Episode 102: Ben and Steve share stories and takeaways from recent travel, including Oak Ridge National Lab, FABTECH, and the Outer Banks. Benjamin shares NVIDIA’s take on the impact of generative AI in robotics. Stephen is peeved by good technology that’s been given a snake oil name. Ben closes with collaborative robot fiber placement.
Articles Featured in This Episode:
Looking for More?
Watch this episode on YouTube
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
Ramia Lloyd:
Welcome to the TechTrends podcast where we discuss the latest manufacturing technology research and news. Today's episode is sponsored by the MT Forecast Conference. I am Ramia Lloyd, producer and I'm here with...
Stephen LaMarca:
Stephen LaMarca, AMT's technology analyst.
Benjamin Moses:
And I am Benjamin Moses, director of technology. Steve.
Stephen LaMarca:
Senior director of technology. Don't try to hide that you're getting old.
Benjamin Moses:
I was just thinking about that. The selfies I take now, I have to worry about my hairline.
Stephen LaMarca:
I love your selfies. I want to take selfies like you.
Benjamin Moses:
I try.
Ramia Lloyd:
They're my favorite thing.
Benjamin Moses:
They're intentful.
Stephen LaMarca:
But it's the fact that you're so methodical about every time you go on a trip, it's like you're documenting it for work reasons. I am actually at this event and on time. Here's my no smile mugshot in front of said event. I love them. It's like we should all really be doing them.
Benjamin Moses:
It's my verification of my expense report.
Stephen LaMarca:
It's like, is this a major industry event? AMT was here.
Benjamin Moses:
Exactly.
Stephen LaMarca:
If you didn't see us, that's your fault. We're here on LinkedIn.
Ramia Lloyd:
Ben's our favorite content creator.
Stephen LaMarca:
One of the most reliable and consistent for sure.
Ramia Lloyd:
True.
Benjamin Moses:
This comes from years of training. Steve, let's talk about our travel.
Stephen LaMarca:
Yeah. We've gone on-
Benjamin Moses:
I heard you were in Tennessee.
Stephen LaMarca:
Was on Tennessee. Before that. So before going to Oak Ridge, which was last week, the week before that, for nine days straight, I was in the Outer Banks with Melissa and my best friend and just a vacation.
Benjamin Moses:
Nice.
Stephen LaMarca:
Nice, solid excuse to be like, oh, we left the dog with my mom. Zero responsibilities. It was just probably seven, realistically, seven days straight of just being inebriated and sandy.
Benjamin Moses:
A good vacation.
Stephen LaMarca:
A proper vacation.
Benjamin Moses:
Exactly.
Stephen LaMarca:
And at the end of it we were just like, oh my God. Well, hold on. Before I talk about the end of it, let me talk about, I did take some notes. When you go on vacation and you want breakfast, because let's be honest, most of the time you're not going to get a breakfast on vacation. The reason is only psychopaths and serial killers wake up early on vacation for breakfast.
Benjamin Moses:
Or nine year olds.
Stephen LaMarca:
So if you have kids, sorry. But the day we left, we were like, let's get breakfast today, because we got up early to leave. Let's have one last nice moment. And I googled where a nice breakfast spot was in the area. Every place was packed because when you go in and out of the Outer Banks, the rental properties are all scheduled the same way. So it's a mass exodus and then it's a mass influx of people coming in to fill the houses.
Benjamin Moses:
Like the movie Jaws.
Stephen LaMarca:
I guess. I've never seen the movie. I'm going to have to watch it now. I know. Shame on me.
Benjamin Moses:
It's fine.
Stephen LaMarca:
So mass exodus, but we're like we're not going to be one of these nerds in traffic. Let's go get some breakfast. Well, 50% of the people also thought that way. So there was a good breakfast joint that we were thinking about going to. And then Melissa was like, let's not spend too much money. We've spent a lot on this vacation. And I'm like, all right. So I found a cheap place with good rating, low dollar sign, high stars, and they had no view and the breakfast was mid at best.
Benjamin Moses:
Oh man.
Stephen LaMarca:
Not even good quality diner food where they know the food's not good, but at least they tweak it by putting a stick of butter in it. Not even that. And I just thought to myself when in doubt, when you're struggling to find a breakfast place and you're up for breakfast, just go to McDonald's. It's consistent, it's reliable. It is always, it doesn't matter if you're in Tahiti or in the Outer Banks, it's always the same. [inaudible 00:04:17] McDonald's is always good about, it's not like Bojangles where one in five locations is going to have awesome food and four of them are going to be utter trash and they're just going to throw the food at you. All McDonald's are always the same.
Benjamin Moses:
Always the same.
Stephen LaMarca:
Always. Should have done that.
Benjamin Moses:
To add that, I recommend a Cracker Barrel breakfast.
Stephen LaMarca:
That's fire too. [inaudible 00:04:38] You're right. Cracker Barrel is like family. I think my aunt and uncle were so prepped. So they have the truffle farm out in, I can't remember, deep woods of Virginia. They have a truffle farm out there and they've got a really nice farmhouse and I think it's in Berryville. No, it's not Berryville, it's Purcellville. They were so excited. They've got all this nice stuff out there. It's very nice farmhouse appointed with nice luxury goods and whatnot. But the one thing that they were the most proud of, they bought two Cracker Barrel rocking chairs for the front deck.
Benjamin Moses:
Wow.
Stephen LaMarca:
And they're nice. They're a nice place to pass out. Okay.
Benjamin Moses:
Let's talk about Oak Ridge.
Stephen LaMarca:
No, hold on. Hold on.
Benjamin Moses:
One more.
Stephen LaMarca:
I got three more things on the Outer Banks. Number one, we survived a tropical storm.
Benjamin Moses:
Oh yeah, that's right.
Stephen LaMarca:
Tropical storm Idalia came through there. It was terrifying. Me and my best friend, Chuck's, his mother, Geralyn, we were up all night. We were just watching the storm. We were watching the water and the sand hit the windows on the fourth floor of this rental house we're in. We're just like [inaudible 00:05:52] and this was after Melissa and I rented scooters and signed all the documents saying that if any damage happens to it, you're covering it.
Benjamin Moses:
These things are getting beat outside.
Stephen LaMarca:
Once the storm, it was still going on, but once it was not deadly out there, I went outside to pick up. Mel's fell over because she had a littler one and I picked it up. Dummy. I should have moved it into-
Benjamin Moses:
Covered area.
Stephen LaMarca:
A protected area. Didn't think to do that. Nope. I just put it back on the kickstand that it had fallen down from in the first place. So it fell down one or two more times after that. Brake lever breaks off.
Benjamin Moses:
Oh no.
Stephen LaMarca:
Fortunately the place was cool, but that segues into talking about Melissa and I rented scooters, these Wild Hogs, these absolute, just the epic example of what motorcycling perfection looks like. Not. These were like these Chinese Wolf scooters and we're like, I don't know what we're getting into, but I don't know if I trust my life with this and I'll do three digits on the highway on a Kawasaki. But this thing, I don't know where it came from. I've never heard of Wolf before. And plus I'd never been on a 50 CC scooter before. I appreciate slow vehicles. The slower the car or motorcycle, the more you can put it, ride it around, drive it around at 10 tenths, feel like Fernando Alonzo all the time. 50 CCs. And me being a couple 10 pounds shy of 300, 50 CCs is not enough. I finally found out what is too slow.
Benjamin Moses:
50 CCs.
Stephen LaMarca:
Too slow does exist and 50 CCs is like a weed whacker trying to move me.
Benjamin Moses:
That's fun.
Stephen LaMarca:
And then my last thing, have you ever been to a rental property or an Airbnb or a hotel room that has a smart TV?
Benjamin Moses:
Sure.
Stephen LaMarca:
This place had multiple smart TVs and one of the apps on the smart TVs. Netflix of course. And the people before you stayed logged in. All right, here's where I differ from everybody else. I feel like everybody else would be like, just use the account. They're already logged in. Who cares? Me being, I feel like a nice person. I'm like, no, I don't want to screw up their algorithm because I don't want to start watching things that are vastly different from an 11 year old princess' Barbie movies and stuff, and I'm going to create a new account or a new profile on this person's account.
Benjamin Moses:
That's nice.
Stephen LaMarca:
And just so they know that it doesn't belong to this family I named the profile Mr. Stupidy Head. And so we had one nice night of watching shows and movies and then the next morning-
Benjamin Moses:
They booted you out.
Stephen LaMarca:
You've been hogged out, log back in. It was like, well, at least they know now.
Benjamin Moses:
That's good. I appreciate that.
Stephen LaMarca:
I feel good about that. I feel like I did something good.
Benjamin Moses:
We have a ton of profiles at home. So I've got an account just for the home stuff. We watch a lot of YouTube, I watch a lot of YouTube and so does Amelia, but an account just for that. And then we have our own separate accounts that we cast to the TV. To your point is I don't want the same history or recommendations that Amelia or [inaudible 00:09:19] get on their channels.
Stephen LaMarca:
And I feel like there's also a lot of people who probably have a Netflix account with a profile that says guest that's just inviting people to use your account when you go to a rental property. [inaudible 00:09:34] Look out for Mr. Stupidy Head. You know I've been there.
Benjamin Moses:
Worldwide Stupidy Head.
Stephen LaMarca:
All right, the end of the vacation, we leave the entire six hour drive back. It's normally four hour drive, but it was a mass exodus, so added two hours to the drive. Traffic honestly wasn't that bad other than everybody else trying to leave the Outer Banks. Melissa and I are just talking on the way back home and we're just like, I don't know if we can go back to work. This was too much fun. We didn't have to walk the dog ever. As much as we love him and we miss him and we want to go back to him, there was no, it felt so nice to have zero responsibility. How are we going to do this? Fortunately our boy Tom Feldhausen at Oak Ridge National Lab was like, "Hey Steve, can you MC this event?" Never MC'd anything before. Can you MC this event at Oak Ridge National Lab? Never been to Oak Ridge National Labs before and it was the best transition to come back to the office.
Benjamin Moses:
Nice.
Stephen LaMarca:
Because I was like, I'm not going to go back to the office. I can't do it. I can't get back to work. I'm not thinking about manufacturing at all right now. And it was just a nice... Oak Ridge National Lab. I think on my post on LinkedIn, Stephanie Hendrickson of Gardner Business Media and Additive Manufacturing Media, she commented on the post, it's like Oak Ridge National Labs is the Disneyland of all things additive and manufacturing in general. And she's exactly right. It is Epcot Center for sure, but Oak Ridge was a real blast. I've got some more notes on there. I want to make sure I cover everything. So I was there for a hybrid manufacturing systems workshop specifically to focus on what are some of the bottlenecks holding back hybrid manufacturing in the production manufacturing industry today.
So a lot of the people there were people that Tom Feldhausen and the gentleman at Oak Ridge National Lab who put it on, Blaine Fillingham, both PhDs, by the way, everybody there's like a PhD, at least almost everybody there is a PhD. And they put on this incredible event. But all the people were there were people that Tom knew on a first name basis that are using hybrid manufacturing technology. And it was just a workshop, like I said, to target the pain points. Well, to identify the pain points before Oak Ridge can target them and help industry leaders figure out what is stopping them. So Okuma and Mazack were there to find out the two biggest hybrid manufacturing machine producers that make dedicated hybrid manufacturing machines. Of course a gentleman from Meltio was there.
I think there may have been a team of people from Meltio and Jason Jones from Hybrid Manufacturing Technology. Made sure to get pictures of their machines, of which Tom gave me a little bit of hell for because he was like, you took a picture of a Haas with a Hybrid Manufacturing Technology's ambit in the spindle. I was like, that's an awesome way to retrofit a CNC machine to do 3D printing. It's the perfect example of hybrid. And then you also took a picture of one of our tour mocks with a Meltio three additive tool in it. And I was like, well yeah, your dumpiest machines do the most impressive work. If you look at $1 million Mazack or Okuma hybrid machine and it does something, you're going to be like, ooh, ah.
Benjamin Moses:
It did its job.
Stephen LaMarca:
It did what it's supposed to do. But then when you take something like the lowest budget CNC machines and they're doing hybrid manufacturing, hybrid metal additive, it's like that's cool. That's why I took pictures of that.
Benjamin Moses:
I do like the idea of retrofitting equipment that's past end of life to continue the life. I do agree that okay if-
Stephen LaMarca:
You want dedicated at some point, but it's like one of the first product demos that I ever saw that really blew my mind was Matt Block of Royal Products. I think he's still at Royal Products did a demonstration of a bar puller and I was like, what is a bar puller? It's such a stupid sounding thing. It's just a tool that goes in a CNC lathe and it's a gripper to pull the bar stock through the chuck of the main spindle to expose more material to cutaway. And he's like, it's nothing special.
But you buy a few of these for your machines, one per machine and these are $200 compare that to a couple tens of thousands of dollars or $100,000 for a bar feeder, like a high-end bar feeder and you can put off that expense, make a little bit more profit so you can then afford the bar puller. It's just a transitional piece of technology. And that's exactly what these hybrid retrofitting tools are. These additive like the ambit and Meltio. And that's why I think the industry, the actual production industry hasn't fully adopted additive yet. And I've got a reason why.
Benjamin Moses:
Before we move on, I do like that term transitional technology. I do like the idea of, so one of the things, the problems that we see in a lot of the road mappings event that we participate in, the challenge is always understanding the risk and understanding return on investment. So being able to use these transitional technologies to say, okay, let's invest $5,000, let's say and test something out. That way we can figure out, do we want to invest in a $50,000 machine later. That path and similar to the test bed that we have, being able to do these small investments say this thing could be worth it and let's quantify that worth through these transitional technologies. I think that's a good approach.
Stephen LaMarca:
I don't think Jason Jones' ambit is as inexpensive as a bar puller, but it's inexpensive enough that a smaller medium-sized job shop can acquire one. If they don't like it, they can always be like, we didn't like it. Jason Jones probably wouldn't comp them, but it would be like, we'll buy it back for you at a reduced price because we need to refurbish it and then it will be sold to another person. But I'm sure there's a deal that can be done through there and that's the item that you want to buy before you go to an Okuma or a Mazack hybrid machine.
One would think, I don't know, I don't run one of these shops, but that's what I would think because it's a major investment. But it was really refreshing going to this event and seeing how much hybrid machines are actually being used. A couple takeaways that I got from the event is the major machine tool companies, again, Okuma, Mazack, Haas probably, Tormach. Tormach is all over the research area.
Benjamin Moses:
Okay, that's good.
Stephen LaMarca:
You can't miss them at NIST. They're at Oak Ridge National Labs, they're at Georgia Tech. They're everywhere anybody is doing anything for any research.
Benjamin Moses:
Their machines are in a lot of test facilities, research facilities.
Stephen LaMarca:
I don't mean this to be offensive, but they were like a laughing stock when I started in this, when I started at AMT and now they're like, they're staying at the same price, but now they have a pedigree. It's really cool what they've done.
Benjamin Moses:
Their transformation has been interesting.
Stephen LaMarca:
And they're really open to everything. So what a cool company. You wouldn't have caught me saying that five years ago, but let me go back here.
Benjamin Moses:
I do have questions.
Stephen LaMarca:
Sure.
Benjamin Moses:
I heard there's a nozzle you're playing with.
Stephen LaMarca:
Okay.
Benjamin Moses:
A hypersonic nozzle. Tell me more about that.
Stephen LaMarca:
But before we get into the fun stuff, just my final closing thought here that I heard from a lot of the people that were voicing, not concerns, but bottlenecks. Again, pain points. The people that are using hybrid when they are trying to train somebody and/or talk to their higher ups about acquiring the next machine, they've all said, can we please call it something other than additive? Let's avoid the words additive in 3D printing because it's scaring people off. Like my boss who signs off on machine tool acquisitions, if he hears a hybrid machine or an additive machine or 3D printing, he thinks, oh man, that's $1 million machine that we're going to have to hire two PhDs to run or else it'll just collect dust, be overhead and take up space.
And a lot of hybrid, a lot of the production quality production grade hybrid machines use DED direct energy deposition, which is using a laser electron beam and in some case an arc for usually powders and metal wire, the most popular are using a laser and metal wire. And they just say, instead of calling it additive or hybrid, can we please just call it CNC laser welding? Because then the boss and the new trainee can wrap their heads around it.
Benjamin Moses:
Gotcha.
Stephen LaMarca:
And we're talking places that have trained somebody that used to be a pizza delivery guy and now they're doing this stuff. Pre-pandemic we were all hybrid machines. They're cool and I hope they're the future, but you need two PhDs years ago. I don't mean to throw them under the bus, but I feel like, and I'm probably being inaccurate about this, but I feel like Roby Lynn and Kyle [inaudible 00:20:05] couldn't figure out how to get the Mazack hybrid that they had at Georgia Tech to work. So they actually did get it working pretty quickly, but not pretty quickly, but they got it working. It's a scary thing. And we should think the industry is talking about renaming additive. Dedicated additive machines, call it 3D printing, call it additive because it's still a buzzword that generates a lot, a lot of publicity. But hybrid CNC laser welding, direct energy deposition, stick to the specifics.
Benjamin Moses:
To piggyback off that comment, one of the other roadblocks with further technology adoption, that was pointed out in a session a while ago. So we talked about return on investment and understanding risks, but the basic understanding of technologies, companies that are lagging to adopt technologies, they're about five years behind where we are today. So I think to your point, understanding the current terms, it's scaling that back to certain audiences or understanding the audiences and what they know of the industry and tailoring the conversation towards that I think is very important because we're using what we think are standard terms, but depending on where they are in their technology adoption profile, they might not have the same understanding that you and I do or the rest of the industry does.
Stephen LaMarca:
Oh, absolutely not.
Benjamin Moses:
I think that's a fair point about adjusting the conversation and the conversation for technology to that specific audience and understanding their audience's needs. So that's an interesting point. So two takeaways. One is technology, transitional technologies. Then tailoring terms or definitions to very specific technology adoption profiles. Thanks Steve.
Stephen LaMarca:
And just the last bit. So you touched on, or you wanted to poke the bear about the nose cone or not nose cone. I'm sorry. So one of the things that Oak Ridge National Labs and Tom Feldhausen does with all his degrees and security clearances, they do a lot of the manufacturing for the research that goes into the US' knowledge of hypersonics. And one of the homecoming queen parts that they had on display there was a 3D printed hybrid manufactured rocket nozzle for a hypersonic missile.
Benjamin Moses:
Cool.
Stephen LaMarca:
This is quarter million dollar part made with multimillion dollar machines.
Benjamin Moses:
Very expensive.
Stephen LaMarca:
That was made by a bunch of PhD researchers. Very expensive, very impressive part. Incredible. And I saw this and got to get up close to it when we were on one of the plenty of tours during this workshop. The first tour we go on, I see this rocket nozzle and I happen to have an empty water bottle in my hand and I just go in my group and it's like, this is pretty cool. What an impressive piece. And I just take, oh man, don't need this anymore. And I toss it in the nozzle, like it's a waste bin. It got no laughs, it got zero laughs. And then to add insult to injury, I reach in there, I can't reach it. It's too deep. This nozzle is huge. So I couldn't get it out. And the next tour group is coming by and they're like, why are you reaching in there? I dropped my water bottle and this nice guy, Brad from Hypermill, which I have another story, we might have to save that for another podcast.
Benjamin Moses:
We'll save that.
Stephen LaMarca:
But Brad's like, "Oh no, no, no, don't. You could ruin your jacket. That's a sharp edge right there." And he's pretty tall and lanky. So he goes in and he's like, that's staying there. So there is a Walmart water bottle, disposable water bottle in a hypersonic.
Benjamin Moses:
That's pretty good.
Stephen LaMarca:
Rocket nozzle at Oak Ridge National Labs, and it's my fault. Please have me back next year.
Benjamin Moses:
RIP water bottle. Steve, do you want to share your hijinks with Tom or do you want to save that for next episode?
Stephen LaMarca:
Oh yeah. So I pride myself in having never sent a accidental text that was supposed to go to your significant other, to somebody other than your significant other. And sure enough, I sent my morning good morning, love. How are you feeling today? I'm looking forward to another great day at Oak Ridge National Lab. Thought I sent it to Melissa. Went to Tom. And fortunately this was at 7:30 in the morning. Nobody else. That day didn't start until 9:00. Nobody else is in the conference room yet. He's in another room, which really close by, but another room mixing his coffee. And a couple seconds later after sending that, I hear, aww, thanks hun. I'm glad you're feeling better. And I'm like, oh no.
Benjamin Moses:
First time for everything.
Stephen LaMarca:
There's a first time for everything. And at least it wasn't like a nude, nobody wants those, believe me.
Benjamin Moses:
That's a rough one to be sending at 7:00 in the morning, by the way.
Stephen LaMarca:
But then, okay, so this was on a Thursday morning. Friday, I'm back home getting ready for the weekend. I'm not doing anything. I worked from home that day because recovering from work travel and vacation travel, and I get a text at almost 11:00 from Tom Feldhausen. Good morning, love. I feel better today. But a little carsick on the way to Florida. He drove his family to Disneyland.
Benjamin Moses:
That's cool.
Stephen LaMarca:
And I kept it going. I said feel better, boo boo face, miss you. So that's I guess in industry Tom is my new work wife.
Benjamin Moses:
Congratulations Steve.
Stephen LaMarca:
Thanks.
Benjamin Moses:
I appreciate that. I was going to talk about Fabtech.
Stephen LaMarca:
Tell me about Fabtech.
Benjamin Moses:
Let's talk about that another time.
Stephen LaMarca:
Another time.
Benjamin Moses:
Let's save that for next week.
Stephen LaMarca:
All right, next time let's make a note. We'll say Fabtech and my run-in with Brad of Hypermill.
Benjamin Moses:
Oh yeah. All right. I'm making notes.
Stephen LaMarca:
Great guy. I can't wait to talk about him.
Benjamin Moses:
Brad.
Stephen LaMarca:
And Hyper Mill. And we'll talk about Fabtech too.
Benjamin Moses:
25 hour trips are not fun.
Stephen LaMarca:
Dude, you're wild. That was so impressive though. Did you have a bag?
Benjamin Moses:
I still check the bag because I was thinking about doing a work event that night and then something else I wanted to change when I got there and stuff. So I had a full bag. I don't carry all the luggage. I carry knives with me too, so I got to check all that stuff.
Stephen LaMarca:
I'm not going to lie. My favorite thing about day trips, leaving early in the morning and coming back late at night, is you only need one change of clothes. So you don't need to bring a bag if you're not bringing a computer. And if you don't bring a bag, it feels like when you walk onto a plane and walk off a plane and walk through the airport with no bags, it feels illegal. It's so nice, but too bad.
Benjamin Moses:
That's all right. It was fun.
Stephen LaMarca:
You learn the way.
Benjamin Moses:
Every trip, I keep learning. It was fun. Ramia, can you tell us about today's sponsor?
Ramia Lloyd:
Today's sponsor is MT Forecast. MT Forecast brings the latest economic news and industry trends straight to attendees. Industry leaders, executives, and key decision makers will explore an agenda that provides a roadmap to better business strategies through customer industry insights, economic forecasting, and deep dives into market data. For years, MT Forecast speakers has been sharing crucial looks into the near future. Go to amtonline.org/events to save the date and register.
Stephen LaMarca:
Good job.
Benjamin Moses:
Thanks Ramia.
Stephen LaMarca:
This is my favorite read. I like how you say for years.
Ramia Lloyd:
Got to keep them interested.
Benjamin Moses:
Steve.
Stephen LaMarca:
I booked my MT Forecast flight yesterday.
Benjamin Moses:
Nice. I got two articles. The first one is about generative AI from gameishard.org. GG.
Stephen LaMarca:
Wow. Okay.
Benjamin Moses:
It's interesting. So it's from Nvidia.
Stephen LaMarca:
It'd be cool if it was .gg.
Benjamin Moses:
So Nvidia to discuss the impact of generative AI in robotics. So Nvidia is in robotics and they do a lot of stuff supporting artificial intelligence, particularly in the hardware, but they have expanded their portfolio. So Nvidia is a big player, huge corporation. They do a lot of stuff. And what the article talks about are general concerns about artificial intelligence related to automatic decision-making. So one, what is generative AI related to robots? So it's autonomously generating new solutions to adapt to changing environments. So if you have a big part of bins, how does a robot know to pick, it's capable of picking things up? So it's learning itself, learning how to pick up things, or it takes a picture of the bin and it says, okay, I can pick this one up, but I know I can't pick this one up. So it makes a decision or learns along the way.
So it is training itself along the way. And a couple of things that it brings up as potential challenges, and it is more on the broader scale of artificial intelligence. It talks about ethical considerations which we've been talking about a lot. Data privacy, algorithm bias. And the last one, I'm not so sold on it, but potential displacement of human workers, which if we already talk about automation, that seems like it's a lower point. But I do want to mention that-
Stephen LaMarca:
You had me on all of those points.
Benjamin Moses:
Except the last one.
Stephen LaMarca:
Until the last one.
Benjamin Moses:
That's a bummer. But the other one, I definitely agree with it. The part bin picking is small when you look at the full scale of autonomous robotics because now you're talking about using vision systems, understand the scene. So when you look at ethical considerations, how do you determine is that a human? Is that a cat? There's still a lot of issues on classifiers determining humans versus other objects. And back to the data privacy, you're transmitting tons and tons of data to a black box somewhere. How do you follow that data and know that it's secure? And then the algorithm bias, that's still a big thing related to the ethical consideration. So there's a lot going on there.
Stephen LaMarca:
What's really cool, and I don't mean to harp back on the trip to Oak Ridge, but in this workshop there was actually a keynote speaker who talked about the AI and where AI helps most in the human brainstorming workflow. It of course helps middle management the most in that position.
Benjamin Moses:
And to be fair, we've been using AI for quite a while, we've been exploring open AI, and so the position that we've been supporting is it supports human development. And a lot of the use cases that we've found most useful is just like automation. It helps support humans to produce more, be more effective, be more efficient. It's an extension of automation.
Stephen LaMarca:
It is a tool. It is an extension of your capabilities.
Benjamin Moses:
But just like we talked about using Ware wrenches versus Craftsman snap-on. If you don't understand that tool and its capabilities, you're going to break it. And when it breaks, there's going to be a problem. So I think understanding the limitations of our AI tools, that's something that doesn't get discussed a lot. Everyone's like, oh, just use Chat GPT. Great. What are the bounds?
Stephen LaMarca:
You got to know how.
Benjamin Moses:
Exactly, how do you want to use it in the context of that small boundary to benefit you? I don't think we talk about that enough.
Stephen LaMarca:
No, no. And sure, there are some things. I wonder where the point of diminishing returns is because there's some things that we know not to do with AI platforms, large language modules. Is that what an LLM is?
Benjamin Moses:
Large natural language models.
Stephen LaMarca:
Model not module. And I put this on do not disturb. And the one thing that we learned right away is if you need it to summarize an article, you have to copy and paste the text of that article. And you used to have to put quotations around it so it knew the range to which to summarize and what not to if you had any other notes, be like put it in this tone of voice or something. You had to use quotations. But one of the things that we tried to do right off the rip was instead of copying the text, let's just copy the URL, let's send it to the website where the article is. The problem that we found out right away with that was it reads the entire website including advertisements. And if it can keep scrolling and if it's one of those websites where if you keep scrolling, it moves on to the next suggested article. You're in for a world of hurt.
And then if you're smart, here's if you do this and if anybody's listening and you're like, oh man, I do that now and you haven't caught on, it's because you're not reading what the AI's output is. Because if you do that and then read the output like what Chat GPT or whoever says the summary of the article is, and you'll notice right away that the figures like numbers and units of measurement that it puts in there are wildly wrong and erroneous and egregious. So that is stuff that by the time it becomes more mainstream, which it is becoming more mainstream than we would like it to be, but when it becomes like Google did in the nineties and was in elementary school computer labs and everybody was using Google, by the time it gets to that point, a lot of those problems that we ran into will probably be patched out.
Benjamin Moses:
I hope so.
Stephen LaMarca:
And fixed. But it's still, I feel like those still belong on a list of best practices. Reread what the output is.
Benjamin Moses:
Trust but verify.
Stephen LaMarca:
There you go.
Benjamin Moses:
Always there. [inaudible 00:34:23] told me about some clever additive technology.
Stephen LaMarca:
Clever additive, okay. I'll tell you about a smart move forward to a smart way to take a baby step forward in the development of additive technology and a really dumb way to name it. So let me find my article. It's from 3dprintingindustry.com. Extra 3D launches new 3D printing technology. The second you see that in any headline immediately go to hit the alarm. Sus. New technology. Okay, let's see. Extra 3D launches new 3D printing technology to eliminate SLA and DLP trade-offs. Eliminate. To eliminate.
Benjamin Moses:
Strong words. Word.
Stephen LaMarca:
If you read that improperly, you have to remember trade-offs at the very end. We're training in that large language module model here. If you read that wrong, you just read to eliminate SLA and DLP stereolithography apparatus. Both of these are vat photo polymerization. You have a vat, a bucket of resin and light is being shot into it to solidify the resin to make a part.
Benjamin Moses:
I like your hand motion.
Stephen LaMarca:
SLA uses a laser. DLP uses a illuminated screen like an LCD screen and lights up shapes that you want to solidify in the resin. And they say that they've created a new one, an entirely new one to eliminate both of these.
Benjamin Moses:
That's crazy.
Stephen LaMarca:
In truth. Ready for the truth. Ready for the crack over the head of the egg of truth. They took both of those technologies and put them in one machine. This is a vat photo polymerization 3D printer that has a laser and it has an LCD screen.
Benjamin Moses:
That's cool.
Stephen LaMarca:
To illuminate light. It is cool because you get the speed of digital light and you get the surface finish and accuracy of laser. So you don't have to worry about a slow print with a laser and you don't have to worry about awful surface finish with the LCD screen with a DLP, they've just put both devices in the same machine.
Benjamin Moses:
Little combo. Combo machine.
Stephen LaMarca:
It's not new. They've just made a new hybrid. And then to add insult to injury, they gave it a stupid name. It's called HPS because got to cater acronyms to the military. The DOD, they called HPS. You know what HPS means?
Benjamin Moses:
What does that stand for?
Stephen LaMarca:
Hybrid photosynthesis. Photosynthesis is an organic thing that happens in chloroplast, in chlorophyll. None of that is part of this. It's not organic at all. This is a machine. This is vat photo polymerization. There's no biology going on in this.
Benjamin Moses:
I feel like they asked Chat GPT to name it and they didn't check it. They didn't check the [inaudible 00:37:31]
Stephen LaMarca:
They just sent it links and didn't read it.
Benjamin Moses:
That's funny.
Stephen LaMarca:
So Extra 3D. Good job. Clever, but nobody likes clever. Get a new name. Good job.
Benjamin Moses:
Good job.
Stephen LaMarca:
You did it.
Ramia Lloyd:
So wait, they eliminated it by combining it?
Benjamin Moses:
Exactly, exactly.
Stephen LaMarca:
They subtracted by adding. Oh, it's hybrid. Wow. Write that down. Extra. Okay. This is the industry we live in.
Benjamin Moses:
This is the last article I have is from Composite World. I think you and I agree. Composites are cool.
Stephen LaMarca:
Composites are tight.
Benjamin Moses:
Like some carbon fiber. So it talks about carbon fiber reinforced plastics.
Stephen LaMarca:
Let's see [inaudible 00:38:14]
Benjamin Moses:
For large aircraft structural parts.
Stephen LaMarca:
Are we talking about fiber placement, bro?
Benjamin Moses:
We're talking about fiber placement. So we're talking about structures the size and support the A350 aircraft. This one talks about new ways for collaborative robots to support creation of these parts using-
Stephen LaMarca:
It's a little spicy.
Benjamin Moses:
Using carbon reinforced [inaudible 00:38:41] fabric. And the cool thing, so tape and tow replacement has been around for a little bit.
Stephen LaMarca:
It is expensive.
Benjamin Moses:
It's difficult, it's big. You have tooling cost.
Stephen LaMarca:
It's basically a gantry machine lathe. But on the gantry there's a robot arm with a end of arm tooling that lays down a fiber tape or tow of carbon fiber.
Benjamin Moses:
So in this example, they use for scale, they've got robots on a linear axis that's 30 feet long. So fairly long track.
Stephen LaMarca:
Or you could get a big machine.
Benjamin Moses:
And the interesting part about this, they have a human working in close proximity of the robot. So the robot is supporting the human as they're placing the material. So for long cuts of four meters in length, the robot holds material above the mold, giving the worker just a right amount of material to allow the curvature. For shorter pieces, the robot holds the material on various spots, pressing tightly on the surface mold, providing the seed points for correct manual draping.
Stephen LaMarca:
This sounds like less expensive, but more complex fiber tape, tow placement. The Russians are going to be all over it.
Benjamin Moses:
I really like this because we've been talking about modifying the definition of what we consider collaborative robots to collaborative operations. This is very supportive and there's always a concern about automation replacing humans. This is literally humans working with robots. What more do you want to work for?
Stephen LaMarca:
This is as collaborative as it gets.
Benjamin Moses:
So I thought it was a fantastic article.
Stephen LaMarca:
So far.
Benjamin Moses:
I think this type of concept has a lot more potential expansion in connection with Fabtech. So a lot of the industry events that we go to is making a specific part, a one part, but when we scale that up, the part always has to go into an assembly. And this is an assembly at this point. So how do I put pieces together? How do I hold that? Some of that could be automated, but at some point a human is probably involved screwing things together, welding, forming, joining, gluing things. So one scenario I see growing quite a bit is how do we present the part to the human in a effective manner for them to be fast, efficient and reduce stress on human fatigue? And I think we'll see a lot more growing use cases for that. Now, do you want to buy a $100,000 robot to hold a bike in the correct orientation for you so you can work on a floor? Maybe not, but maybe-
Stephen LaMarca:
If you sell $10,000 bikes. Yeah.
Benjamin Moses:
Or maybe holding a transmission so a human can install the sensitive parts that you need. So I think there's a lot more potential. And what we're seeing is the reduction in cost of certain robots and the increasing capability. So between that intersection and the ability of, hey, let's test this out on a $15,000 robot and see if it works, and then buy something that's going to last longer back to the transitional technology. So I think we're at a very interesting crossroads where we could see a lot of more interesting use cases for that.
Stephen LaMarca:
This is awesome and I appreciate it because it's bringing fiber placement back into the spotlight that it really never had. It's not popular, but it's such a cool technology.
Benjamin Moses:
Agree.
Stephen LaMarca:
It's also untouchable in terms of its pricing. I was talking about $1 million machines for hybrid. Fiber placement machines are tens if not hundreds of millions of dollars.
Benjamin Moses:
They're big.
Stephen LaMarca:
Per machine. I think there's nine in existence in the US.
Benjamin Moses:
Depending on the scale.
Stephen LaMarca:
They're all in research or Boeing.
Benjamin Moses:
So depending on the use case too. So if they're military use cases, then it gets tough getting access to that. But they're used on large barrels of fuselages.
Stephen LaMarca:
If this technology is viable or if this method of implementing technology is viable, I think it will bring more popularity and publicity to fiber placement, dedicated fiber placement machines and maybe bring their costs down because with more demand, they're going to need to supply it more. So immediately there's going to be a price spike, but then it might come down. There's certainly going to be more people inquiring about it.
Benjamin Moses:
I think so.
Stephen LaMarca:
But if robots are good enough, your off the shelf robot is good enough to do work like this, it's going to go against what I said, and I think it might actually increase the prices of robots.
Benjamin Moses:
It could be. And there's two benefits to this. One. What's the return on investment? I see a lot of operational throughput increase. So instead of having a human do everything, having the robot support it, so you're getting a faster output. And then the next operation is can we incorporate some type of inspection, automated inspection as part of this? So there's a lot of opportunity to increase the throughput of these parts.
Stephen LaMarca:
Cool.
Benjamin Moses:
A lot of good articles.
Stephen LaMarca:
This was awesome. Great episode.
Benjamin Moses:
Ramia, where can they find more info about us?
Ramia Lloyd:
amtonline.org/resources.
Stephen LaMarca:
Bing bong.
Ramia Lloyd:
Like and subscribe.
Benjamin Moses:
Bye everyone.
