Artificial Intelligence Podcast: ChatGPT, Claude, Midjourney and all other AI Tools

Can AI Solve the McFlurry Problem with Kurt Lueck

• Jonathan Green : Artificial Intelligence Expert and Author of ChatGPT Profits • Episode 339

Share episode

Welcome to the Artificial Intelligence Podcast with Jonathan Green! In this episode, we dive into the intriguing world of AI in manufacturing with our special guest, Kurt Lueck, co-founder of Automation Intellect, a company dedicated to revolutionizing machine performance monitoring.

Kurt discusses the challenges faced by manufacturers and how AI can play a pivotal role in diagnosing and solving issues with machinery. Drawing parallels between the infamous broken McFlurry machines and complex industrial equipment, Kurt explains the importance of understanding fault mechanisms and how AI can streamline this process.

Notable Quotes:

"There's no machine that someone can't find a way to break it that I've experienced." - [Kurt Lueck] "Having a machine tell you it took half a second, now it's taking six tenths of a second... can be significant." - [Jonathan Green] "We're building a coach IQ that allows you to have conversations with your machines." - [Kurt Lueck] "The best use case of AI is data analysis." - [Jonathan Green] Kurt highlights the potential of AI to simplify complex data sets in manufacturing, making the process of identifying and addressing inefficiencies much more accessible. He emphasizes the value of providing manufacturing personnel with AI tools that allow them to ask questions and interact without fear of judgment, fostering an environment of continuous improvement.

Connect with Kurt Lueck:

Website: Automation Intellect Kurt shares insights into Automation Intellect's innovative Coach IQ module, which leverages data analytics and conversational AI to enhance machine performance and ease manufacturing processes.

Connect with Jonathan Green

Can AI solve the McFlurry program? Let's find out. Today's episode is brought to you by the bestseller Chat, GPT Profits. This book is the Missing Instruction Manual to get you up and running with chat g bt in a matter of minutes as a special gift. You can get it absolutely free@artificialintelligencepod.com slash gift, or at the link right below this episode. Make sure to grab your copy before it goes back up to full price. Are you tired of dealing with your boss? Do you feel underpaid and underappreciated? If you wanna make it online, fire your boss and start living your retirement dreams now. Then you can come to the right place. Welcome to the Artificial Intelligence Podcast. You will learn how to use artificial intelligence to open new revenue streams and make money while you sleep. Presented live from a tropical island in the South Pacific by bestselling author Jonathan Green. Now here's your host. Now this has been something on my mind for a long time, Kurt. They, McDonald's has always had the broken McFlurry machine. And there was recently a massive lawsuit that just was settled very recently because what McDonald's has always done is that when it's broken, it puts an unreadable code on it and no one who owns a McDonald's other franchise franchisees are allowed to know what it is. Another company came in. Decoded all the error messages. That's it. And they just put a thing on top of it, a little screen that reads the error message and translates it to English. It just tells you which part is broken. And the McDonald's corporate went crazy, right? Because they charged people these massive fees to repair the machines. And it's this really interesting, 'cause people always complain. The McFlurry machine's always broken. And now you know why. And I think about. How many different things I have that the error codes are inscrutable. So there's this challenge for the manufacturer to make an error code that is simple enough that someone can understand it, but complex enough that they can then solve the problem themselves. And that's something that I'm very fascinated about by what you do because you work with much, much bigger machines. I have my printer always says. Just pull the paper out and push the button again, it will solve itself, right? Like it doesn't gimme any more complicated things, but I'm so fascinated because there's no machine that someone can't find a way to break it that I've experienced for sure, and especially in the field that we're in. These are very complex machines and there's. Also, there's just so many people with their hands in the cake. Yeah, it's a huge problem now. Now, I will say, just jumping in a little bit to our real world the biggest thing is around faults. Like a machine is running, it's running. But most modern machines have, have a series of, and it could be th a thousand different faults. And so those faults are spitting out and saying, Hey I'm having, I might still be able to produce the part, but I'm having a little problem here. And the problem can grow and they could be, and they can even be what we call micro faults. So a micro fault is a situation where the fault occurs and then it solves itself, but it takes up just a little bit of time. Then over time it grows and it, what went from five seconds of downtime grows to 30 seconds. But if it happens many times over the day, that starts to build up to be actual real downtime and higher volumes. But the operators, they have learned how to operate these machines. They come in every day. They come in at, five in the morning and they know these machines like their baby, and so they know when that fault occurs, I just press this thing or I push this thing or I jiggle this and I do that. And and it starts to work again. And so part of a big part of what automation intellect does, and then we're starting to get into to. To doing more of this with AI as well, but that is to identify these micro stops and put some language to, hey, that's taking a lot longer and it's growing over time. That was a long answer to your McDonald's question. I think it's a really good answer. So one of the, and this is a big issue I deal with as well with software with people, is that you have a small thing that delays you. But it delays you over and over again. And I'm one of those people that it's really hard to switch gears if I think I'm doing an app, if I think, for example, I'm doing a podcast appearance or having a guest on, and then it cancels at the last minute. It's very hard for me to be productive with that time doing something else because I've. Gotten in the zone, my preparation zone and I'm thinking, planning on doing one thing. I I'm very linear as in I can do one project and then the next, but I can't do two things at the same time. Yeah. And that's just how I operate. So I like to do something to til it's finished. Yeah. Which is why if something ends up with a problem that I get really stuck.'cause I keep trying to fix it even though it's not a good use of time. So I can see how these little inefficiencies build up. And we see them all over the place. One of the places I see them in software a lot is with email. A lot of people, we get so much spam email. Yeah. And irrelevant email. And it's, you don't realize it at first because it just takes you an extra minute. But now we've all been dealing with it for 30 years. Yeah. And we don't realize it. Come look at my Gmail. Come look at my Gmail. It's a mess. And it's, there's an entire industry of specialists to design systems and like strategies, not even like software systems, just ideas for how to control your inbox. And then there's software implementation ai. These little problems, I can see how they can become significant. And the important thing is if we're trying to detect when a problem is getting bigger, we often don't notice it. So we're not very good at noticing small changes. It's like that story about boiling a frog, I think it's not even true, but that idea that if you just do it really slowly Yeah, slowly make it worse, people don't notice. Whereas if you just cranked it up really fast, they go, it's really hot on you. You jump out. Yeah. Yeah. Yeah. So I can see how having a machine tell you it took half a second, now it's taking six tenths of a second. You wouldn't notice it, but if you wait till the time you notice it, the problem is at three seconds, five seconds, 10 seconds, and yeah. Yeah. So it's very interesting. What do you find is the biggest problem? So when you look at machines and people use them, is it that op, like people use the machine and they just don't wanna say anything because they wanna hit their productivity marks so they just wait too long to fix it? Or that they just think of it as oh, this machine has character. Like I just know I have to kick it on the side every three days and it works fine. Think it's a little of both. And I think it's just that I think your discussion on the fact that we just deal with small problems and we just deal with 'em. We don't actually take care of them. But then I also think you have to look at the fact that if you ever gone into these manufacturing facilities, they are multiple football field size, most of them, most of the ones that we deal with. And so you walk in there. If you're a continuous improvement person, you're like, you're really probably focusing on the ones the lines that are the most critical and like you said, the one that has the biggest problem that just screamed, I'm a problem and I have to be solved. So I, our software's absolutely gonna tell you that, but I think we also, for those people who can put a structured approach in to say, let's also, let's take the time to also fix the ones that are growing, the ones that aren't screaming yet. So that, yeah that, but it's I think it's just a. And it's not just our clients. It's every client in manufacturing has these problems. There's just so much going on. There's so many deadlines that have to be met and there's not enough people. This is the, gets into almost the political discussion about just we can't find enough good workers and manufacturing is really struggling with that problem. Again, that's where we try to help that. Yeah. I think that more and more we, everyone wants to do a software job and not do a physical job. And there's so many TV shows about manufacturing, how to make a motorcycle, how to make a cake, how to make, create different things. But we almost see them as not real anymore. But I had, I used to have a friend who worked in a factory, he used to put in the back seats of Saturns. Now you're dating yourself, Jonathan. I know I'm old. He loved it so much. He would talk about it all the time. He was like, he would demonstrate, he goes, I put it in the chair, like the backseat, like this, that I bolt this. And then I wait for the next one. And he had and it was very interesting to me.'cause I would struggle to do, I struggle with repetitive tasks to do the same thing over and over again. My mind will drift. But he had this. And I thought it was a really impressive skill, this just ability to put in the same backseat 50, a hundred times a day every day for months and months you had to, until they, they went outta business. He did it. He was so excited to be in the union. He is like my uncle's in the union got me in. So people in manufacturing league, it's so fascinating, but a lot of us on the outside don't really understand it and the demands of a physical product because. When you're in information or software space, you don't have as much like issues with stuff sitting on the shelf or delivery costs. Yeah. So all these things can add up because maybe you're five minutes late, but now you're five minutes late to give the delivery to the truck and now they hit their drive limit for the day and suddenly you're delivering it late and now you have a punishment for that. Yep. So there's a lot of things in manufacturing that people outside that world don't understand. Yeah. We live in a, what is it called? Like last minute of delivery or like this idea that we, or just in time. Just in time, right? We, most products and most warehouses and most stores, they get the product. They want it there right before they sell it. They don't want it sitting back. You don't want, you don't want it, you don't want raw materials sitting there.'cause it costs you so much. And that's what we found, like with that in the United States, we are so efficient. But that's what, that's also our weakness. That's what happened with Covid. It's something got broken in that just absolutely efficient process, and then it just took forever to get it to back into that smooth rhythm that it has. But anyway. One of the areas where artificial intelligence now can be super useful is with kind of diagnostics. Helping you to solve a problem. Like I've always been fascinated, there's this book called The Checklist Manifesto, which was a massive study on how when you make surgeons do a checklist at the beginning and end of a surgery, the cases of patient death plummet. Like patient survival rates at the worst hospitals in the world. I'm talking third world countries where they don't have supplies, they don't have anesthesia. The infection rates went down to almost zero. We have this, and the reason I think I was in manufacturing, so a hundred years ago when you'd manufacture a car, there was one person, it was like the maestro. And so one person would make a car from the beginning to the end. Yep. And when they switch to you make the back, you make this, you make that the. Each person doesn't ever part in the same way surgeons see themselves as above the checklist. And so when they convince'em, go, listen, it just could be five points. The nurse says it out loud and it will save half of your patients major difference in results. It was a massive campaign to try and create this shift. And it's part of, it's because pilots, when a pilot has an emergency, they pull out a checklist. Yeah. They can switch doctors to that mindset. So there's this idea that. Giving people the steps to work through is so helpful.'cause even if you think, you know it, sometimes you forget step six outta seven just because it's an emergency. You haven't had to do this one in a while. So where I found AI to be so helpful is it never forgets the checklist, right? So whenever I'm working through a process I'm trying to solve something or fix a small problem, it could be super helpful to work hand in hand with me. Can't solve the problem itself. It doesn't have hands. It's not a robot yet, but. And I'm very interested in some of the things you're doing with that area where you can start to give someone guidance to say, start by looking at this part. And they say, I tried this, it didn't work. And go, okay, try that. Yeah. So some of those things are very interesting. Can you talk a little bit about how AI can help someone diagnose cooperatively? Yep. So I'll give a little bit of. Background context of what we're doing and then try to answer your question. So remind me if I don't answer your question, come back to it. We're building so our software essentially we install a small edge device on these multimillion dollar pieces of very complex machinery. And it gathers data like good parts and. Status of the machine. Is it running, is it not running? Faults, it sends all this up to the cloud. And so that's great. So we install this software and you, you now go from almost no visibility to what's going on in at each of your pieces of machinery, your lines, and your, frankly, your whole plant. Okay? Once again, great. All we've done though is really give you visibility to it, and you will get a huge benefit from that. But it doesn't make you do anything. And so we see at times two different scenarios. One is a client that basically loves our software and they actually have some business analysts and they start really digging into the data and they'll continue to climb and get better benefits. And then the other is. They just get the software and they don't do anything. And that's, it's very much a struggle. So years ago we realized that we need a value team that's gonna come in and it's gonna help you with the software becoming your, an extension of your continuous improvement team. And that role, that's what we're trying to replace, not really replace, if they're listening, we're not really replacing you. But the value team's job is to basically look at this data and to figure out where are the issues, how can we begin to push you towards, Hey, you should do something. You should do this. But the benefit of having an actual value team member is that you can actually have a conversation with them. Guess what about for us? I know that LLMs and generative AI has been around for a while, but for us it's really come on in the past few years and light bulbs went on within our organization that I think we could, I think this has a really good use case. So that's what we've started trying to create with our, our the module is called Coach iq. And so we have a chat that basically allows you to have those conversations. So I don't know, did I answer, did I get there? Yeah, that's exactly what I gonna talk about. Yeah. Which is that the best use case of AI is data analysis. Yeah. So if I'm trying to. Do something with software. I don't know how to do any coding in Python. I get an error, I can copy and paste it in. And it'll say, this is the fixed copy and paste this code. And it al it eventually hits a point where it can't solve a problem and that's when I have to go to an expert. Yeah. But it definitely solves like those lower level problems and it allows you to deal with, there's a big issue I believe, which is too much data. So we have these dashboards with dozens of metrics. Yeah. And it's. There's too much information. It's kinda like I feel when I look in a cockpit. There's much data Yeah. That I wouldn't be able to interpret at all. That's why it takes years to become a pilot and there's so many levels of training. There's so many moving parts'cause it's a lot to learn. So learning how to interpret data, know what's important, and then react to the right data. It's a very. Specific skill or training that we don't really get in a lot of educational paths? I certainly didn't, yeah. My degree never covered that. I wish it had. Because you can look at, for example if I just look at all of my posts for the last month and try to figure out which one was the best I will always be wrong. So I was doing analysis recently and the po, I thought a post failed and it was my most popular post the last six months. I go, oh, no one liked this one. And then I looked at the numbers, so I was that wrong. I'm that far off. But I can send a picture of the screenshot to an AI and say, looking at this data, because there's all these little things. They say, oh, what time of day is the best time of day for you to post? It's I don't know. You like what? I'm not watching that. I don't think I post when I have an idea. Then. And you get these habits, you schedule at certain times. And I have all these habits. People will say I love to launch a product on a Tuesday, and I used to only email at 11:00 PM at night. And people always say, why do you email at the middle of the night? And I go, all my customers are under 30 and if they're awake at 11, they're watching infomercials, like they're ready to buy something. And I'm low ticket stuff. Yeah. Now I, so I think about that, but I don't know if I'm right. So the ability to have it, say, here's all the data, but here's what matters. Like to simplify it because we've just gotten so obsessed with data that you have enough information, you can't really do anything about it. Like how often do we have spreadsheets where half the columns don't fit on the screen anymore? And I have so many dashboards that you, it's really hard to say, this doesn't matter, this doesn't matter. I just need I don't need 20 of these, I just need two. So I think that's a very interesting use case, which is. You have a lot of data, you're not doing anything, and we are all guilty of this. Every single one of us has these massive bits of data that we just don't have the time to go through. We don't wanna spend time doing it. So having an AI fill in that gap to say, I. Okay. This is what you need to be looking at. These things don't matter. I think that's a really cool use case because it's where AI is the most accurate. A lot of people think AI is good for making pictures and writing blog posts, and that's where it's the worst. Interpreting data analysis, spreadsheet analysis, all the stuff that we're terrible at is humans is what it's the absolute best at. Yeah. So I really like, that's what I'm really fascinated about what you're doing, which is that it can. Help you to go from I don't know what to do because I have so many data points. I don't know what's the right thing to work on. Yep. And I think I think there's also some level of I, people feel like they should know what to do and they're actually, when they get our software, I think they're afraid to ask that question. Okay I know you've given me all this dashboard stuff. And I still don't really know what to do with it. Where should I start? And so if we could give somebody a private connection to a coach and they can ask dumb questions and good questions and start to be able to really engage with it, that's our goal. It's funny you mention that because a lot of people, the reason they don't use chat, GBT. There's this thought that a lot of people have, which is that it's keeping a record of who's dumb and who's smart. And there's this fear that and I hope it's not that. They're like, if you ask a dumb question, Nick, you get put on like Santa's naughty list, right? And I know this, right? I felt the same way I was first using a few years ago. I go, I don't wanna. I know there's no dumb questions, but this computer never forgets. And I don't want to. So it's very interesting that you, we often will, and this is a very human thing, right? Like we'd rather never ask for directions and admit we're lost. Like this is a standard thing. So I think it's really clever to try and find ways to say, listen. This is only, you'll see this, no one will ever know. It doesn't keep a record of your questions. So this is your place to ask. Yeah.'cause there's nothing worse than when someone explains something and you, they say, you understand? You go, yes. And then you realize you absolutely didn't. I have no idea what you just said. Yeah. When I was teaching English as a foreign language, which I did for 10 years around the world, working on my first postgraduate certificate, they said after you give someone instructions, make them say the instructions back to you.'cause people say yes all the time, even though they don't because they don't wanna look dumb rather than because they know. Yeah. And you. That's really powerful all the time. You say what did I say to you? What are you supposed to do? It's definitely affected my parenting. When I give my kids instructions, I say, I'm gonna do it, dad. Okay, what are you gonna do? And they're like I wasn't listening. Can you repeat that? Yeah, I was thinking about Fortnite skin. Sorry, dad. You know what's funny is on the same topic, but I have noticed that I'm kind. To my chat, to my coach iq. I say, please, can you tell me this? And somebody pointed that out. I'm like, I don't know why I do this. I just, I'm kind to it. And I actually have read some blogs about the whole topic, and I don't really know. I still don't know exactly why I do that. I think I'm a generally kind guy, but I know this is just a machine behind it, so I'm not sure why I'm being kind to it. So that's one of my favorite topics because there's two reasons people are kind. One is fear. That one did. AI will take over and remember the kind one. It's a real thing. And the other one people anthropomorphize or they just go to their default setting of how they treat it like a person. The way AI are programmed and their core programming is that the way they know they've given a correct answer is affirmation. So imagine that the way you get paid is by customers saying Good job. And you have two customers, one who comes in every day and is neutral, doesn't say anything. Yeah. Another customer who says when they're happy, they says, good job. Who are you gonna give more attention to? Yeah. That's what's happening there. So it's a very interesting way. It's that because it, the only way they can know if they've given the correct answer is if you say, that's a good answer or good job or, thank you. Yeah. And that's also, bad job. You go, that's wrong. So that's part of its learning algorithm. It's how it's, basically seeks affirmation and tries to move away from like disappointment and never wants to disappoint you, which is why it'll lie to you. If you don't have good training in there, and that's why the default model will lie. Interesting. So what you're doing, it's very interesting that a lot of people do it without realizing it's actually giving you a better result. So you're getting a positive reinforcement, so it's a good thing that you're doing and it doesn't. It doesn't mean there's something wrong with you. It means that you're getting a superior result and you gotta keep getting better results. Why would I stop doing it? It's, it's, I don't know about the computer side of it, but it's definitely a life thing. I try to tell my kids treat people well, because in the end, you want them to treat you well. And especially with adults, if you treat them terrible, they're not gonna give you the benefit that you could get and you won't know what you're losing. But yeah. Yeah. It's. It's one of the hardest lessons with my kids 'cause I have five. And One will complain, the older siblings bullying them and then they bully their younger sibling. Do you don't like when it's done to you, but you still pass it down? Do it. It's so hard to break that chain. And it's exactly that thing where you start to develop these habits. But I think it's really important to see that there are. Uses for AI that are outside of what people are thinking about that are so powerful and can make such a big difference, that the stuff in the news is usually the least useful. It's the most exciting. There's this inverse relationship between exciting and value, like the most profitable days of the year every. Every few months I have an optimization week, which is why I go through my entire business and look for anything that's not, that's broken or like sometimes people in my system, like a customer will end up off a end up off a cliff to nowhere where they suddenly stop getting emails from me or they did just hit a dead end because there's something wrong where there's a broken link somewhere. And no matter how many times they go through this process, and it's. It always makes you feel bad.'cause it's like just finding everything I've done wrong. And that's part, and you fix all these things and it's, that's very, that's a sign of a good, healthy business or person too, is like the ability to look at yourself and go, okay what isn't working? That's a scary thing for people to do, but it's a healthy thing. So we hear about the 80 20 rule all the time, but nobody I. Everyone who teaches it never does it. Like very few people sit down and go, okay, what did I do last week? That was a waste of time. I was doing something earlier today that was a waste of time, and I was talking to a friend. I was like, I need to stop doing this because it's like a two hour job that will generate $7. What am I doing Right? Something broken for years and it's like someone brought it up again. I was like, oh, I haven't fixed this for two years. And, but it's easy to get caught up in these loops, right? We want everything to be perfect, but it's really, it is very, you just have to look yourself in the mirror and say, what am I doing wrong? What needs to be improved? Looking at each little piece, and it's the most boring stuff to talk about, which is I look at all my funnels, I look at email deliverability, I look for all these little things that are half a percent here and half a percent there, and are people getting lost and is there something getting lost in the communication? And for a while I was getting a lot of followers from the wrong countries. I was like, where are these people coming from? Because right. Native language is in English, so they, it just and I just had to like figure it out because it was just becoming a bigger problem. First you have a couple of people and then you start getting all these constant emails and you have to figure out what's causing this.'cause now it's slowing down the machine, right? And it's all part of that optimization. So it's not the most glamorous, but it's also usually the most. Valuable. It's the consistency is far more valuable than like a moment of fame or virality. And so I really excited about what you do. So for people who interested or working in the manufacturing field, and we have a lot of followers in that space, a lot of people are in tech, like manual tech companies. Where can they find out more about what you're doing? See how AI can actually really help in the world of manufacturing now. Yeah, for sure. Just hit www automation intellect.com. That's our, just our base website. And there's a sign up for a demo. And we'd be glad to, to talk with anybody who wants to kick the tires and look at what we're doing. Core model is really just around, like we've talked about, just helping people to understand how the machines are performing. We call ourselves a machine performance platform. That's essentially what we do is we pull in all the data about what's going on at the manufacturing site and let you look at it from a coach iq, what we're really doing. Are two things. One we have we're using large language model and really some traditional just business intelligence as well to allow you to look at all your past 30 days of performance or look at your faults in detail. If you use our coach iq, it basically allows you to, because people don't know how to do it, that, when we first rolled it out, it's okay, I know that I know what chat GPT is 'cause everybody's heard of that. But what does that mean for automation, AEX product? What do I do? Like I do, I just ask it. So what we've tried to do is to. If you look at the UI is to to simplify it. So you can just say, I wanna, I want to know about this topic. And so that's like the last 30 days of production. Or I wanna look at all the changeovers that are recurring, or I wanna look at all the manual input. So we've basically given them selections of data that they can look at. So that's that I will put. Coach IQ into two different sections. One is that more data analytics, and then it will go get that data. It goes out to the database that it looks in the background. It's going, okay, what is it that they've asked for? Okay, this is about data. This is the subject they're looking for. Now we go to the database and that's nothing to do with large language model. That's basically our. Just running some SQL in the background. We go get that data, bring that back up, throw it into the model, and then we do a first pass interpretation of what that data is. And at that point then it goes into that second phase, which most people are really familiar with that conversational aspect. So we really try to break what we are doing into two different pieces. The data analytics and the conversational, I'm most excited about what people do with it. Like how do they have a conversation about it? We're we're really excited about where that goes. Does that make sense? Yeah, I love it. Okay. I think that there is this hurdle. I don't know how it works. I dunno what to ask it. And easing people into, that's really critical. So I think that's one of the things I'm really excited about. What you're doing is that, yeah, you're helping people who think, oh, that's not for me. That's for these types of people. I don't wanna learn a programming language. It's you can ask it really, once you get a feel for how to ask it. Simple question. It's like that. Once you get that first taste of it, then you start to get used to using it. It's that initial hump you have to help people over. So I love what you guys are doing. And I put the link, of course, in the show notes and below the video for you guys watching on YouTube. So thank you so much being here. Kurt, this was an amazing episode. Yeah. Of the Intelligence podcast. That's awesome. Thank you. Thanks for listening to today's episode starting with ai. It can be Scary. ChatGPT Profits is not only a bestseller, but also the Missing Instruction Manual to make Mastering Chat, GBTA Breeze bypass the hard stuff and get straight to success with chat g profits. As always, I would love for you to support the show by paying full price on Amazon, but you can get it absolutely free for a limited time@artificialintelligencepod.com slash gift. Thank you for listening to this week's episode of the Artificial Intelligence Podcast. Make sure to subscribe so you never miss another episode. We'll be back next Monday with more tips and tactics on how to leverage AI to escape that rat race. Head over to artificial intelligence pod.com now to see past episodes. Leave, review and check out all of our socials.