META This! Ep. 9 - The Future of Robotics: Insights from Yarek Niedbala of KUKA Robotics
Illustration by @max_gps
META This! Series Ep. 9
Podcast Transcript:
Stefan Guttensohn:
Hello, I'm Stefan Guttensohn and you're listening to the META This! Podcast. Welcome to another episode where we explore the advancements, trends, and strategies in various creative-led industries such as technology. Today we're diving into the fascinating world of robotics.
Yarek Niedbala is a distinguished expert in robotics and automation with a rich and diverse career spanning multiple countries. He holds a systems design engineering degree from the University of Waterloo, complemented by business studies degrees in France, and the United States. Yarek's journey in robotics began with his contributions to the development of a PC-based controller for robots and CNCs at Hewlett-Packard's meteorology division. His career took a significant turn when he joined KUKA, a renowned German robotics company, which I'm not mistaken, I believe is one of the biggest robotics companies in the world, and he rose to the ranks to become vice president of sales for KUKA Robotics Canada.
Today he is vice president of technology advancement, where he leads a team of engineers in developing AI-based robotic solutions and advanced processes for various applications. His function spans a number of industries, including what I would call general manufacturing, automotive, architecture and construction, entertainment, agriculture, aerospace, and even woodworking. That we're going to have to ask him about. In today's episode, Yarek will introduce us to robotics. We'll be discussing the latest trends and where we are. Welcome, Yarek. It's a pleasure to have you on the Look Legal Podcast.
Yarek Niedbala:
Thank you, Stefan. It's a pleasure to be here and a pleasure to talk about a subject that's very dear to me, robotics.
Stefan Guttensohn:
Well, let's begin with the most obvious question. What is a robot?
Yarek Niedbala:
Yeah, that's a good question, Stefan. Over the years, the definition has not only, I would say changed, but also there are many different definitions depending on where you're looking at robotics from. For myself, coming from an industrial background, a robot for me is a reprogrammable device, physical device that performs physical tasks or a series of tasks. I think that's a pretty short but sweet definition, and I think it applies pretty well to the kind of robots that we deal with day to day, whether it's articulated robot arms or AMRs, meaning autonomous mobile robots. That's a good definition, I think.
Stefan Guttensohn:
You mentioned reprogrammable, so if you have a device that is only programmed to do one thing and it cannot be changed, that in itself is not a robot?
Yarek Niedbala:
Right. So think of it as in automotive industry, it's a good example. Robots tend to be programmed to manufacture a car model, and that typically it's a span of seven years. That's the duration of a program. However, at the end of those seven years, those robots are still alive and kicking and ready for more, and it'll be a pity if you could not reprogram them to do the next model for the next seven years. You know what I mean?
Stefan Guttensohn:
Oh, I see what you're saying.
Yarek Niedbala:
Oftentimes, it's good to think of it as being a device that can actually do more than just one task. You know what I mean? For example, you take a robotic articulated arm in the automotive industry, that arm itself is basically just a bolt. By itself, it's not that useful. You do need to add a gripper to it, you need to program it. You need to add sensors around the cell, maybe conveyors that bring the parts in, conveyors that bring the parts out. Of course, there's safety. You have cage and so on and so forth.
Let's say at the end of a particular product that is building, you want to be able to repurpose that robot to basically do other things, meaning that you can actually take that robot arm tooling and replace it by something else. So instead of a spot gun now, you basically have a gripper, a vacuum gripper that now it's picking boxes rather than welding cars. You know what I mean? That's I guess what I mean by reprogrammable more than doing something autopilot that's basically kind of a rather narrow task for it to do. You know what I mean?
Stefan Guttensohn:
So a robot would be some sort of machine. It's a device. It's a machine that gets instructions, that has some form of sensor in order to carry out its instructions to do something around it, and it has to engage in some sort of activity, so it has to manipulate its surroundings, and without those elements, you don't really have a robot then.
Yarek Niedbala:
Yeah, I would say so. Yeah.
Stefan Guttensohn:
Where are we today with robotics? So my understanding is that the robotics industry is also advancing. Well, maybe not as rapidly as artificial intelligence, but is going through a new dawn, if you will. Is that accurate?
Yarek Niedbala:
Yeah. Robotics continues to evolve, quite frankly, continues to evolve, and new branches get formed. Even if you take a look at industrial robots, which is probably one of the largest fields of robotics, although right now from what I've seen recently is that service robotics, actually are really picking up. But even within industrial robots, there is already a bit of a shakedown because collaborative robots have become quite an important part of industrial robots.
And collaborative robots, I mean robots that have sensors that allow the robot to sense whether they have collided with a person and then stop. And that means that now you can have, in a given workspace, you can have a robot potentially working with a collaborative robots without any fences, whereas in the past, industrial robots, they're rigid, they are heavy duty, and you put your arm in front of them and they're going to just go right through it. That's why they have to have fences around them.
So this new generation of robots, of industrial robots called collaborative robots or co-bots for short is I would say a big disruptor that is allowing businesses that before may have not considered getting into robotics to now jump on the bandwagon.
And I may also add the other big trend is also ease of use. So one of the things that originally collaborative robots were associated with was this fence-less environment, being able to work side by side with humans, and as these robots became more and more present in the industry, it became clear that it's not just that aspect of it, but also the fact that they were easy to program and easy to integrate with, let's say grippers, PLCs, and other equipment.
In fact, another big trend is the development of ecosystems where you basically get a robot and then you go to the ecosystem portal and you start picking which gripper and which vision system and which sensors you're going to add to your cell or solution, knowing that these particular items have already been validated by the robot manufacturer, and so they're going to be easy to integrate. They're essentially plug and play.
Stefan Guttensohn:
Could you give us an example of a robot in a particular ecosystem?
Yarek Niedbala:
So the ecosystem itself is, call it this portal where you actually pick the different items that basically make that robot useful for a particular application. Once again, a robot is nothing but a bolt. It's just a part of a machine. You do need a gripper, you do need conveyors. You need sensors to know when a part has arrived at the end of the conveyor for the robot to pick it up. There's also safety equipment that needs to be added. All of that is basically part of a complete cell.
And so all of these different types of equipment in the past, you basically would have to engineer the interfaces between them, whereas with the collaborative robots and the ecosystems that the manufacturers have developed, essentially portals on their websites that allow you to pick the different components and allow them to plug and play. So it's almost like if the different devices talk to each other very easily, making the job of integration of developing these solutions much easier.
Stefan Guttensohn:
What in particular makes a co-bot so unique? You mentioned there's a plug and play aspect, but it sounds like there's an interactivity.
Yarek Niedbala:
Yes. Yeah, there are several things because you have sensors in the robot or the ability to sense, it allows you to, number one, have people close to the robot working close to the robot because the robot now, if it happens to bump into someone, it'll stop. It'll make sure that it's not going to injure a human.
In addition to that, because the robot can sense, you can actually teach that robot different positions or how to move by taking the hand of the robot and moving it around so it records the positions. Whereas before, you'd have to use what we call a teach pendant and keys or joysticks to basically move that robot. And it was easy for me, having an engineering background, but it might have been a little bit more daunting for someone that has not done it before. Whereas this, by just grabbing the robot and moving it around, it may be a little bit more natural for them and easier.
Stefan Guttensohn:
If someone has one of those robo-vacs and they want to move it to a different place, instead of just programming it to go, they might just pick it up and move it over. It's as simple as that.
Yarek Niedbala:
Exactly.
Stefan Guttensohn:
And identifies it surroundings and continues.
Yarek Niedbala:
Exactly. That's right. Yeah, and once again, when I'm talking about collaborative robots, I'm really talking about still six axis articulated robot arms.
Stefan Guttensohn:
Let's break that down quickly. I'm not quite clear on all the fields of robotics that exist. You mentioned industrial robots. What is there in terms of robotics?
Yarek Niedbala:
So there are quite a few fields of robotics. There are special purpose robots that have been developed to do a very specific task. They're still reprogrammable, but it's a very specific task. So they have a form that is quite unique. An example could be oftentimes a snake robot to basically go inside pipes. It literally looks almost like a snake. It is comprised of multiple joints and multiple links between the joints.
There are also now cable robots. So imagine almost like almost like a replacement for a gantry where the end of arm, let's say, is moved around three-dimensional space, but in six degrees of freedom just by basically pulling different cables. Those are, for example, a little bit more rare.
And you have service robots and a service robot could be something as simple as these days, you go to some of the restaurants and you see a trolley that basically is bringing food automatically to their guests or taking their dishes away. By the same token, these same kinds of robots, service robots would be used in hospitals to distribute medication or take linens from the room to the laundry area. So those would be service robots.
Stefan Guttensohn:
And these all fit within the category of industrial robots that you had mentioned before. Is that correct or is it?
Yarek Niedbala:
No, I would say that those would be what we would call service robots.
Stefan Guttensohn:
Oh, I see. That's separate and apart from industrial robots?
Yarek Niedbala:
Industrial robots tend to be more associated with manufacturing.
Stefan Guttensohn:
And what is currently propelling the advancement in robotics? You mentioned how co-bots is one of the trends that you're seeing. What is it in particular that is so revolutionary?
Yarek Niedbala:
It is all about sensors, the actuators, and of course you mentioned it yourself earlier on, is AI, right? So when it comes to sensors, oftentimes those will be cameras, 3D cameras. They have become a lot more robust, more powerful, and of course less expensive. So now you can develop applications that before would've been prohibitively expensive to develop.
You have the actuators, which let's say you can call these basically the motors and inside each one of the joints of our robot, that's basically one example of an actuator. But in different types of robots, like in cable robots, those would be basically the cable tensioning mechanisms that exist. Once again, there's an advancement in those devices that allows, let's say, to have robots that not only have motors and gears inside of them, but also torque sensors. And all of this is basically minimized to make small compact module that allows the robots to be small and not heavy, therefore making a great collaborative robot.
Stefan Guttensohn:
The age of hydraulics is over. Is that it?
Yarek Niedbala:
Well, actually, very interesting that you bring that up, Stefan, because the first robot that was developed by... Well, let's say commercialized by Joseph Engelberger back in the sixties, it was first invented in the fifties, but commercialized in the sixties was actually the Unimate robot. That was a hydraulic robot. But yeah, I think that sometime in the seventies already we were looking at electromechanical robots. Not so many hydraulics anymore.
Stefan Guttensohn:
Does that mean that the cost has also been affected? I'm assuming from what I'm hearing here, that there is a positive cost consequence to all of this?
Yarek Niedbala:
Yes, yes. Cost has diminished because technology has diminished. Certain things, of course, are not going to necessarily diminish in price. For example, the links of our robot, they're made often out of metal and metal is metal. You're not going to be able to get that price down significantly. But processing power on the controller side, yes, you can. And some of these actuators that we were discussing also can be reduced in cost because once again, advancements in processing and materials as well.
In addition to that, we're talking about sensors, actuators, AI. There's also end effectors, once again, end effectors can make a huge difference on the type of applications that a robot can perform because it does give the robot flexibility. For example, an arc welding robot has a very simple end effector, which is just a torch, and that robot is going to do one task only and is going to arc weld, unless of course you dismount that torch and put something else at the end of that arm.
But in some cases, what you want is you want to have more flexibility, especially when you're handling parts. You can develop a vacuum gripper, but if that vacuum gripper can change its shape to be able to pick up different types of parts or different sizes, now all of a sudden you are able to handle a greater variety of parts. And so now all of a sudden, by having end effectors that are more flexible, you are allowing companies that don't necessarily mass produce like the automotive industry to now be able to handle small batches.
In fact, sometimes batches as small as batch size one. A good example would be in the cabinet making industry, you have a process where the designer designs a cabinet. That information is sort of exploded into parts and sent down to the CNC machine that takes a sheet and cuts it, and the robot has to unload that. If the robot knows that it's programmed to basically pick up always from the same location, that's simple, but unfortunately, that only allows the robot to do that particular cabinet. But now if the pattern that gets cut on the CNC is different every time, the robot needs to be able to not only see what has been cut, so that's where the vision becomes important, but also have a gripper that can pick up large parts, but also small parts. So the ability of that gripper to change is quite important as well.
Stefan Guttensohn:
The CNC machine, that is the tool that is used?
Yarek Niedbala:
For cutting these, it's basically a CNC router. It's basically used for cutting these sheets into the different parts of a cabinet so that it can be later on assembled.
Stefan Guttensohn:
Oh, so that's how woodworking comes into play here. I was curious about that.
Yarek Niedbala:
That's how woodworking comes into play. And I am telling you that's an industry that I have followed pretty closely because the woodworking industry is very different than the automotive industry in terms of not too many companies are mass-producing to the same level as the automotive industry. In fact, there are many shops that are doing batch size one production. This is where flexibility has permitted these smaller companies in different industries to now start using robots.
So robots right now, industrial robots are no longer just for the automotive industry. They're being used in many different industries, and as you mentioned, construction, agriculture, and so on and so forth. Industries that require more flexibility and be able to change quickly between batches.
Stefan Guttensohn:
Oh, I see. And to clarify, when you say batch size one, you mean there'd be one order out for say, 50,000 nails or 50,000 widgets and that's it. And that's the extent of the order. So that would be?
Yarek Niedbala:
Right, that would be one batch. Yeah. So batch size one, so you've been to 50,000. Basically that's the size of that batch. But because of the flexibility, it is possible in certain processes to basically do a batch size one, meaning that the batch contains one unit. So let's say today or this hour, I'm going to be building this cabinet, and then the next cabinet that gets built is a completely different cabinet.
Stefan Guttensohn:
So you can customize, you can have a series of different cabinets built by the same machine because it adjusts for based on the instructions?
Yarek Niedbala:
Based on the instructions, based on the data. And that's actually another, let's say a big, big advancement. It is just how much data we can handle now versus what we could several years ago, because a lot of production now is more data-driven, right? So imagine being able, like I said before, to take someone's concept like someone's cabinet and explode it into the different parts that are used to assemble that cabinet and that information gets sent out to the factory floor. And that is basically what the machines use. That information is what the machines use automatically to self program themselves.
Stefan Guttensohn:
And I would imagine that that results in efficiencies and significant cost savings. Is it affecting the cost of furniture?
Yarek Niedbala:
It certainly should, yes. Actually, is it? This is interesting because it is either affecting the cost of furniture or let's say the price of the furniture, or it's affecting the margins that some of these companies can make, or it's allowing some of these manufacturers to be able to compete, let's say manufacturers in the US to be able to compete against manufacturers from China.
Stefan Guttensohn:
Oh, interesting.
Yarek Niedbala:
In fact, this is another very important point that we have to consider. A lot of the discussion about reshoring is now possible because automation does allow you to be more cost-effective and be able to now not compare your manufacturing costs on a labor basis, but on an efficiency basis because you are minimizing the labor at that point, the labor content of the manufacturing. You know what I mean?
Stefan Guttensohn:
Yes. And labor is an issue in a lot of industries nowadays from service up to especially highly skilled in various sectors. So what I understand then is that robotics is at least a partial solution to the problem where we have these costs of robotics being not controlled, but reduced from what they were. I would imagine that the cost of robotics a decade ago was far greater than it is today.
Yarek Niedbala:
Yes.
Stefan Guttensohn:
And so you have robotics becoming far more accessible and it's dealing with labor issues, particular what I would expect, manual labor, people on the assembly line floor.
Yarek Niedbala:
And it's not just simple labor as in pick and place, but it's also quite frankly, complex labor. And here I'm talking about processes. I can tell you that in the US in the next five years, there's going to be a deficit of 300,000 welders. That's a lot of positions that are not going to be filled, and there needs to be a solution, and robots can actually fill that solution, which means that the welders that remain behind they're going to be able to be essentially managers of these robotic welders. They'll be able to take their skill, which is to weld, understanding the materials, understand the weld process, and apply it without necessarily doing all the welds themselves.
They can still do some welding of course, there's going to be instances where welding is going to be required, but the repetitive welds is something that no welder wants to do, and that's going to be done robotically. It is already being done robotically, but in the case of this huge deficit ahead of us, it's very important to have a solution.
Stefan Guttensohn:
And how pervasive is robotics nowadays? I would imagine that you would find a robot in just about every industrial facility.
Yarek Niedbala:
Yeah. So in North America, there's about just over 50,000 robots that get sold. Just in the US you're looking at just over 40,000. And a lot of those robots, they still go into the traditional industries like automotive. Electronics is also another big one. But right now you can see robotics in a lot of different industries. Anything from fast-moving, consumer goods, consumer products, metal working, job shops, and as I mentioned before, in industries that are quite unique like architecture and entertainment and agriculture, construction as well.
Stefan Guttensohn:
Let's just talk about that for a second. How do robots play a role in architecture?
Yarek Niedbala:
I was lucky to be invited to an event ROB|ARCH 2024 that took place in Toronto. It's an international event that brings a lot of architects into one spot, and it is unbelievable the workshops that they had. Our company provided robots as well for them to be able to do these workshops. And it was mind-boggling in how many ways they could use these robots.
One example that comes to mind right away is being able to use the robot for a million large parts that basically become ornaments. Once again, it'll be a batch size one. Someone basically develops a sculpture and rather than go at it with hand tools, they basically use the robots to essentially mill that sculpture out of whether it's foam or in some cases out of stone.
In fact, I don't know if you're aware of this, but the Canadian Parliament right now is being restored using one of our robots that is milling sandstone. So a lot of the statues on the outside of the building because of environmental conditions and so on and so forth, and age, of course, they have been deteriorating and they need to be replaced. And we basically have our robot milling these parts, milling these statues. So it's basically building restoration, which is part of architecture.
Stefan Guttensohn:
I mean, I can imagine the application of robotics to entertainment, but I'm curious, where have you seen robotics applied in entertainment?
Yarek Niedbala:
In entertainment? One project that I myself took part in was a filming robot where we ended up taking one of our medium-sized robots. We put it on a rail or a slide so that the robot had a longer reach, let's say, and there was one of these green rooms where they were filming all kinds of programs or sports or scenes or videos, and the robot would basically be filming. You essentially were able to program the robot to take really interesting angles and essentially became a camera man. And that was actually basically just for filming.
But for commercials, sometimes you have these really wild shots using high-speed cameras. For example, milk is dropping into a bowl and you see it splashing and it looks really cool, but to make it interesting, now the camera is basically doing like a 360. That actually is something that can be done with a robot. And I have seen that done with a robot as well because of how fast the robot can move and how precise the robot can move.
Stefan Guttensohn:
So what makes this robotics is the fact that these scenes are programmed in advance. So there's a determination made about where the camera needs to be, and that's programmed in, and then when the filming starts, the camera moves based on the instruction, so there's no one manipulating it at that time.
Yarek Niedbala:
Correct.
Stefan Guttensohn:
Is that what makes it a robot?
Yarek Niedbala:
Yeah, correct. Now, and here's when we start talking about sensors. Now instead of actually having to pre-program that robot, let's say it's a more free-flowing video that someone wants to do. So what if that person now in their pocket, they have an emitter and there's a receiver that's picking up that information, and based on the location of that emitter, that information is sent to the robot that in real time is following that person.
Stefan Guttensohn:
You mentioned before the camera. How did that particular project begin? Who came up with the idea? Did someone call you and say, "This is what we want?" Was there a collaboration?
Yarek Niedbala:
In that particular instance? Yes. The owner of the business had a vision and they reached out to us. Quite frankly, these kinds of applications have become popular enough that now customers are actually calling us because they want to see something similar, not just for filming because you're talking about entertainment, not just for filming, but also for handling actors in different scenes. For example, the movie Gravity with Sandra Bullock, when she was twirling around in space, she was actually being twirled by a robot. Sherlock Holmes and Moriarty at the end of that scene of the movie, when they were basically falling from a building, that was once again also done with two robots. So the robots were actually manipulating the people at the end of their arm.
Stefan Guttensohn:
I'm trying to envision opportunities and location-based entertainment where I would imagine that the need for robotics would be really high. And how would someone who's trying to make an environment happen determine what they need? How would they even know if robotics is feasible from an economic perspective? Would they just call and say, "Here's our idea. Could you provide us with a budget? Do you have a sense of how much this would cost?" How do these conversations happen?
Yarek Niedbala:
Yeah, no, that's exactly how they start. They do reach out to us. We do have a discussion with them, and oftentimes what happens is that we typically end up working with one of our partners that already has experience in the entertainment industry. And once again, it depends on what kind of a robotic application within the entertainment industry they're looking for. Because there are a variety of them.
For example, we just talked about two of them. But as you know, Disney and Universal Studios, they use our robots for puppeteering. Essentially, they need to have a certain level of flexibility. Rides no longer can last 10, 20 years. Right now, people are changing their tastes. New movies are coming out, and the rides need to change to basically keep bringing people back. And the best way of changing those rides is the flexibility that is afforded by robots.
Stefan Guttensohn:
I have to ask you, in your experience, has the life cycle of a robot die or ended before it's retooled or reprogrammed? Or more often than not, these robots are reprogrammed a number of times before the end of their life cycle?
Yarek Niedbala:
Yeah, that's a very good question, Stefan. And I would say that because robots can last a very long time, it of course all depends on the kind of environment that they're in. The type of environment, the type of speed and movements that they make, all of that has an impact on them. But I have seen robots, I'm talking about several years ago, I have seen robots that have been working in industry for well over 20 years. And yes, during their lifetime, for sure they get repurposed.
For instance, some of the robots that go into the automotive industry, as I mentioned before, typically it's a seven-year cycle. So they often go through two cycles, let's say on the spot welding line where precision is pretty important. And at the end of that, because of usage and so on and so forth, that robot can still be useful for something else. And so now all of a sudden, instead of doing an accurate application, it can be used for something a little bit less precise, like for example, palletizing or handling in general, you know what I mean? And of course also in industries that don't necessarily require that sub-millimeter precision.
Stefan Guttensohn:
So using the example of a palletizing robot. That's indoors, so it's not suffering from weather conditions outside, it's in a controlled environment, but it has a lot of moving parts and it's obviously mobile. Would you be able to provide a life of a palletizing robot generally 20 years, 10 years, two years?
Yarek Niedbala:
Yeah, I would say that a palletizing robot should last those 20 years. Palletizing typically happens, it's over two shifts, right? Yeah. So that's basically, that's typically what, 4,000 hours a year? So just to give you an idea, the major maintenance on a robot, which is the oil change, happens every 20,000 hours. So in that particular robot, you'd be basically changing oil after five years.
Stefan Guttensohn:
Wow, 20,000. As a lawyer, I'm very aware of the amount of hours that I have to put in and 20,000, that wouldn't even be five years if you're working at a regular pace. That's incredible.
Yarek Niedbala:
So they definitely last a very long time, and they can be repurposed, and that's one of the benefits of robots. Absolutely.
Stefan Guttensohn:
What are the challenges in robotics? What are the challenges that you are facing right now? And I don't mean just the technology challenges, but perhaps perception from the public at large, economics, high lending costs?
Yarek Niedbala:
So oftentimes there is a little bit of a... Depending on the industry. It is funny because in some industries, robotics is just by default, that's what people use. It's almost like, I don't want to say almost like toilet paper, but it's an essential, just as toilet paper would be.
But in other industries, we see a huge potential. But I would say that oftentimes, especially in Canada, people tend to be a little bit more conservative, like for example, compared to the US. And part of it is also in the US, companies tend to be larger, they have more resources, and the batch sizes tend to be larger as well. In Canada it's smaller, smaller batch sizes. Companies tend to be a lot of small to medium manufacturers a lot. And so it has been historically more difficult, but we are seeing that change. We are seeing more and more of those smaller companies start adopting robots as well now.
Stefan Guttensohn:
Have you seen any difference between the United States and Mexico?
Yarek Niedbala:
Interestingly enough, in the last few years, what we have seen is a huge growth in Mexico, and that once again is part of the near-shoring. So we have reshoring, but we have also near-shoring. And while we see Canada being rather steady, we see the US increasing, maybe not in the last couple of years, only because they're having been some economic challenges, let's say. But I would say that Mexico, we have seen a huge growth, and a lot of that has to do with near-shoring. So taking advantage of both less expensive labor and potentially some easier regulations along with using automation.
Stefan Guttensohn:
I want to finish by asking you the question that I think is on a lot of people's minds right now, and that is there an associated danger with the growth of robotics? Are we coming to an age where robotics might be a threat to civilization?
Yarek Niedbala:
I would say that if we're talking about industrial robots, I would say we are still pretty far from that. Reason being is that robots need to be intelligent in order to be a threat to humans and not, especially if there are no rules that are implemented. Reality is that AI, to me personally seems to be a little bit more of a threat simply because I see the types of, let's say, tasks that it is able to do that allow for more decision-making, creativity, processing. And quite frankly, that to me seems to be a much larger threat.
Now, keep in mind that robotics and AI are slowly starting to merge. Right now for example, just to give you an idea on the industrial side, on how far we are from robots becoming a danger, a simple application like bin picking. You have a bin with a bunch of parts in it. A kid can stick their hand without even looking and grab a part and then accurately place it into a machine. Okay? A little kid can do that.
For a robot, that's a very difficult task. It requires a 3D vision system. It requires a path planner that will automatically guide that robot into the right location. And remember we were talking about that gripper? Well, you need to have a gripper that has the flexibility to be able to grab whatever part is in that bin. And yeah, that's a very big challenge. And this is just thinking, it's something that we do without even thinking about. It all depends on how much AI evolves and how it evolves. What kind of rules are going to be in place for robotics? There's already a set of three rules that Isaac Asimov put together. And of course the first one is that the robot may never injure a human being or through inaction allow a human being to come to harm. Now the question is, are we going to be implementing that for AI? If we do, and if we stick to those rules, we should be okay.
Stefan Guttensohn:
Thank you. Well Yarek, I appreciate your time. This was Yarek Niedbala from KUKA Robotics. If you'd like to speak to him, we'll have his contact information underneath the podcast. And once again, Yarek, thank you. Until next time, everyone.