Episode 76: Ben and Steve discuss subscription services, the lengths they’ll go, and how to stop them. Benjamin says NIST has some cheese for metals-based AM research. Stephen insists robots will soon have watch-and-learn capabilities. Ben is a fan of retrofitting old machines with current features. Steve explains why a robot broke a little boy’s finger playing chess. Benjamin highlights a recent cybersecurity partnership. Stephen has six sigma all figured out.
https://www.dpaonthenet.net/article/191820/Robots-can-learn-household-chores-by-watching-you.aspx
https://screenshot-media.com/technology/ai/robot-breaks-finger/
Explore, watch, read, learn, join, and connect at https://www.imts.com/
Tune in to the AM Radio podcast https://www.additivemanufacturing.media/zc/am-radio-podcast
For the latest in Manufacturing Technology news https://www.amtonline.org/resources
Transcript
Benjamin Moses: Hello everyone, welcome to AMT's Tech Trends Podcast where we discuss the latest manufacturing technology research and news. Today's episode is sponsored by IMTS+. I am the director of technology, Benjamin Moses, and I'm here with-
Stephen LaMarca: Technology analyst, Stephen LaMarca.
Benjamin Moses: Steve, how you doing today?
Stephen LaMarca: What a dopey way to introduce myself.
Benjamin Moses: Eh, it's all right.
Stephen LaMarca: That was I think the worst in a while. I got my name and title right, but it's like, how full of myself did I sound right there?
Benjamin Moses: Fair point. We've been talking about subscriptions recently.
Stephen LaMarca: A lot of subscriptions have been coming up.
Benjamin Moses: Yeah.
Stephen LaMarca: So much so that there is a subscription service for unsubscribing to subscription services. Seriously, there's a lot of sibilance right there.
Benjamin Moses: The circular loop of subscriptions.
Stephen LaMarca: Alliteration and sibilance is our problem of today's episode, but-
Benjamin Moses: There's been some weird ones too. So I think you talked about BMW.
Stephen LaMarca: Dude, BMW... man, this one. So I was swiping through Instagram before bed last night, and I see this meme video of some dude with a really jacked arm is holding the wheel to a BMW and then all of a sudden the vehicle stops. Through the windshield, all of the scenery stops moving and a little credit card swiper pops up and is like, "Please pay to continue driving." And swipes a card and it continues driving. It's like okay, I don't get this at all. What's going on here? Then I go to bed, and then I wake up this morning to, I think it was either the Hustle or somebody that says, "BMW's making you subscribe to use your seat warmers."
Benjamin Moses: That's rough.
Stephen LaMarca: I could care less.
Benjamin Moses: Sure.
Stephen LaMarca: Immediately... okay. So I'll be honest. Immediately, I played devil's advocate because we're in the manufacturing industry, we have an idea of where this came from. We've seen over time since the assembly line, the Ford assembly line to where we are now, you no longer replace parts or when you're assembling a car, you're no longer putting individual parts in at a time. You're doing pre-manufactured assemblies that are installed into the car and think about how much money a car company could save if instead of having all of these different seats, you have one seat, and all seats have seat warmers now. Companies like Kia and Hyundai love doing this with their headlights and the radar systems. It's like even our base model Hyundai Sonata has the advanced tech package that you would find in a Mercedes-Benz, a loaded Mercedes-Benz not a base model. You have to pay extra for this.
Well, it's in a base Sonata because Hyundai's like, "Well, we want to be ahead of the tech. We want to keep up with technology, but it would cost us too much money to make a base model because now we're having to make two models of this headlight, when instead we could save money by making one that cost savings will trickle down to the consumer," because they're a decent company, "and eventually it'll come around and the auto critics will recognize this." They have and we have, and we're in the industry and we realize yeah, it saves you money if you just make one thing as opposed to many different things.
BMW sees this finally and they're like, "Okay, we'll stop making different tiers for different price point customers, but no, we want to charge people differently. We want to separate the plebes and the riffraff from the premium consumers that we want." So now they're doing a subscription service. They haven't rolled it... this leaked in the US. They've rolled it out in some country that's not necessarily a third-world country, but I forget what country it is and they're starting to introduce it into the UK, which I'm like, "What is wrong with them? How did they wave the white flag so quickly to the Germans?" And that's not going to stand in the US, and they're... it was just a leak. It wasn't even an official rollout yet and the US is already not having it, and I couldn't be more proud.
Benjamin Moses: Good. Yeah, it is interesting. The big concern I have is I've been buying a lot of used cars for the past bunch of years, and I would hate... and to be honest, my last couple purchases have been minimal research. I know roughly the car, the horsepower, some reviews about suspension, things like that. So the critical parts that I'm interested in, but now I've got research of, is this feature a paid subscription service that I have to not only buy, but subscribe to afterwards? Some of the cars I've wondered what type of collision detection systems they have. Have to do a little more research to understand does it auto stop because we're in that weird transition where not everyone's the same, but now if I buy a used car in a couple years from now, now we're going to have to research are there any subscription services I have to worry about and are they still maintained.
Stephen LaMarca: Yeah. Sirius XM, everybody knows how awful that is. Just get Spotify or Pandora. What are you doing with this stuff? And that's expensive.
Benjamin Moses: Yeah, it's not cheap. I have that for my wife's truck.
Stephen LaMarca: Bless you. Man, you didn't need to flex on everybody today. Geez.
Benjamin Moses: One thing that I have been doing a lot is I've got a ton of subscription services, which my purchasing habits have shifted a lot. So when I was buying stuff, it was a lifetime contract. Basically, like a video game. If I bought a video game, I would technically have it forever, and I had a physical disk then. Over I'd say the past 10 years, there's been a shift of all services being more of subscription or pay to use. Some video games are still purchased for lifetime, but the digital platform that they're on is a little dicey, but I have been doing a lot of say, self-reflection on all the services that I have and how much I'm paying for those services because it does add up at the end of the year very, very quickly.
Yeah, monthly you may pay a hundred dollars for everything, but that's $1,200 that you may not need every year. So I've been one, rotating all my video services. So I had HBO Max, realized I'm not watching that anymore. Let me turn that off, and then I sign up for Apple TV, which is not bad. Once I finish those series, I'm going to turn that off and then go back to HBO when, was it, Game of Thrones or Lord of the Rings comes out. I don't remember which platform they're on, whatever. I've definitely been rotating my services as I use it. I haven't gotten into borrowing logins yet, but that may come up soon.
Stephen LaMarca: Dude. Man, they're trying to get ahead of that. Don't get me wrong, I've been borrowing logins forever. Especially in the college days, come on, man. It's a struggle, but that now it's cool that there's the... well, I guess it's cool. There's unsubscription subscription services. You can subscribe to a service that doesn't cost you any money. Doesn't mean it doesn't cost you anything but doesn't cost you any money and they will go through your social security number and find out everything you're subscribed to and give you a report back and they're like, "Tell us what you want to unsubscribe from and we're going to do it. You don't have the email to that account anymore? No worries. We got you."
Because it's crazy. If you want to unsubscribe to something, let's say you did get HBO or whatever way back, long ago, and you used obviously a personal email account, but a burner email account and you don't have access to that email account anymore. How do you get rid of that service because they won't let you in? So I'm like, "Yeah but you have my account number and I have that account number. So because I have my credit card number, will you let me unsubscribe?" And they're like, "No, we have access to your bank account, but you don't have access to your email. So we can't verify that it's you." And it's really sketchy like that. They make it really sketch. So it's like this is needed, but it's not really, but anyway, these unsubscribe services don't charge you any money, but I'm sure they data mine the heck out of you.
Benjamin Moses: They're getting your information.
Stephen LaMarca: Which is fine. If you have nothing to hide, data mining's not a big deal. Let's take our tinfoil hats off, but even still, you don't need the unsubscribe service.
Benjamin Moses: There's another option.
Stephen LaMarca: My mom recently had to go through all of this stuff because she's protecting her nest egg. She's retired. She realizes she's got... she had, excuse me, a subscription to Angie's List.
Benjamin Moses: Wow.
Stephen LaMarca: Yeah, back when they paid.
Benjamin Moses: That's interesting.
Stephen LaMarca: And because she's grandfathered into it, they were still charging her even though now it's free because they're data mining. Another one, Rosetta Stone. Her and my dad before he passed went to... they did their final, big trip together. They went to England and France and while there, she wanted to make sure she brushed up on her French before going over there. She paid for Rosetta Stone. Doesn't have access to it anymore, but they're still charging her, and Rosetta Stone is not cheap.
Benjamin Moses: That's expensive.
Stephen LaMarca: They would not let her unsubscribe because she doesn't know what email she used and what burner account she used, and my mom always had this bad habit of creating an email account with the different ISPs that we would use growing up. So-
Benjamin Moses: There were a lot back then.
Stephen LaMarca: Which is really weird that was a thing like, "Oh, we don't expect you to have an email. So we're going to create an email for you, and by the way, when we go out of business, all of the subscriptions are still going to keep going and charge to your credit cards." So that's essentially what happened, and my mom has had a few meetings over the years with her wealth management team and CPAs and stuff like that. They're all like, "Yeah, just if that ever happens, if you ever find another one, use the nuclear option."
Benjamin Moses: What is that?
Stephen LaMarca: Nuclear option?
Benjamin Moses: Yeah.
Stephen LaMarca: You call your credit card company, find out which credit... it's easy to find out which credit card is being charged because you just look at the statement, but you call that credit card company. Be like, "My credit card was stolen." Not lost, not misplaced, not expired, stolen because if it expires or if you lost it, they keep all of the numbers the same because oh, they don't want any inconvenience. They don't want to inconvenience you with having to resubscribe to anything. So they keep the numbers the same, but change the security code, and it's like it's not that they don't want to inconvenience you, they want to keep you hemorrhaging money.
So to stop that, you have to claim my credit card was stolen, and then they change all of the numbers. You don't want to cancel your credit card account either because it's bad for your credit, but you report it as stolen. They send you a new card, they give you new numbers, and almost immediately she was notified by all of these subscription services that, "We're having difficulties getting your payment. You need to fix this." And it's like, "No, I don't."
Benjamin Moses: I'm going to fix it by leaving it alone.
Stephen LaMarca: I did fix it.
Benjamin Moses: That's an interesting option. Yeah, I had to think about that and to be honest-
Stephen LaMarca: It is a matter of time before the credit card companies change that one though.
Benjamin Moses: Yeah. We'll see.
Stephen LaMarca: That's terrifying.
Benjamin Moses: Yeah.
Stephen LaMarca: I hope they don't. Maybe I'm just being... maybe I have my tinfoil hat on right now, but that's the nuclear option.
Benjamin Moses: That's the nuclear option.
Stephen LaMarca: Don't be afraid to use it.
Benjamin Moses: It was getting out of control. Push that button.
Stephen LaMarca: Dude, hit the button. Get the football out.
Benjamin Moses: Steve, can you tell us about our sponsor for today?
Stephen LaMarca: IMTS+, the people behind technology. The stories driving the future of manufacturing. The thought leaders and people like us creating the products, the opportunities, and solving the challenges of our industry. Explore a new digital destination designed for the manufacturing technology community, where you can watch, read, learn, join, and connect. Go to imts.com.
Benjamin Moses: Thanks, Steve. I got a bunch of articles today on a variety of topics. The first one is from NIST, awards nearly 4 million to support metal-based additive manufacturing. I like this on a couple of layers, one-
Stephen LaMarca: Just awards or they were awarded?
Benjamin Moses: They're awarding. So they have a couple of research grants that they've awarded, but also their own internal stuff.
Stephen LaMarca: 4 million, dude.
Benjamin Moses: So they've got... the article lists four different awards. Actually, this totals about 4 million. So this is how much money they're giving you.
Stephen LaMarca: That's eight titanium blisks for the F35.
Benjamin Moses: Good callback. So of course, NIST is managed through the Department of Commerce, and what they're looking to help address is the current and future barriers to the widespread adoption of metal-based additive through measurement science research. I think that's a subtlety that... actually there's a lot of barriers for additive, and they're looking at the measurement side of it. So through their own research and with these grants, and this is addressing barriers of adoption of additive manufacturing, including measurement science to support equivalence-based qualification and model-based qualification, characterization of AM materials, and standards to support consistent data exchange characterizing new advances in AM production. So they actually hit onto very key elements that we've talked about before. Characterization of AM materials. So we talked about standards for AM materials, but obviously the value of AM is being able to pivot from material to material and develop some new materials. So the characterization's very important.
Equivalence-based qualification and model-based qualification. So in aerospace we do that a lot where I have a previous design, I do a modern modification. I say, "This new part is equivalent to the old part. Therefore, it's qualified by similarity." And I think that's a very good approach, and consistent data and then obviously we're talking about, I have a machine. I'm printing parts, how do I get data off that machine? So I think this is very holistic approach of the need for additive to be more production ready. Obviously, it's still supporting low volumes and tooling, but when we look at additive metal-based, obviously we're still stepping into production environments, and I think these are foundational steps that NIST is looking at.
Quick highlight on some of the research, The Research Foundation for the State University of New York, they're getting $900,000 on enhanced nondestructive testing techniques. Colorado School of Mines, my favorite school of mines, project to examine new optical metrologies to enable real-time process feedback and control. They're getting also about 900K. Auburn University established data-driven framework with computer vision and machine learning for nondestructive qualification of AM materials, and GE Research out of New York teamed up with University of El Paso, established Intelligent Stitch Integration for Testing and Evaluation. So pretty broad spectrum of projects that the investment is serious. So I do like the potential for a lot of these turning into standards, which obviously NIST is going to crank out from this research to further push the adoption of AM.
Stephen LaMarca: Right because... I have a question for you. We love standards, and I love that the need for standardizations in materials is not only there, but it's gaining traction and it's moving forward, but my problem is that's confusing to me and might be confusing to others is standardizing a material isn't really beneficial to additive because one of the beauties of additive is not only being able to use multiple or any material, but being able to use the exact and perfect material for whatever part you're trying to make. I don't think a standard would necessarily benefit a class of material, but is the standard for materials a standard in the quality of the material, like the powder, like the grain size, stuff like that?
Benjamin Moses: So I think what they're looking to is understand the characterization and understand what are the key elements that they need to make sure that they're quantifying through standards. So they talk about splatter particles, grain size, potential defect detection, oxide thicknesses. So they want to understand what are the key elements, and then probably develops standards on those key elements that roll up into different materials. So I agree with you. Well, it's two sides. If you are looking at standardizing the material itself, so what does a stainless steel 321 in powder form look like? That's a standard versus the components of that material could be different constituents, but how you characterize each of those constituents could be standardized.
Stephen LaMarca: Right. How close are we to AM machines or 3D printers? Instead of having a powder vat or powder hopper for Inconel, having multiple powder hoppers of just raw, perfectly grain-sized just elements and the machine just blends the material for you in the machine.
Benjamin Moses: For AM, for-
Stephen LaMarca: Are we a hundred years out on that?
Benjamin Moses: For metals, you're a ways out. For plastics, I think that's achievable pretty soon because you-
Stephen LaMarca: Wow. Really?
Benjamin Moses: Because you could do by the pellets, I think. With metals, it's a little... because you have your elements. So you're mixing smaller size than your-
Stephen LaMarca: It'd be dangerous if you had a hopper of potassium. You better keep moisture out of there.
Benjamin Moses: Yeah, I agree. That is an interesting point though, when can we mix on the fly.
Stephen LaMarca: Because I'm sure there's some big brains thinking about that. It's like, "Let's just break it down to the elemental."
Benjamin Moses: So you could do multi-layers right now. So you could shift from I'm doing Inconel 625 and then shift to something lighter like a 300 series a couple of layers later and shift back and forth. That exists now, but to have one part, one material from five different vats and then shift that different print and then just basically make your materials of the fly. I don't know, man. Seems a ways. That's like Star Trek type thing. You've got an article on machine learning.
Stephen LaMarca: Oh, I do. Let me queue it up because that one, that was a good article, Ben.
Benjamin Moses: I like what NIST is doing. We need to visit them more often. They're way down the street.
Stephen LaMarca: Yeah. I was telling people over the weekend that NIST has multiple clones of Newton's apple tree. They legit have the Newton's apple tree, a clone of it, which is just a graft, but because it's fun, you just say clone. It's effectively the same thing, but yeah, whenever somebody does something special or somebody donates a lot of money to them, it's like here, have Newton's apple tree. It's like wait, what? Okay. So yeah, I have an article that touches on machine learning for automation via vision systems, and the article's called, Robots Can Learn Household Chores by Watching You, and I just thought it was really cool because we know that robot arms, industrial robot arms can be taught by clicking a button and guiding the arm, physically guiding the arm to do a task or setting checkpoints as it does a certain motion to avoid collisions and do this. You can teach a robot arm, but it's cool that this article talks about you just need a robot with a vision system and it needs to watch you, and they're talking about for household stuff, but let's take this to an industrial level and-
Benjamin Moses: Yeah, watching them do an assembly. Pick and place and screw things together. That's an interesting technique. Creepy at the same time, but it's fairly interesting.
Stephen LaMarca: It is a little creepy. Don't get me wrong. That definitely crossed my mind like, "Oh my God." Yeah, but I think it's cool and the article also goes on to say that the robot can also learn by trial and error.
Benjamin Moses: Yeah, definitely. So I've definitely seen that technique a lot. A lot of pick and place from a random bin of parts. It learns by taking picture, and actually there's a couple of techniques a bunch of years ago where they had it train itself and they're determining how long and how many pieces it needed to train before it got to 90% capable, and then I feel that the training sets now are much smaller and they're much more efficient to get to that, but that is an interesting new technique for teaching automation is one, you could just do backend programming that always exists, which is great. I like that.
You could teach on the fly by moving the pendant around, but now teaching by watching basically, learn as you go, and I think that's another way to increase your speed of automation when you look at low volume, high mix scenarios. If you're changing different assemblies a lot, if you do a different part number every couple of shifts or every shift, but you can teach the robot for one part or a couple of parts and then have it complete the shift the rest of the time, I think that's a fantastic way to implement some automation.
Stephen LaMarca: Heck yeah.
Benjamin Moses: I've got one on, a little old school.
Stephen LaMarca: Oh, wait a minute. Wait a minute. I thought of something, but I didn't want to interrupt you. I thought one of the coolest demonstrations of machine learning, maybe not machine learning, maybe that had nothing to do with it, but automation and vision systems being able to detect defects was back at a MTConnect Technical Advisory Group tag meeting before it turned into the Standards Committee and we got it ANSI certified, I think it was at Notre Dame where we saw this. You were there. There was a vision system that showed a picture of two parts, but a 2D picture of parts and the top of the part looked almost exactly the same as the bottom of the part if the part got flipped over, and looking at the pictures, as a human, you couldn't tell the difference, but they went into the process of programming the robot's vision system to be able to tell. It was fascinating for me because I was looking at that and it's like, this robot's eyes, not even talking about thermal vision or anything fancy like that. This robot's eyes are legit better than a human's.
Benjamin Moses: That's fair. Yeah.
Stephen LaMarca: That was years ago, man.
Benjamin Moses: I've seen a couple use cases on one, optical measurement, so that exists. We've seen a lot of use cases where they're actually able to measure parts, see if it's in spec and out of spec. We're also seeing a lot of cases where you can teach. So checking assemblies or teaching pass/fail criteria on the fly. Back to your point on learn by observing for machines, there's a couple companies that offer basically EnCE issued training sets. So as the part's being inspected, the operator's saying pass/fail, and the machine vision's taking a picture and recording the pass/fail criteria, and I think it needs a small data set considering all things that maybe it's 20 or 30 data points and then it starts applying that data set for new parts.
So it's able to... basically, you're able to teach the machine for that set. So back to your point earlier about implementing automation on the fly, when you talk about automating inspection, that's a very interesting and real-time use case where you could check 30 parts and have the machine vision check the lot for the rest of the shift. So I thought it was fascinating. Thanks, Steve. Good article.
Stephen LaMarca: Thank you.
Benjamin Moses: I've got one on retrofitting breathes new life into old machines. I like a good retrofit.
Stephen LaMarca: Dude, heck yeah. I know I keep talking about this, but the first time that I went to, and the only time and but I plan on going back, the first time I went to the American Precision Museum in Vermont. I loved that they showed me Serial #1, the Bridgeport E-mill. The very first Bridgeport E-mill, and then just behind it, when you're looking at the first Bridgeport E-mill at the exhibit, at least it's the way... they've probably changed it now. If you turn around, right behind you is a functioning, being operated Bridgeport E-mill that has been not only running MTConnect and not only outfitted with a digital readout, but automated axis. They put stepper motors on the axis. So it's fully CNC. You don't even have to operate it. You can do it yourself. It's a manual machine, and they've retrofitted to CNC. So I love a good retrofit too. Let's hear it.
Benjamin Moses: So this article from IMTS talks about a manufacturer over in Ohio. They make process equipment for physical separation steps in the corn wet-mill process, and they've patented a bunch of things for corn to ethanol, dry grinding process. What they're making is very large bowls, basically machine bowls to the scale of... they have three boring mills with more than 10 feet of travel. You talk about traveling in the feet range-
Stephen LaMarca: That's a lot.
Benjamin Moses: That's big. 90 inches of vertical travel and five very large vertical aids.
Stephen LaMarca: Is this like a gantry machine? No, it's a lathe.
Benjamin Moses: Oh, no. They have some boring mills.
Stephen LaMarca: Okay.
Benjamin Moses: So yeah, it's... and when they talk about stuff that scale, it's very, very expensive. So their plan is actually to purchase a lot of used equipment. So they're able to buy one of the 10 foot boring bars for $35,000. A little bit less than my used car, so that's impressive.
Stephen LaMarca: That's a lot of money for a boring bar.
Benjamin Moses: Well, brand new, 1.3 million. So when you're talking about the cost of the machine-
Stephen LaMarca: Dude, that's a huge savings.
Benjamin Moses: Right, but that also includes a dated machine. So that's, I think, the big driver for the article is that they're able to get the good quality metal, good quality ways, all the physical characters of a good machine and retrofit a new controller or retrofit other modern equipment on it. So that's their business strategy, and I thought that was a very intelligent way to approach, maintain the revenue because I'm assuming producing process equipment, that's not super high margin stuff. So all your components and cost to keep your equipment cost low makes a lot of sense, and you're able to buy a big machine for 35,000 and then you could spend another 100K on retrofitting it and it's still significantly cheaper than a brand new machine. Not that I'm saying you shouldn't buy new machines, but if you have the capability to retrofit something, that's not bad.
Stephen LaMarca: $35,000 is a fraction of one and a half million.
Benjamin Moses: I thought that was a great article about being able to if you got old piece of equipment is revitalizing that equipment to make it one, more usable, but also be able to standardize all your controllers, being able to get data off your machines. So that's a very good look at... there's a lot of manufacturers that have equipment that's been running since World War II that run great. It produces a very, very good part, very consistent part, and we'll talk about some quality stuff at the end of the episode too. They're able to maintain the level of quality and process control that they want, but it's mainly maybe manual operated or maybe it's very old controller. So being able to retrofit the equipment to modern standards, that's fantastic.
Stephen LaMarca: Well-
Benjamin Moses: You got one on a seven-year-old that doesn't get lucky with a robot.
Stephen LaMarca: Yes. So this actually came across my social media just like the BMW thing did. Last night or may have been Monday. Anyway, I saw something shot across my bow on... overseas, there was some chess, not chess tournament, but a chess demonstration and some university, I think, was demonstrating their robot that can play chess. The seven-year-old boy goes up to make his move on the chess board, and then his hand gets pinned down by the robot and they said that the video was graphic and it really wasn't that bad, but I think they have to say that when it happens to somebody that's a minor.
Anyway, yeah. The kid gets a broken finger right from it and when you see what happened, the little boy makes his move on the chess board and then does something that's technically illegal in professional chess, then decides to not make that move and move it back, and at that this point, the robot had already started moving and decided to take the spot where the boy decided not to move and ends up... the boy's arm and the robot arm collide. This robot ends up breaking his finger. So immediately looking at this video, actually not even the video, looking at the still shot, I see that's not a collaborative robot. That robot doesn't have collision detection. That's a full-on industrial robot. That's mistake number one. That's red flag number one.
Benjamin Moses: There's a lot of mistakes in the scenario.
Stephen LaMarca: I'm not going to give the little boy crap because he's seven years old.
Benjamin Moses: There're parents that should be involved in the situation.
Stephen LaMarca: Yes. That's-
Benjamin Moses: Mistake number two.
Stephen LaMarca: Mistake number two. Good job. See, parents know this stuff, I don't. Mistake number two. I am not going to fault the boy's illegal move because the illegal move was the parents letting the boy... then I see another picture of the same... it's the same picture, but it zoomed out. Not only is it not a collaborative robot because an industrial robot does not have collision detection. So if it was a cobot, it would've felt that, "Mm, this doesn't feel like chessboard. This feels like small human bones. Let's back up. Let's throw a flag and let the operator come over and reset everything." That didn't happen because it's an industrial robot that is assuming that it's in a perfect world with no soft targets around it. Anyway, I see the zoomed-out picture. There's a safety enclosure around the chess boards. So not only is it not a collaborative robot with a void of any collision detection, but this little boy was given permission to go within safeguarding, safety fences. So-
Benjamin Moses: He broke all the rules on the chess board and the safety rules.
Stephen LaMarca: That's like sending... would you send a seven-year-old kid to go inside a CNC machine?
Benjamin Moses: No. Come on.
Stephen LaMarca: Inside the safety enclosure?
Benjamin Moses: Exactly. That's unfortunate. There's a lot of things they have wrong.
Stephen LaMarca: It's unfortunate, and I feel bad for the little boy. Shame on the parents, but I feel bad for the parents too. That kid's not going to want to play chess again. Some kids grow up and as adults are terrified of dogs.
Benjamin Moses: He's afraid of chess boards.
Stephen LaMarca: This kid's going to be afraid of chess, going to be afraid of robots.
Benjamin Moses: That's unfortunate.
Stephen LaMarca: Probably going to be a politician that is anti-manufacturing.
Benjamin Moses: That's a good investigation, Steve. I appreciate the thoughtfulness of getting past the headline of robot just breaks seven-year-old boy's arm.
Stephen LaMarca: Right. We saw the problems.
Benjamin Moses: Yeah. So the last article I have, Steve, is Rockwell Automation Partners with Industrial Cybersecurity Expert Dragos, and the big takeaway... it's from Manufacturing.net. The big takeaway here is we're starting to see a progressive trend of automation companies or companies dealing with physical and software automation partnering with security experts. Episode or two ago we talked about some other partnerships related to that and how important it is when they're transferring data, being able to make sure it's secure. Now, these guys are looking to partner for a couple levels. One, they have integrated control systems and incident responders from Dragos, field service engineers, and the global network from Rockwell and the coordinate from Rockwell project management for any issues.
So it's a growing trend where I think in our sector, we're slowly talking more and more about security since the tail end of it is all the digital information that we've been collecting. Now, we're seeing a lot more remote connectivity with our manufacturing equipment. So it's a short article, but I see a growing trend in automation companies doing a lot more collaboration with security companies and vice versa, and I think we'll see a lot more of that in the future. As NIST pushes their cyber security framework and the Department of Defense pushes their CMMC, there's a strong need for security in manufacturing.
A couple episodes ago, we talked about the military printing on ship. The Navy printing replacement components for drones and I think that's a fantastic use case. It's a drone so some of it's quite critical because it's for drones. Being able to print stuff, but then also the metrology side of it. It's still questionable, but being able to test that, but the whole data stream to get to the ship and then how do you control all the data on the ship. So it's very interesting look of where security plays in the future of manufacturing and there's a lot more discussions and collaborations to occur. So that's a good one. So to touch base on quality, Steve.
Stephen LaMarca: Yes.
Benjamin Moses: You got something that I'm a little nervous to talk about because it's a little outside of my strength, but-
Stephen LaMarca: Well, it's confusing. It's okay to admit that it's confusing. I'll admit that I'm going on eight years in this industry and I still, at least until this day-
Benjamin Moses: Today.
Stephen LaMarca: Maybe. I've felt uncomfortable explaining to somebody when they ask what Six Sigma is.
Benjamin Moses: Okay.
Stephen LaMarca: The first time somebody explained, at least attempted to explain to me what Six Sigma is in layman's, they were like, "Oh, it's using data monitoring. It's monitoring data to ensure quality on either a manufactured good or even a customer service." And I'm like okay, that sounds like what everybody does. Everybody looks at data to ensure, well, at least attempt to ensure quality control. Now, as I learned more, it's like oh, if you achieve a certain high degree of quality, now you're starting to meet Six Sigma, but anyway, I'm at my desk late last night working on the Tech Report, and there's three people left here in the office. It's Nina. She's about to leave. Actually, no, Nina had just left. It's just me and Jesse Tran, the FNG, the new guy. Jesse is hanging out with... he's about to leave. He got all his stuff together. He's got a little lunch pail.
He's about to leave and he's hanging out with me in the tech pod while I'm finishing up the Tech Report, and he sees a book, I think, in between two desks, unoccupied desks, RIP, that says something about Six Sigma. It's a Six Sigma book, and he's like, "Stephen, what is Six Sigma?" And I'm like, "Can you look at another book and ask me?" Of all the books to look at-
Benjamin Moses: There's a bunch of books in our area, and that's the hardest one to ask.
Stephen LaMarca: At this moment I had just finished the Tech Report and I'm like, "You know what? Let's learn this together. I'm going to put an end to this confusion and use Google, and I use an AI tool to summarize long-winded articles."
Benjamin Moses: We'll get in that after your article. That's really cool.
Stephen LaMarca: We can talk about that later. It's experimental. We're just fiddling with it, but I'm like, what better time than some academic essay of an... multiple, I looked at three articles trying to explain Six Sigma and they were all like you had to scroll, scroll, scroll just to get to the bottom of them and you're like, "I can't read this in five minutes."
Benjamin Moses: Can I point something out real quick?
Stephen LaMarca: What up?
Benjamin Moses: I'm glad that you took the time because I would've told Jesse go read the book.
Stephen LaMarca: So out of this research that I did with Jesse's inspiration, I wrote a small, two paragraph article, articlette summary of explaining WTF is Six Sigma, and all right, I'm going to read it out loud.
Benjamin Moses: Go ahead.
Stephen LaMarca: First paragraph, the Six Sigma quality improvement approach was developed in 1986 by Motorola. The main goal of Six Sigma is to reduce defects in manufacturing and service processes by using a data-driven approach. The Six Sigma methodology includes a number of quality tools and techniques that can be used to improve process quality.
Benjamin Moses: Nice. That's a good overview.
Stephen LaMarca: That's the intro. Now, let's do the outro because the intro still leaves me with some questions, and the main question is, why Six Sigma? What is Sigma? Why are there six of them? We mentioned tools, we just mentioned quality tools used. Are there six tools? No, there are five. So now it's just got more confusing trying to answer this what should be a simple question. Paragraph two. Final paragraph, don't worry. The term Six Sigma comes from statistics and refers to the fact that in a normal distribution, 99.99966% of data points will fall within six standard deviations of the mean. Six Sigma is a quality control program that seeks to reduce defects in manufactured products and services to a level of just 3.4 defects per million opportunities.
Benjamin Moses: Yep. That's solid.
Stephen LaMarca: That's it, dude.
Benjamin Moses: You got to get the math in there to answer that.
Stephen LaMarca: Why has it taken me nearly eight years to get to this? And it's not my fault. Don't call me dumb. Nobody has been able to distill Six Sigma to this small and acute of a level in layman's terms.
Benjamin Moses: Yeah. To be fair, we don't talk about it that often.
Stephen LaMarca: We don't, but we should still know about it. The first person that I know who tried to tackle this, the explanation of Six Sigma, was Jules. She tried her best. She did a great job.
Benjamin Moses: Okay. That was a while ago.
Stephen LaMarca: She still made no sense. It was a while ago.
Benjamin Moses: That is a fair point because the term Six Sigma does get tossed out a lot, and it's more of a journey that people try and achieve. So the different sectors approach it differently. So if you're producing a million parts, if you're in semicon and you're producing a huge volume parts or automotive, making sure you have three defects per million makes a lot of sense because you're going to get to a million parts. When you scale that down to low volume production, then you get to more acceptable range of trying to achieve Three Sigma where you have a certain number of defects and again, it's back to opportunities.
Stephen LaMarca: Oh yeah. Jesse and I were looking at my computer and I saw a chart of the different levels of Sigma. Also, I found out there's a seventh level too, which is crazy high quality, but anyway, I'm looking at these levels and Three Sigma like you just mentioned, perfectly fine.
Benjamin Moses: Right. Exactly.
Stephen LaMarca: Why do we need to go three more? It's because Motorola is Japanese, and they care about quality better than anybody else does.
Benjamin Moses: Yeah, and part of it is a journey to improvement. It's not necessarily always achieving that. It's identifying hey, this thing, we have more than what we want for waste. Let's try and solve it. Manufacturing is just a big journey, man.
Stephen LaMarca: Dude, this is more than a journey. This is some religious thing.
Benjamin Moses: Yeah, that's true.
Stephen LaMarca: This is some sort of religion. Six Sigma, what is this utopia these people are speaking of?
Benjamin Moses: Yeah, there's a lot of stats behind it and it is worth diving in the next layer within that to help understand the type of things that they're measuring, understand the standard deviations, and understanding the low limit, high limit, and then they're talking about just normal distributions too, which covers a wide variety of scenarios, but if it's not normal based, then you have to adjust for it. It's a big world. So I would definitely recommend talking to a quality manager about that because they can definitely spend hours on hours talking about that. So buy someone a beer, get ready to take a nap.
Stephen LaMarca: Yeah. Okay. I did misspeak. Motorola, I thought was a Japanese company. It's not.
Benjamin Moses: No. Back then, they were US based, right?
Stephen LaMarca: Yeah.
Benjamin Moses: Yeah.
Stephen LaMarca: Wow. What happened to defunct as of 2011?
Benjamin Moses: Oh. RIP Motorola.
Stephen LaMarca: A lot of good it did.
Benjamin Moses: Best phones ever though.
Stephen LaMarca: They were pretty durable. Not Nokia though.
Benjamin Moses: Mm, fair. They had some solid flips.
Stephen LaMarca: I'm confident Japan has Seven Sigma.
Benjamin Moses: Steve, we covered a lot of different things. Thanks.
Stephen LaMarca: We covered a lot.
Benjamin Moses: Where can they find more info about us?
Stephen LaMarca: amtonline.org/resources. See you there.
Benjamin Moses: Awesome. Bye everyone.
Stephen LaMarca: Bye.
