Hi, I'm Kence Anderson. In this video, we're going to talk about the skills gap. I want you to reflect for a minute on your own organization. Who in your organization, if they left, would leave a huge hole in your production capabilities, and as they walk out the door, what high-value skills would leave with them? I was talking to Max Petri an R&D and innovation executive at PepsiCo snack foods a couple of months ago. He told me something that surprised me, but also makes a lot of sense. He said, our most valuable asset is the skills of our operators. We're here talking about the skills gap. I want to ask Max about this because he's been on the actual snack food manufacturing lines and also now manages a group that works on process improvements and innovation. I want to hear Max's take on skills. Yeah, the most important asset for us at PepsiCo is the skills of our people. In producing our products, specifically the skills are our operators who are running those lines. Wow, that's an amazing statement and I've heard similar things, but you say it so clearly. But really you're suggesting that even more than the equipment that it's made on, this is like millions of dollars worth of equipment and process lines is the operator's skills in running those lines? Yeah. I look at the same way as data. You'll hear a lot of people say that the most important asset for a lot of companies is their data or their capabilities they have with their equipment or something. But you need to have someone who knows what to do with it. Yeah. If you don't have people who are actively utilizing that to turn that into a finished product, to turn it into a finished service, then it's just going to sit there. Just think about that for a second. Let that sink in. This is a $60 billion a year revenue generating company. He said, more important than the equipment that makes the products that they sell, more important than the business executives, more important than the facilities that make the products, are the skills of the operators. That means that these skills in some ways are priceless. When I walk into many factories, I asked, how long does it take to learn how to do this? They say, well, you could be reasonably good at it in a couple of months, maybe a year. But often, it's years or even decades before operators have enough practice across enough different scenarios that they can execute their skill at an expert level. Tell me a little bit about what it takes to be a good machinist. How long does it take to train? A guy coming in off the street, a young guy just out of high school with no training, at a minimum, on a very simple operation, it's probably going to be six months training. Okay. A guy coming out of a trade school, we could do him in maybe three months, half the time. Right. Because he's familiar with some of the operations and the machinery. A guy completely green, it's ground up. You're having to teach everything. In that training, there's so much safety involved in that training because you want to walk away like I did with all the digits intact. That makes sense. It's a big deal. So it's safety first, and then you get some of the basic operations, but then it's going to take years to get to the nuance. Yes, to really get good, you are five plus. Five plus years. Yeah. I mean, it's when I can walk up to a guy and hand him a print, and say I need this, and he turned around and produces that, if he can do it in five, he's a really special guy. The difference between a novice and an expert operator is what the skills gap is all about. In the skills gap, we're talking about how to identify problems in situations where autonomous AI will be useful. We already talked about identifying the value of the problem. Now we're taking a close look at the importance of skills. Skills are units of competence at performing specific tasks. We'll be talking about skills in a lot more detail in course number 2. Unlike high-tech and in some cases, finance, the manufacturing industry runs on the exchange of skills, not on algorithms. That's why autonomous AI is such a good fit for this industry. The difference between skills and algorithms is like the difference between classical music and jazz music. I'm a saxophone player and I love to play and listen to jazz music. When you're playing classical music, you read the notes on the page and play them in sequence exactly as written. Jazz improvisation is very different. With jazz, there's a structure and there are rules that guide the selection of notes to play, but musicians rely on intuition that they built from extensive practice to determine what notes to play. Algorithms are like written classical music. They codify exactly what to do in each situation. They provide step-by-step instructions for how to do things. Skills are more like improvised jazz music. They require practice that builds nuance intuition about what to play in a situation. Algorithms are more rigid, while skills improvise through uncertainty. Every time I interview an operator or an engineer that controls the process, they described how they practice and practice to build the skill. But what about all the people who haven't practiced that much? Can they do the job too? Yes, they can with the help of rules and recipes. We all do this when learning new tasks. We follow rules that help us make decisions well, until we build our own intuition as we gain expertise. That's how a skill is built. In manufacturing, operators, supervisors, and engineers practice building, refining, and teaching high-value skills for controlling the equipment and processes. I think you have a really interesting perspective on using AI to upskill workers. Can you tell me about that? Yeah. Just like most manufacturing in the United States you see, and we were in Navasota, Texas, which is between College Station and Houston, which is away from Houston. When you look at manufacturing, you have to find people to come and work in these remote areas. Yes. It's hard. People want to live in cities and those areas. Our goal, I think this was exact goal I spoke to you and Kevin about was I said, how do we take AI so that I can go to the Navasota ISD and go take a person who doesn't go to college or the military and say, "Come work for NOV and we can get you to be effective in the workforce very quickly. You don't want to have to spend 6-8 months learning how to use a piece of equipment." How do I use AI to be those training wheels to get you up to speed very quickly. Tesla is doing it in the automobile industries. But how do we do this in machining, or forging, or something else? We say, "Look, don't worry about it. We'll put these training wheels so you won't hurt the product or the machine.". Right. We can learn enough to be able to do that. That was the intent and we felt technology was there to be able to help us get there. I love your perspective on that and that was music to my ears when we talked about this years ago. Because a lot of people are afraid that autonomous AI is going to take jobs. But really most of the executives that I'm talking about, think like you, that it's really about my changing workforce. Many experts retiring, I have folks coming in. How can I use autonomous AI to capture and codify the best of my current experts and pass those skills on to the next generation? Skills are different from assets, they can be accumulated and store to accelerate improvement in manufacturing. Assets like part blueprints, manufacturing recipes, they're very specific and they're very proprietary. Skills on the other hand, can be common to similar processes stored and customized to the process of specific companies, manufacturing lines, and product types. This suggests to me that any mechanism for an improvement in manufacturing must be able to adjust, adapt, and manipulate skills. Similar to how financial systems can be controlled with algorithms. How is it possible to accumulate and store skills that exist only in the minds of experts using software? That's why we call an autonomous AI design a brain. It's an agent that can acquire, perform, and manipulate specific skills. Machine teaching provides the organizing structure for those skills. These skills can be stored perhaps along with specialized virtual environments for practicing skills in certain scenarios. Imagine training the best experts in your organization using machine teaching, which enables them to codify their most valuable skills and then design an AI that will practice and master these skills. This is especially important to the many organizations that are at risk of losing large amounts of experts as their experts retire. In this specialization, we'll treat skills as the right level of abstraction for advanced problem-solving instead of algorithms or rules, or recipes. Whether it's our supply chain team where they worked on these lines for decades in building that inherent expertise and knowing how to run every piece of equipment on those lines. Or its our R&D folks that have been running experiments and developing these processes for years, it's very difficult for us to harness all that capability across all of our fleet in multiple countries, multiple plants at one time. The application of AI and machine learning gives us the ability to have essentially your expert operator or your R&D technical expert right there on the line with you. Yeah, that's what I find really exciting about this course where we are specifically dealing with autonomous AI. AI that can learn to perceive and then respond and act like a learning control system. Then it opens up possibilities like you're saying, where it can be another teammate on the line. It can explore areas and help you see what might happen as it interacts in a simulation. But it can also learn from you. You can say no, I heard you have experienced here, these are the specific skills that you need. Absolutely. As we're building the architecture for these brains, that's something that's really important to us as you have the ability to just let the brain go. But we think we actually have an inherent advantage and being able to build in some of that knowledge that we have from these experts on our team that had secure execution so great. Let's spend a little bit of time talking about the human aspect or the workforce development aspect of it. One of the things that I hear all the time is a fear that autonomous AI is going to take the jobs that these people who have built their livelihoods on operating these plants and equipment, what do you say to that? No, I definitely don't think that's going to be the case. For us, we've optimized our lines over the years and we've got the right number of people run into lines today. Really the way I look at it is the application of autonomous AI is on these lines, is going to really enhance their capability, and be like that second person that's helping them running these lines day in, day out where it'll enable them to focus on the things that are most important, and also build their own knowledge and capability as they're on the lines. You still have a very important need for human touch. A lot of the things we do are very manual on these lines, but this can take some of those things off their plate and really enable them to be successful every day, every shift. I hear two really important things. One is it's a better-together story, which I do totally believe in. I believe that autonomous AI and expert operators together are going to be better. We see this in chess where teams of chess and AI players will beat humans, and then there'll be AI players. But I also hear you saying there's a physical and sensory aspect to it where you have the right number of people on the line, like there's only so much that you can remove people from the equation? Yes. We've had that in my opinion years ago. We're already there. We're already in a situation where you're in a line, you're having to check and see how things are running on a certain part of the process. You're loading up, seasoning or equipment on another piece, you're fixing something somewhere, you're grabbing measurements for quality. A lot of that stuff still needs to happen. But the application of new instrumentation or AI can really help automate some of that to make your job a little bit easier and should make you better at your job too. Yes.