hype. And, but the problem is like every time I dig in deeper, it's just like, it's the hype sort of just compounds. So it's not really working personally, but I think focusing on, and I know focusing on the objective and make sure you achieve the objective like that, that's kind of like stuff that everybody says about, have been saying about technology for a long time. And I think the practically speaking, beginning to test both, I would say, this technology's ability to reflect and respond and also to listen and digest. And I think taking notes for meetings is probably one thing that I'd recommend as kind of a first start, right? Because often what happens during meetings is that there's a huge amount of insights and I sort of, you know, that 8th waste, wasted human ingenuity. At some points, if you listen to everything that's said in a meeting, there's a bunch of little nuggets there that if you were to collect that as an organization and pile it together, there's a whole bunch of knowledge that could benefit the rest of the organization. And I think so practically speaking, play with the tool. I think I'd recommend for organizations just to check to see if there's ability to play with it through the API rather than going direct, just because of data security. And that's something that is— we spoke about in our last podcast. or last show. But yeah, so, you know, it's something that can't be ignored. I think, you know, as mentioned last time, you know, with experimentation that's done and a report coming out from Harvard Business School, how people's productivity is going up between 30% to 80% for office tasks. So, you know, given how much time we waste doing those kinds of things, we can't afford to— no organization today can afford not to pay attention, but you need to pay attention in the right way and be cautious. So, you know, you need to experiment, but do it cautiously and be wary of the fact that it is distracting and can point your organization in the wrong direction. So yeah, continue to be excited. And I'm excited about continuing on this series. I think we have some great, great guests coming up. I know that Jeff Winter and Joachim Hensch and Ira is gonna come on and Catherine Brierley. So we have a bunch of neat shows. Coming up in the near term, really, really jazzed to sort of continue on with the series. I think we're even, you know, up it to a couple shows per week. So for those who want to follow along, you know, stay tuned and keep an eye out for our posts. And feel free to make suggestions about other topics and people we should be speaking to.
Yeah.
This technology is advancing very quickly, and I'd love to hear from folks who are playing with it right now and trying to, you know, achieve practical things in a manufacturing setting, either, either on the, on the, the manufacturer's side or technology companies as well.
Yeah, thank you, Martin. That's a great point. So hopefully for those baby boomers and Gen Xers who've been listening today, and really for anybody from any generation, hopefully you've gotten some really, uh, uh, some nuggets from today's conversation that, that will help you to move forward and really experiment and learn from ChatGPT. I want to thank everybody that's joined the broadcast today, that's been listening. In particular, I know Joanne and Mutano, if I pronounced that correctly, thank you for joining in and your comments on LinkedIn. If you like the sort of conversations that we have, as Martin said, we're going to continue this broadcast, the ChatGPT series. And our club, the Industry 4.0 Club, also has other conversations around women in manufacturing, the future of manufacturing, and people, process, and culture. Please join our LinkedIn page as well or visit our website at www.industry4.0club.com. With that, I want to thank everybody again for joining us and wish everybody a wonderful rest of the day. Thanks, Brad. Thanks, Barton, for your time.
Thanks, guys.
