AMT Tech Trends: Tool But Verify

AMT Tech Trends · Tool But Verify

Episode 92: Ben tries his best to help Steve understand project management, then the two discuss ethics in AI. Benjamin contemplates how to improve industrial robot efficiency using large language models. Stephen is excited about next-gen aramid fibers that can conduct electricity. Ben explains augmented intelligence and its usefulness with QA/QC. Steve has been hearing about metamaterial concrete for ages now; is it more than vaporware?

Connect with the Manufacturing Industry here https://www.amtonline.org/events

Discover the past, present, and future of American manufacturing with https://www.mmsonline.com/madeintheusapodcast

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

Produced by Ramia Lloyd

Transcript

Benjamin Moses: Hello, everyone. Welcome to the AMT Tech Trends podcast where we discuss the latest manufacturing technology, research and news. Today's episode is sponsored by Modern Machine Shop's Made in the USA podcast. Warn them later. I am the Senior Director of Technology, Benjamin Moses, and I'm here with ...

Stephen LaMarca: Steven LaMarca, AMT's technology analyst. Almost forgot. Almost forgot.

Benjamin Moses: I stumbled a little bit because my notes are all screwed up here and I crossed the wrong thing out here.

Stephen LaMarca: That stumbling is contagious.

Benjamin Moses: Steve, we had an onsite training last week and one of the topics that we hit on was project management. I know you and I want to talk about-

Stephen LaMarca: First off, let's not call it onsite training. The title was the AMT's Cornerstone Week.

Benjamin Moses: Cornerstone Week. Yeah. That's correct.

Stephen LaMarca: And even though AMT has a lot of employees dispersed throughout the United States, all of them were flown to AMT's HQ, our office. And we basically had a week of, almost a week of team building activities, educational seminars, and overall it was just an awesome, great time.

Benjamin Moses: Good. I'm glad you enjoyed it.

Stephen LaMarca: I really liked it. One of those happened. One of those seminars happened to be on project management.

Benjamin Moses: Yep. We underrate the value of project management.

Stephen LaMarca: I would say I don't. I might even overrate the value of project management because I don't fully understand. After the project management seminar, I had probably not just one. I would say I probably had two or three hour long discussions with Moe on and offsite work about project management and my just blatant misunderstanding. I even had Chris and Kristen, two amazing project managers. The three of them, you and you, the four of you guys are AMT's rockstar project managers. And I had the three of them, Moe, Chris, Kristen, all the best project managers ever trying to explain to me. I wouldn't say I'm a total idiot, but trying to explain project management. For example, I was just not picking it up at all. For example, Kanban. I love me some ... Every time the guy said Kanban, our instructor said con bun-

Benjamin Moses: You giggled [inaudible 00:02:53].

Stephen LaMarca: I repeat. No, I repeated it.

Benjamin Moses: Okay, good.

Stephen LaMarca: So, "Hey we use Kanban." And I was like Kanban.

Benjamin Moses: Yeah.

Stephen LaMarca: I think I was starting to annoy the people around me, but I ... And I don't want you to go along into this because we only have, let's try to keep this episode to a half hour, but I can't tell a difference between Kanban and Scrum to save my life. If somebody held a gun to my head and was like, "What's the difference between Scrum and Kanban?" I'm dying that day.

Benjamin Moses: Let's use some simple language in the stuff we want to talk about today. Let's talk about inventory of tasks.

Stephen LaMarca: Yes. Okay.

Benjamin Moses: So that's kind of related to Kanban.

Stephen LaMarca: I like that. I like the sound of ... That makes sense.

Benjamin Moses: There we go.

Stephen LaMarca: You have a number of tasks, a stockpile, if you would.

Benjamin Moses: Yeah.

Stephen LaMarca: Which ones have to go to shipping and receiving and which ones are overhead?

Benjamin Moses: Which one? You got to pull from somewhere. So let's create an inventory of tasks. Now we were talking about-

Stephen LaMarca: IOT. Great. We just made it more ...

Benjamin Moses: Put more acronyms in this thing. So I think a lot of people that are managing their tasks. Everyone has a task list somewhere. That's your inventory. That's where you're pulling your Kanban from.

Stephen LaMarca: Yeah.

Benjamin Moses: Some people have sticky notes. Some people have a digital version. We've been using Workflowy for a long time.

Stephen LaMarca: I love Workflowy. But that's just a list. It's just a listing tool.

Benjamin Moses: Yeah.

Stephen LaMarca: It allows you to just write lists. You can do some macros with it.

Benjamin Moses: Right.

Stephen LaMarca: There are some macros. I'll be totally honest with you, I don't use the macros, because it as a listing tool is just so powerful. But go on.

Benjamin Moses: At home, I use sticky notes and a whiteboard.

Stephen LaMarca: Respect.

Benjamin Moses: So on the weekend, I have a list of things to do. I create a big list of inventory and I move things if I want to do them or if postpone them to other stuff.

Stephen LaMarca: Nice.

Benjamin Moses: So that's my kind of note about project management that ... The conversation we had last week was more of the culture and kind of topics, how we actually go about it. And then to be honest, a lot of companies when they use a Kanban system on their own factory floor, because you can do that for a work. Right? Does that matter? Managing workflow on the shop floor or just tasks in office.

You can use sticky notes to indicate you have inventory stuff and where it sits in the process. So I've seen a lot of factories just use a giant sticky note board for managing their workflow through their factory. So the tools that we were talking about, there are digital versions, and it is nice to have a lot of the digital tools so we can see where things are across the organization. But in the end, real life risk digital. What's the best tool do you need at that time?

Stephen LaMarca: Right. And I think we had a really great discussion about, because one of the things our instructor said was, "Sticky notes are better than any software." And it was like, so what software do you recommend? Somebody was asking what softwares do you recommend, which ones are the most successful to you? And it's like software offers some added flexibility that you can't get out of just a physical whiteboard and sticky notes. But at the end of the day, there's something inexplicable to the feel of writing down something on a sticky note and putting it physically adhesing it to a board and taking a step back and looking at. There's a physical experience to it.

Benjamin Moses: Agreed.

Stephen LaMarca: I meant to get his sources, if he has any on ... Because he made a bold statement saying that an actual whiteboard and sticky notes are more effective than any software that he's used or any software that was used in whatever source he has should he have one. And I would love to read just that article.

Benjamin Moses: That'd be fun.

Stephen LaMarca: And to be fair to Moe, going back a little bit. To be fair to Moe, Chris and Kristen, they all concluded because of my inability to comprehend project management, they all concluded you just like to get project management and to be good at it and to understand what goes into it, you just have to be thrown in. You just have to be baptized by fire.

Benjamin Moses: Right. Just do it.

Stephen LaMarca: And you can't do it with one project, because I was telling them about how I did the AMT's 2020 priority project and of the three priority projects, mine actually got done. But nobody would still call me a project manager because it was just one project.

Benjamin Moses: Sure, sure.

Stephen LaMarca: Being a project manager, it's like saying, I actually have no example. I have nothing to say here. But that's like saying you're a school teacher but you only teach your child and it's homeschooling. You're not a school teacher. You just teach your child. To manage a project or be a project manager, you can't just manage one project. You have to manage multiple projects. But anyway, going back to the digital versus physical sticky note. This got me thinking because I wanted to prove him right or wrong, and I think a lot of people in our audience did, because we're all smart cookies here except for the ones that can't get wrap their head around project management.

It was really insightful to the concept of digital twins. I agreed immediately with our teacher, our instructor, Alan was his name, with that's a good point. I play a lot of racing games and in those racing games, I have my own car and I've even spec'ed my car in the game to be almost exactly like my car in real life. And I got to say, the car in the game feels just like ... it handles and it's numbingly slow just like my car in real life. But I get to experience my car and feel the comparison between the real life and the digital version. Because I don't live in California or anywhere near it, I can take my car in the game to Laguna Seca and actually put some laps around the track.

Benjamin Moses: That's cool.

Stephen LaMarca: I've put laps around the Nuerburgring and I'm no way am ever going to have ... I'm going to have a Ferrari before I'm able to ship my car to Germany to drive it around the Nuerburgring. There's the benefits of a digital twin. But at the end of the day, there's also other cars. For example, a Ferrari. I'm never going to own a Ferrari, but the closest thing I can is driving it in a game.

Benjamin Moses: It's someone else's digital twin.

Stephen LaMarca: And it's a digital twin. I take it. I have confidence that somebody did the due diligence to ensure that the Ferrari in the game feels and handles as closely as possible, as digitally possible to how it feels in real life. And that's enough for me because I'd rather not pay for those $4,000 oil changes. 4,000 is probably light. It's probably more than that.

Benjamin Moses: It's probably more.

Stephen LaMarca: There are some other examples of a digital twin that I had in here. No, I didn't.

Benjamin Moses: No. Was that the only one?

Stephen LaMarca: That was the only one. It was comparing that to post-it notes versus a whiteboard. Now, the flaw with the whiteboard and just the flaw with the car is-

Benjamin Moses: The sticky notes fall off.

Stephen LaMarca: ... I can't take my car to the Nuerburgring.

Benjamin Moses: Correct.

Stephen LaMarca: I can't go to Laguna Seca. I only have these awful roads around Tysons to appreciate my car with. The digital twin lets me take it anywhere, but it's only digital. It's not the physical real life thing. The whiteboard and the sticky notes are only here in the office. If you come up, if you think of something in the middle of the night and you complete something or want to manage something, I guess. Are you going to get out of bed, get dressed at 2:00 AM to come in and move a sticky note from one area to the next or add another sticky note to it? No.

Benjamin Moses: No.

Stephen LaMarca: You're going to want a digital solution on your phone ready to roll.

Benjamin Moses: Right.

Stephen LaMarca: Anyway, we've gone, I knew he would go too long on this.

Benjamin Moses: I think the key is being able, having the tool transportable and communication. If you wanted to see how another project is doing, you just poke in there. If you know had the sticky notes, you'd have to walk to a room and look and figure out and read people's handwriting, which is .. Don't try to read mine. My handwriting's terrible.

Stephen LaMarca: I think the point is, why not both?

Benjamin Moses: Why not both? That's fair.

Stephen LaMarca: The [inaudible 00:12:19] would tell us when we're talking about digital twin this much, he'd be like, "You can't use the word digital twin unless you have a physical original."

Benjamin Moses: Yeah. That's fair. That's cool. I like that concept. So project management and the core of project management is a couple of layers. The main thing, it's about communication, staying organized, but it's a productivity tool.

Stephen LaMarca: It's a tool. It's a philosophical concept to be used as a tool.

Benjamin Moses: Exactly.

Stephen LaMarca: And people who can wrap their head around that philosophy and are effective with it get paid a lot of money.

Benjamin Moses: The next topic you want to talk about-

Stephen LaMarca: I have another tool to talk about.

Benjamin Moses: Yeah? Tell me more.

Stephen LaMarca: It's the tool of our generation, our lifetime. It's the next greatest tool going back to ... Let's think about the greatest tools of all time. Of all time. The first, a chimp picked up a stick and used it as a club probably to beat another chimp to death with, but I digress. Then a Neanderthal created fire, and then probably a more advanced Neanderthal created the wheel and then we became homo sapiens. And I know we've harnessed electricity and all this kind of stuff, but the new tool, our new stick is AI.

Benjamin Moses: Artificial intelligence.

Stephen LaMarca: ChatGPT, baby. You've heard it all over the place. We're going to keep talking about it, but there's a lot of fear right around it. There's a lot of concern. I can't wait to get into how the art community is handling it, rather losing their minds over it. Right now they are super butt hurt. But anyway, I digress. The point is just to reassure everybody and calm everybody down, it's a tool.

Benjamin Moses: Correct.

Stephen LaMarca: It's just a really good tool. It's a tool like a paintbrush but a paintbrush that also has Autodesk Fusion 360 in it. But it's also easy to use. It's that powerful.

Benjamin Moses: Go ahead.

Stephen LaMarca: No, you go on. It's a tool. I think they get it.

Benjamin Moses: And one thing that I think a lot of people are glossing over is people are using tool and posting their results wherever. The more we've been using AI in a lot of our content, the one thing I've always come back to is always check your work.

Stephen LaMarca: Yeah, absolutely.

Benjamin Moses: Before you even copy and paste out of the screen or hit the next button in your tool, incremental checks throughout the entire process is pivotal. And I think a lot of people are forgetting that. And actually I have an article that I'll talk about this a little bit later, is the human is pivotal in terms of you can trust but you got to verify.

Stephen LaMarca: Yeah. It's so important to project, going back to project management. Promise, I won't go back again. But it doesn't do the work for you. It still needs to be checked. It gets you really close. It fills gaps. It breaks down roadblocks, especially writer's block a lot, but it doesn't do everything for you. We've been using OpenAI's Playground for a while now, but that's more like comparing that, they both use the same engine, both ChatGPT and the Playground are the same engine, but the Playground is really like a command prompt. It's like the blank screen and you enter a code or a command-

Benjamin Moses: The good old days of DOS.

Stephen LaMarca: ... and hope that you get a result that you want. And if you don't get the result that you want, you keep tweaking the input until you get the output that you want. ChatGPT is like that but for dummies because ... and there's the problem, and we'll get to that. But ChatGPT is so powerful because it makes the programming of the AI's output conversational.

Benjamin Moses: Correct.

Stephen LaMarca: Not everybody can speak program code. Not everybody can ... And the computer can't, the Playground rather can't understand what the heck you mean half the time. Less than half the time. But the way to ensure that people are programming it properly is to fix the language barrier, fix the lost in translation. And that's why conversational is so important. I think what's special about so special about ChatGPT isn't so much the AI because like I said, we've been using it forever and it is special and magic. But what's even better than that is the conversational programming on top of that. And pre-pandemic, we were ... Tim was trying to get us hyped up about something, some technology that was being applied in a conversational way and I don't remember what it was.

Benjamin Moses: Who knows?

Stephen LaMarca: But man, ChatGPT really stole it from them, whoever that was.

Benjamin Moses: And to be fair, there are more competing conversational tools now. Google released one. I'm sure Microsoft is right behind.

Stephen LaMarca: Oh my God. That's scary. That's the scary part. You have every reason to be scared about that stuff. When companies only see dollar signs and they look at something as influential and disruptive as ChatGPT with only dollar signs in their eyes, that's when you're going to have bad things happen. And that kind of scares me, but ChatGPT isn't the one.

Benjamin Moses: The capitalism to push out some of these tools. And to your point, I think that does alarm me a little bit because the need to get to market as soon as possible. I think they're to your point of the underlying trust and ethics related to artificial intelligence, basically the training tools, the training models they used to get their end state, that's the biggest concern I have, is what corners were cut or what database was used that's maybe an extension as opposed to building their own tools or their own learning set.

Stephen LaMarca: I had a lot of my faith restored, not that it was ever gone, but I listened to a little bit of the Lex Friedman podcast where Lex was interviewing, really having a conversation with ... Lex Friedman is an MIT AI researcher, was having a conversation with the CEO of OpenAI and the OpenAI CEO was the ... Picture all of big techs CEOs, and now picture the complete opposite of them. Not super humble and just really down to earth-

Benjamin Moses: That's cool.

Stephen LaMarca: ... without the facade of trying to be down to earth. And he was just super humble about the product and was like, "Listen, our AI is still very juvenile. It's infantile. What we're seeing now is incredible, sure. But it's still in the early stages and it's got a long way to go, and that's why we can't fathom charging people too much for it yet." He said that in so many words, but it was just really reassuring hearing him talk like that.

But getting to the ethics about it, I personally find that hilarious because you and I know that it's a tool. It doesn't do the work for you. You need to do have input. But the art community, I think it was last year that an AI artist won several awards, if not all of the awards for modern art or whatever. And the art community was outraged because it was an AI. It wasn't a person. It was an AI that created these masterpieces that were so good, they beat out all the human competition. And the art community naturally was outraged, but not because of this new paintbrush, which is effectively what it is that made this incredible artwork.

But there was a lot of concern about, well it was very American, the argument because it was like, "How are we going to sue to get money out of this?" Because they were like ... The AI tool scoured Google images, looking at photos that belong to other people of things and looking at other art pieces for inspiration. And it's like, wait a minute. Then that means every artist after the internet age needs to be sued for royalties because if somebody ... if you painted something nobody would come after you and be like, well, so if you painted the Washington Monument, the National Park Service wouldn't come after you and be like, "Hey, we own that. You need to pay us royalties because this obviously inspired you." I think I used a better example yesterday.

Benjamin Moses: Used Niagara Falls.

Stephen LaMarca: I was like, if you painted or drew and got a lot of money for drawing an obelisk, the first thing you'd picture, because this is how our minds work and this is how artificial intelligence works, is an obelisk. What the hell is an obelisk? Oh like the Washington Monument. Then you need to be sued because the Washington Monument doesn't belong to you. Whoever does own it, the government, you're going to give them taxes anyway for any money you make. But they need more because that's their piece that inspired you. No, that doesn't work like that. Muses are not paid, and they don't get royalties. Anyway.

Benjamin Moses: Is our community just angry just to be angry? Is that the thing? I feel like that's a cultural thing for our community. They like to be angry.

Stephen LaMarca: Yeah. They get really worked up at politics, which I mean who doesn't? But then they pour their life into creating art, and they know darn well that nobody goes into art and is successful while they're alive. You don't start making money until you're dead.

Benjamin Moses: Oh, that's a bummer. I'll take that off my list of things to do then.

Stephen LaMarca: I think that's why they're so worked up.

Benjamin Moses: I think so.

Stephen LaMarca: Dude, you sleep in the bed you make.

Benjamin Moses: I also think they got a bad beat from NFTs.

Stephen LaMarca: No. I don't think NFTs affected them at all.

Benjamin Moses: No? Okay.

Stephen LaMarca: NFTs are so dumb. They're not dumb. They just have nothing to do with art. NFTs, if anything, will help art I think and artists-

Benjamin Moses: Maybe. I doubt it.

Stephen LaMarca: ... in protection of IP. It's like blockchain.

Benjamin Moses: Correct.

Stephen LaMarca: NFTs and blockchain are very special tools.

Benjamin Moses: Application for art.

Stephen LaMarca: The original application that brought them into the public eye is awful, and crypto bros ruined the blockchain.

Benjamin Moses: I'm glad they're gone from video games.

Stephen LaMarca: Unfortunately, Crypto bros are not gone, because the crypto bros die out in one of two ways.

Benjamin Moses: Tell me.

Stephen LaMarca: They either go broke and they become the bum on the street, or they become successful enough to become an investment bro.

Benjamin Moses: Oh, no.

Stephen LaMarca: And they're just as bad, if not, even worse because they're not as clout chasing. They hide in the shadows and next thing you know, they have all the money.

Benjamin Moses: Steve, let's move on to our sponsor.

Stephen LaMarca: Yes. Modern Machine Shop Made in the USA podcast. Tune in for Modern Machine Shop's Made in the USA podcast to explore manufacturing issues faced by companies making an intentional choice to manufacture in the U.S. Featuring commentary from OEM leaders, Made in the USA blends its nearly centrally century long expertise with a unique audio storytelling experience to shine a spotlight on the past, present, and future of American manufacturing. Find Made in the USA on Apple Podcasts, Spotify and all major podcast platforms. Follow Modern Machine Shop on Twitter, Facebook and LinkedIn.

Benjamin Moses: Thanks, Steve.

Stephen LaMarca: Somebody I'll watch on YouTube because I like nice shoes and boots, and I've got a pair of made in the USA Wolverine 1,000 Mile Cap-Toe boots. The Wolverine boots are made in Michigan.

Benjamin Moses: Nice.

Stephen LaMarca: And the leather that they use is Horween Chromexcel in number eight because that's the only color you should be buying. Number eight, which is like a burgundy. It's like a purpleish brown. It's really awesome. That leather comes from Chicago and from Horween in Chicago. That's where the tannery is. And I think they source the hides from Texas. Anyway, super made in USA. I love them. But anyway, you got to take care of them because at the end of the day, it's organic matter.

Benjamin Moses: Right.

Stephen LaMarca: It will decay and rot and fall apart if you don't take care of them. So I watched some YouTube videos way more than I probably should have recently on how to take care of them. And one of the videos that I watched was from a channel called Rose Anvil, and the guy has an entire series comparing and ranking the best made in the USA shoes and boots.

Benjamin Moses: Oh cool.

Stephen LaMarca: And he calls it the Madeusa series mat for made in the USA.

Benjamin Moses: Yeah. That's pretty good. I like that.

Stephen LaMarca: Went on all of that to tell you that dumb nugget of information anyway.

Benjamin Moses: That's a good punchline. So I've got two articles. They are artificial intelligence related, but one is related to automation and the other one's related to in metrology. So it's interesting how they both kind of dive into or the need for artificial. The first one I'll get into is actually from Forbes. Can large language models enhance efficiency in industrial robots?

Stephen LaMarca: Forbes is talking about this?

Benjamin Moses: Forbes is talking about it.

Stephen LaMarca: Who's the author?

Benjamin Moses: I don't know.

Stephen LaMarca: Is that Jim?

Benjamin Moses: There might be a Jim involved.

Stephen LaMarca: Okay. Respect if it's Jim because he knows what he is talking about. Other Forbes ... Other Forbes contributors, we'll see, but go on.

Benjamin Moses: We'll check those work. So the interesting thing that he talk about is we've come in the to know in the past few years how important automation is for the growth in the manufacturing industry, particularly in the U.S. as there's a big initiative obviously to reshore back into the U.S. or reshore locally, domestically depending on your location. And automation is one of the enabler technologies to help us scale up but also be price competitive to compete in the future against other global markets. And they talk about tools that enable better human to machine interfaces, better HMIs. And robotics is one of those areas where they're thinking about where tools like ChatGPT or other tools related to conversational commands to help implement robotics.

Stephen LaMarca: Nice. You know it.

Benjamin Moses: And I thought it was a very interesting article about the tools that we have available to accelerate the adoption of other technologies. So using conversational tools as leverage for improving your physical manufacturing line. And it goes to a couple of scenarios. One is speech to robot programs. So issue a command similar to what, Siri, Alexa. Yeah, that's the one.

Stephen LaMarca: This is the article? Okay.

Benjamin Moses: Maybe. No, that's not the one.

Stephen LaMarca: Oh darn. Sorry. I threw off the beat.

Benjamin Moses: So issuing verbal commands, again getting to programs. That's been around. The adaptation into the manufacturing sector is a little different because of noise and the clarity, and there's one of the limitations we'll get to later on. But being able to text to program or speech to program, that's fairly-

Stephen LaMarca: That's incredible.

Benjamin Moses: That's fairly interesting. The other one is line by line program explanations. So telling you what the programs are doing. One interesting thing that they talked about in this one is the learning curve of an operator or person on the floor for robotics. So if I'm on the floor and I want to see what someone else did, going line by line, reading other people's code is terrible. And if you're in software dev, you have to do that. It's a nightmare reading someone else's code. People would rather just scrap someone else's code and rewrite it from scratch than trying to figure out what they did. So being able to be told what a block of code does is super efficient and reduces a lot of the learning curve for a lot of people.

The other one is adaptation. So automation's very robust depending on scale. The bigger the volume, the easier it is to justify and implement. But if you have low volume, high mix, how do you change from one to another? And they're looking at different tools using conversational language to say, okay, instead of this part or this series, switch over to another part or another command. That's another benefit, is being able to modify your existing programs very, very rapidly. So those are three scenarios were beneficial, but there are limitations to this. And we did briefly touch on this earlier is yes, it can get you a result. The question is, is that result accurate?

So the article talks about putting in buffers before it physically moves an object. So have it create a code then we've got to manually verify, is that code correct? So the limitation is everything has to be manual checked, which is 100% true. To be fair, it's across everything, but it's not too ... I think it's unfair to say the tool is super robust right now to get you to moving an arm right away. And I think that's where the limitations of, yeah, it can accelerate you but someone has to come back and verify. Someone has to come back and say, "Yes, this won't crash. Yes. This did not miss something that the program didn't see or things like that."

Stephen LaMarca: That's incredible. And it goes back to what you said earlier, trust but verify.

Benjamin Moses: Yep.

Stephen LaMarca: Oh man. I think that is so wild, because the conversational ... If when you take the conversational aspect of ChatGPT or something like ChatGPT and pair that with voice recognition software that both Google and Apple have. And then OpenAI's ability to, their AI's ability to take what you want and be able to write code to it and be able to send that to a robot arm, even though it's not ... the computer might not be directly connected to said robot arm assuming they're both on the internet of things, and that robot arm receives the program code. You could do something like a machinist, or a machine tool technician could be working on a machine and drop their wrench, but they're trying to hold something in place and be like, "Hey can you grab that? Can you grab that with the FANUC arm?"

Assuming it has a gripper, and there'd be some processing lag. Can you imagine? In the future that the problem is going to be the processing lag of somebody saying something to a computer. The computer has to process what they need, simulate it and then send the final verified program, assuming that the computer was able to verify itself or send it to somebody who can verify it for them and then send that program to the robot, and the robot grabs the wrench and hands it back to the technician.

Benjamin Moses: I have a task for you.

Stephen LaMarca: Yeah?

Benjamin Moses: At the robotic block party.

Stephen LaMarca: Don't ask me to get the pocket or the test bed to do this.

Benjamin Moses: No.

Stephen LaMarca: I can't.

Benjamin Moses: At the block party in San San Francisco.

Stephen LaMarca: Yeah?

Benjamin Moses: Can you ask about this?

Stephen LaMarca: Absolutely.

Benjamin Moses: I think it's a worthwhile conversation.

Stephen LaMarca: This is incredible.

Benjamin Moses: Where they are in their development cycle to see, are they thinking about this line path?

Stephen LaMarca: I think they definitely are.

Benjamin Moses: Of course.

Stephen LaMarca: But I'll keep asking it because from when my visit to FANUC, FANUC has done a lot to make programming robots as easy as possible with touch screens and icons instead of code. You can take the icon of the commander, the thing that you want the robot to do, and drag it to a line of sequences. So you can program your robot by a sequence of pictures of things. And while that's really good as an educational tool, all of the people who are in robotics are like, "This is great for education but nobody wants to program a robot like this."

Benjamin Moses: Fair.

Stephen LaMarca: Nobody wants to work on a robot like this. Yeah. It's a great question I will be asking them.

Benjamin Moses: I think the key takeaway is a variety of tool gets you there. No one tool ever solves one problem.

Stephen LaMarca: Right. That's why it's called the toolbox. It's supposed to hold multiple tools.

Benjamin Moses: Steve, tell me about materials and fiber.

Stephen LaMarca: All right. Do you like body armor?

Benjamin Moses: Yeah.

Stephen LaMarca: I think body armor's pretty cool, right?

Benjamin Moses: Yeah, sure.

Stephen LaMarca: It's pretty tight. Kevlar.

Benjamin Moses: John Wick-

Stephen LaMarca: Aramid fiber.

Benjamin Moses: John Wick has a cool one.

Stephen LaMarca: Yeah. I mean, assuming you don't have to use it, but it's a cool thing.

Benjamin Moses: Somebody does.

Stephen LaMarca: Somebody's got to use it. Hopefully the people on your side. Do you like carbon fiber?

Benjamin Moses: That's the best.

Stephen LaMarca: Carbon fiber's pretty tight too. Right? So the Korea Institute of Science and Technology has developed a new composite fiber made of aramid and carbon nanotubes. So they've apparently like the aramid fiber called golden silk or super fiber blended with carbon nanotubes. The aramid fiber on its own is called golden silk or super fiber. And then you mix it with carbon nanotubes or lace it with carbon nanotubes, if you will. And they call the new composite black fiber.

Benjamin Moses: Wow.

Stephen LaMarca: Sounds super sick. But basically it's a lighter weight, stronger composite. Could be used for body armor. Could be used for car hoods, but you can also send a electrical current through it.

Benjamin Moses: Oh, fascinating.

Stephen LaMarca: This new composite fabric, if you would, or fiber, doesn't have necessarily have to be fabric. It's going to have so many, it's going to unlock so many different applications. So you could integrate sensors into an automobile hood for lighting purposes, because we've all been to a car show and lights everywhere, are the thing. Cars and gaming PCs are very similar, and many RGB lights as you can stuff in it, the better. But also there's the defense industry application.

All soldiers want body armor, number one. And they all want it to be lighter because everything that weighs them down weighs them down. So lighter's always better. But the fact that you can run intellectual current through it and you could theoretically make your body armor also a PCB or circuit board, you could have something. You could have sensors or you could have your body armor quite literally a wearable. And so if a soldier takes a hit and the sensors theoretically might be able to tell that projectile penetrated. Seek medical attention.

Benjamin Moses: That's good. No, I didn't think about the sensor, the wearable sensor side of it. I was thinking about for the long time, actually forever they're coupling of the battery or the energy at point of use. So a walkie-talkie, right? You pick up a walkie-talkie, a big portion of that walkie-talkie is the battery itself. So if you're able to segregate those two, so the battery is somewhere else, somewhere more centrally located or somewhere that's easier or consolidated to a battery pack across several devices. That's what I was thinking of is being able to move the energy storage and point of use equipment, so it's easier to use going forward.

Stephen LaMarca: I don't like that idea.

Benjamin Moses: All right.

Stephen LaMarca: And I'll tell you only because batteries store energy and energy always finds a way to release in a manner that you don't want it to or that it wasn't intended.

Benjamin Moses: That's fair.

Stephen LaMarca: When Tesla was having problems with their cars catching fire, it was like, oh, batteries are so dangerous, we got to stick with gas. What do you think gas is? Gas cars never caught fire and exploded before? You ever heard of the Pinto? Any kind of energy storage can go horribly wrong.

Benjamin Moses: Correct.

Stephen LaMarca: So I would never want to put energy storage with body armor, but maybe for something else.

Benjamin Moses: So use the body armor for ... or that material for transportation of the energy.

Stephen LaMarca: Yes. Nobody wants cables.

Benjamin Moses: Right, exactly.

Stephen LaMarca: In a world where we're trying to go wireless and wireless electricity transfer doesn't exist yet. Sounds great. That sounds great. Energy transfer, you can still have issues, but not as bad as energy storage.

Benjamin Moses: Yeah. Agreed. We're on the same page, Steve.

Stephen LaMarca: Yeah.

Benjamin Moses: Can I go over?

Stephen LaMarca: Yeah, your turn. It's your turn.

Benjamin Moses: The role of augmented intelligence in quality inspection. So we've been talking about artificial intelligence. This article talks about augmented intelligence, and I think the reason they bring up this subtle difference in the term is the tools that we're using now is making a lot of the decision for us. The artificial intelligence, when you give it a command, it's making a lot of decisions to get to an end state. The role of augment intelligence is identifying or using humans as the final decision making person or final decision maker in the process. So then in the scenario, this is from Metrology News.

Stephen LaMarca: Nice.

Benjamin Moses: They go over using obviously artificial intelligence for vision inspection or different inspection techniques. Say they're inspecting a mirror or a glass. Instead of using AI to say here's a defect and the here's a area of concern and a defect, they're using the tool to say, here's a area of concern. Operator, can you check this? Or technician, can you check this? And then they would be the final decision maker. So instead of having the operator check the entire section, having AI say this of the whole 300 square foot area, two square inches of it, you should check. Go check this area. And then the decision's made by the technician or operator. So I think it's an interesting subtlety, and the thought process behind it is humans are still very good on the manufacturing floor. We're talking about some advanced technologies, but as we're talking about-

Stephen LaMarca: I like this a lot.

Benjamin Moses: Yeah.

Stephen LaMarca: I'm going to cut you off and say this article is, it's almost like an academic paper even though it's just an article.

Benjamin Moses: Sure.

Stephen LaMarca: It's just proving our entire episode right. It's supporting us.

Benjamin Moses: It is.

Stephen LaMarca: Because AI isn't everything.

Benjamin Moses: Correct.

Stephen LaMarca: It doesn't do all of the work for you. It gets you a long ways away. It does all of the nasty parts that nobody wants to do. And then you get to clean it up, make it look pretty at the end. Yeah. Your example or when you were talking about this, I kind of thought of how spell check original spell check on Microsoft Word, it's not like autocorrect, which as soon as you hit space away from that word, autocorrect just corrects the word. Oh, I never spelled a word wrong. All my spelling's perfect. I never see red lines. But that came from the augmented spelling, spell correction, which is redline. The computer does not recognize this word. Is this a real word? And if it is a real word, if you right click it, you can see all of the words that we think it might be or you get really depressed if it's nowhere near the word that you wanted to, that you meant, and now you're Googling.

Benjamin Moses: I've gotten in less trouble spell-check, being spell-checked than autocorrect.

Stephen LaMarca: Yes.

Benjamin Moses: So I think that's a parallel example everyone can relate to.

Stephen LaMarca: Yes. This is where augmented is superior. Wow.

Benjamin Moses: Congratulations Steve. Appreciate it.

Stephen LaMarca: This is amazing.

Benjamin Moses: Steve, I think we should get to your article so we can let people get on with their day.

Stephen LaMarca: Yeah. Okay. My last one. Again materials. Researchers create self sensing metamaterial concrete that produces power. So I remember all the way back when I was in college, a long time ago. People were really excited about some European country, maybe Germany, maybe Sweden or Switzerland, was experimenting with power in the roadways. As a car is driving down the road, either the road absorbs the kinetic energy, the downward kinetic energy from the weight being transferred onto that material of road, to either power something or at least it being used as a sensor to be like, there's a car coming. Street lamps need to turn on this far in advance so the road is lit for the car. But also it saves power because you don't need to leave the entire road of street lamps on if nobody's driving on it.

But there was also a futuristic aspect to electric cars can be charged by as they're driving down the road through conductor induction coming from the road. This is getting a little bit further into the future.

Benjamin Moses: That's cool.

Stephen LaMarca: About talking about the metamaterial behind it. I still think it's really far away, but it's being done by the University of Pittsburgh. And you and I love Pittsburgh. Pittsburgh is the most cost-effective city in the United States, if not, the world. We've always had a good time when we've gone for work events. It's a really, I just love ... I'm nerding out about Pittsburgh. It's really nice. You get all of the glamour and the fanciness of other major cities, but it doesn't cost anything to have a good time there. It's also small enough where you can walk and see the whole place. And then in terms of cost-effectiveness or to add on to cost-effectiveness, one of the major brands at Harbor Freight is Pittsburgh.

Benjamin Moses: Pittsburgh, my favorite brand. Not Wera, not Snap-on. Pittsburgh.

Stephen LaMarca: You don't need a Wera.

Benjamin Moses: No. Sometimes you don't. Steve, we covered a lot today. I was riding down a couple of things. So artificial intelligence was our theme throughout, but we hit on robotics, metrology, material sciences, energy. Also key takeaway, love your project manager.