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
Ramia Lloyd:
Welcome to the TechTrends 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. I'm Ramia Lloyd, associate content manager, and TechTrends podcast producer.
Benjamin Moses:
Nice.
Ramia Lloyd:
And I'm here with...
Elissa Davis:
I'm Elissa Davis. I'm the digital community specialist for AMT.
Stephen LaMarca:
Stephen LaMarca, senior technology analyst.
Benjamin Moses:
Benjamin Moses, senior director of technology. Guys, that was the driest intro ever, but-
Stephen LaMarca:
We're old.
Elissa Davis:
Yeah.
Ramia Lloyd:
Yeah.
Stephen LaMarca:
We're old.
Ramia Lloyd:
It really just depends the way you flex it on us.
Stephen LaMarca:
We're seniors.
Benjamin Moses:
A lot of seniors, here. I want to jump into some test bed topics. I know, because we've been making quite a few advancements this past couple of months. Steve, fill us in on hardware updates.
Stephen LaMarca:
So we made huge advancements in the last two months, towards building, literally a digital twin of our test bed. And speaking of testing, that has tested our computers that we currently have over at the test bed. Which is a hand-me-down, a decommissioned AMT development computer, which is a strong computer. But we've tested its limits, and we found them very quickly, and we learned pretty quickly that we need a proper, dedicated computer for doing digital twin simulation, modeling, and AI work. And when Sharab looked at, in the NVIDIA Omniverse Isaac Sim system requirements, it laid out in a nice little table. Minimum requirements, recommended requirements, and ideal requirements.
Benjamin Moses:
Right. I like those tables, because they're common, across everywhere.
Stephen LaMarca:
Yeah, across games.
Benjamin Moses:
Across video games.
Stephen LaMarca:
Like, video games tell you... Well, non-console games. PC games tell you, "You should probably... You need this, to run this game. You probably should have this, but if you really want the best experience, you want this." It's amazing how software just does that. It's very nice.
Benjamin Moses:
I think you're going to hit on one of the points that I hit on about that table, too.
Stephen LaMarca:
I don't know what point you're talking about, but we'll probably get there. The recommended, meaning like, "This is what you should have, if you really want to get the full, the best experience with our software possible for Isaac Sim, we specced out some computers, and went to IT and talked to them. And we were like, "Okay. We need a computer from our supplier," which is Dell, "that has all of this hardware and capability." And they came back to us and it was like, "This is going to be a $9,000 computer."
Benjamin Moses:
So to be fair, that's the recommended setting.
Stephen LaMarca:
Recommended. Then we looked at ideal. Ideal.
Benjamin Moses:
Okay, you looked at ideal.
Stephen LaMarca:
Then we looked at recommended, which is in the middle. Sorry, I've messed that up. Ideal is best, mid-tier, recommended. And then minimum, which we already have. We have minimum, and it's not enough, so we're looking at recommended. We looked at the recommended PC from Dell, and they were like, "$5,500." And it's like, "Oof." It's still a heartache.
Benjamin Moses:
That's the funny part about the table, that I see across the board everywhere is, the basic is like a thousand dollars, but even on the gaming side, the high end is literally like $5,000. More.
Stephen LaMarca:
Yeah.
Benjamin Moses:
So the total cost, it's already, the spread on the table you've got right there.
Stephen LaMarca:
Which is funny, because for gaming, as an individual gamer with a limited pocketbook. Like $5,000, that's a lot of money for something that's going to be out of date in four years.
Elissa Davis:
It's expensive to be a nerd.
Stephen LaMarca:
It is so expensive to be a nerd.
Benjamin Moses:
That's the realest advice ever.
Stephen LaMarca:
But for computational power in a corporate structure, and in something like manufacturing, modeling, AI work, digital twin. Chump change, that is chump change for them. But the test bed runs on a limited $10,000 annual budget, so we can't be dropping 90% of that on a computer that we're going to use once a week.
Elissa Davis:
Understandable.
Stephen LaMarca:
The computer is designed, those system requirements, that software is used, or is to be used in a professional and industrial setting 24/7. That thing doesn't get turned off, and it doesn't have downtime. It's constantly running. We realized we don't need that, number one. Number two, this is a corporate price. Like, they know this is a corporate tool, Dell knows this is a corporate tool. They know that if you want to play with the big dogs, and if you are a big dog, a 9K computer is... You can do that, if you need it. It's not hard.
Benjamin Moses:
Well, to be fair, it's coming pre-built, right?
Stephen LaMarca:
It's got some pre-built, as well.
Benjamin Moses:
With a warranty, and all their services, right?
Stephen LaMarca:
Not just a warranty. Warranty is like, you can go on Amazon and buy something. We could have got a warranty for the GPU that we bought for the, spoiler alert, PC we're building. But it's not the same as what Dell would offer, which is a three-year service plan. If something goes wrong, if something dies on it, it's no questions asked. You send it back to them, or they come and get it, or diagnose it. And if they determine, "Yeah, something crapped out on our end, we're replacing it." That's good for three years, and that's worth a lot.
But one, with our limited budget, and with the way nerds can be very frugal using a website like pcpartpicker.com, we were able to build a computer, spec out a computer, that is like... Almost all of the components are in the recommended tier, and the GPU is in between recommended... No, all of the components are in ideal. The GPU we bought is between recommended and ideal, because the GPU is like-
Benjamin Moses:
That's the biggest cost, actually.
Stephen LaMarca:
The computer without the GPU, that we built, is insanely powerful. And it's $1,500 without the GPU. After buying the GPU? $3,600. That's $2,100 for something that just does a bunch of really dumb, easy calculations, a lot at once, really fast.
Benjamin Moses:
I think that's the biggest takeaway, because I've been thinking about the test bed in general, and the ability to get to... Because all of our equipment is accessible, right?
Stephen LaMarca:
Yeah.
Benjamin Moses:
It's low footprint, low cost, for the most part. Based on our budget, our year-over-year spending. But getting into the workstation environment, especially on the test bed environment, being able to just... I wouldn't say hack your way through it, but build your own, to see. Is this investment worthwhile, to say, "I need to go to Dell." Because to your point, we need to see if it's valuable, and what type of scenarios you can run, and then get into a production environment. So this is a really useful experiment.
Stephen LaMarca:
And I just have to throw in a slight towards Apple here, and if you think that the Dell price tag is high, that's cute. Because that's like a third of what you would get for the equivalent computer from Apple. Apple would take our build, a $3,600 build, easily add a zero to it. That would be a $36,000 computer from Apple.
Benjamin Moses:
Add another thousand dollars, if you want wheels on top of your workstation. And you do get into some really interesting points that we'll hit on later, is about computing, the need for computing. Because we're talking about cooling, and energy usage, and stuff like that.
Stephen LaMarca:
Right.
Benjamin Moses:
There's a really big difference on... Because that's a workstation GPU, basically, right?
Stephen LaMarca:
Yeah, a workstation GPU. So it's not for gaming. It could game, but it's not designed for that. And that's an interesting topic, because I don't think a lot of gamers can wrap their head around that. Like, "It's twice the price of a gaming GPU, and it doesn't game as well. That doesn't make sense to me." Well, it's not for doing that. It's for a different reason. It's for thermals, it's for stability. The industrial workstation GPU is running 24/7. It doesn't stop. Your gaming GPU is more powerful, but it's only gaming for six hours on end, until your eyes bleed. And then you turn it off and go to bed.
Benjamin Moses:
Yeah.
Elissa Davis:
My brother had a crisis this week. Well, last week, my parents were in Hawaii, so my brother was dog-sitting for them. He also cat sat for me, while I was at MTForecast, so props to my brother for driving from Ashburn to Fairfax, to do that for me. But he has a gaming headset that he uses, because he's built his own computer, he has a computer at school that he built. He needs that for gaming. And he texted our family group chat yesterday and he goes, "I forgot my headphones at our parents' house."
Benjamin Moses:
He's distraught.
Elissa Davis:
And he lives on campus.
Ramia Lloyd:
Stressed.
Elissa Davis:
In the dorms-
Ramia Lloyd:
Struggling.
Elissa Davis:
... at Mason. And so he was like, "Okay, I don't know," but he doesn't have a car. So he's like, "I don't know when I'm going to get to our parents' house, until probably at least this weekend." And so he was like, "Okay, I guess I'll just have to get an adapter, like a Bluetooth, like an aux cord, to plug in. And I was like, "Well, if you need headphones, I have a ton of Bluetooth headphones, you can just borrow a pair of my earbuds." He goes, "No. My computer is not Bluetooth compatible. I have to have the aux cord." And I was like-
Benjamin Moses:
Man.
Elissa Davis:
... "That sounds... For the computer that you built, it doesn't have the Bluetooth compatibility." That's wild to me. But he was just like, "It's fine. I'll just see..." He's going to see if his boyfriend can take him to my parents' house today, to get his headphones.
Stephen LaMarca:
I don't think you're giving your brother enough credit. He was saving you from a mansplaination, of basically that, any real gamer does not connect through Bluetooth. It's not stable enough.
Elissa Davis:
No, I know.
Stephen LaMarca:
You want to hard-wire that cable, and he saved you from that explanation, to which I'm here to fulfill.
Elissa Davis:
My brother has four sisters. He would never try to mansplain anything to me. He knows better than that.
Stephen LaMarca:
God bless him. What a gem.
Elissa Davis:
But yeah, so that just made me think of that. Because I was like, "Oh, he built this whole computer," and then it was like, "Oh, I have four pairs of Bluetooth headphones. Do you want to borrow one?" "No, it doesn't work." And I was like, "What?" Because you know, growing up in the age of plug-in headphones, you're just like, "Ooh, these Bluetooth ones are so cool," but gamers hard-wire into the set, or the computer, whatever it's called.
Benjamin Moses:
A whole set of podcast party, you can see all the cables I have, coming out of a machine. There's nothing wireless in my setup at home. So mansplaining, and other lingo terms, that should die. Steve, you hit out some corporate terms?
Stephen LaMarca:
Corporate lingo that needs to die. And I tried to pick Elissa's brain earlier, about things that need to die on LinkedIn. Because that's where my, this is why I'm bringing it back up. Immediately after, maybe not Saturday after IMTS, but Sunday after IMTS, my LinkedIn feed was flooded with literally the same two phrases. For the entire next two weeks.
Benjamin Moses:
Was it GTK?
Stephen LaMarca:
"And that's a wrap. This is one for the books." You don't have a book, you don't write a book. There's no book. Stop saying it's for, "This is one for the books." What books? Everything's digital now.
Benjamin Moses:
This the one for the Kindle?
Stephen LaMarca:
"That's a wrap." That's a wrap. What does that even mean? What is the etymology of the phrase, "That's a wrap." What's being wrapped up?
Benjamin Moses:
Why are you so upset about this? I don't understand.
Stephen LaMarca:
I saw it so much. It was the only thing on my LinkedIn. It was the only things that I, "And that's a wrap," from literally all of the people in the world. It's like, you can't have an original thought? You just came from IMTS, and you're using literally the same thing everybody else is saying?
Elissa Davis:
So as the person who runs our social media, I'm going to tread lightly here, but on a platform like LinkedIn-
Stephen LaMarca:
I need aspirin.
Elissa Davis:
... on a platform like LinkedIn, it is very group-sought-ish, where like-
Stephen LaMarca:
Hive mind?
Elissa Davis:
Yeah.
Ramia Lloyd:
Yes.
Elissa Davis:
So if someone sees it, someone else is going to post that. Whether they even realize they're doing it or not.
Stephen LaMarca:
My?
Elissa Davis:
Yeah, they're going to post that, right? And sometimes, there are times when I'm sitting there staring at the computer like, "What is a good phrase?" And I'll be like, "That's a wrap." I'm like, "No. We use that too much." But I have to think about us as the brand, not just myself. If it was just myself, I'd probably post "that's a wrap," to be honest with you. Because I'd just be like, "Okay." Because that's wrap leads into the recap, right? So I have to be careful, because I run our social media, and I don't want people to stop tagging us and stuff.
Stephen LaMarca:
So you run social media-
Elissa Davis:
Yeah.
Stephen LaMarca:
... but you're not an influencer.
Elissa Davis:
No.
Stephen LaMarca:
I saw it from a lot of the influencers, too.
Elissa Davis:
Of course. Yeah.
Stephen LaMarca:
It's like, you're supposed, it's your job to have an original thought.
Ramia Lloyd:
Do notes. As a person-
Elissa Davis:
Do notes.
Ramia Lloyd:
... who probably put, "That's a wrap," in her LinkedIn post. In my case it counts, because "that's a wrap" originated in the film industry, and I worked in media.
Elissa Davis:
That's true.
Ramia Lloyd:
That's all I did, before I [inaudible 00:13:42]-
Stephen LaMarca:
What does it mean?
Ramia Lloyd:
It's like, a wrap.
Stephen LaMarca:
What's being wrapped?
Ramia Lloyd:
Like you're wrapping up the end of the filming progress. The filming is wrapped.
Stephen LaMarca:
But we still need to break down the etymology of wrapping up. What is literally being wrapped up?
Ramia Lloyd:
You wrap it-
Stephen LaMarca:
Apart from the day?
Ramia Lloyd:
That's the-
Stephen LaMarca:
Oh!
Ramia Lloyd:
... you're wrapping up, the carpet. You're wrapping up your filming.
Stephen LaMarca:
Maybe. Is that it?
Elissa Davis:
I don't know.
Ramia Lloyd:
No, it literally means you're-
Stephen LaMarca:
It's so believable.
Elissa Davis:
I think of like, red carpet. You're wrapping it up. Right?
Ramia Lloyd:
It's like, it's from film.
Elissa Davis:
Rolling it up.
Ramia Lloyd:
You're wrapped up. So when the film ends, and you're finished filming, you have a wrap party. You're just wrapping it all up. That's the end of it.
Stephen LaMarca:
Thank you.
Ramia Lloyd:
So that's my explanation for it.
Stephen LaMarca:
That's a good answer, thank you.
Benjamin Moses:
Are you less angry now?
Stephen LaMarca:
Yeah.
Benjamin Moses:
Can we still kill it?
Ramia Lloyd:
We can. Yeah. But I'm telling you, that's why I even say it.
Stephen LaMarca:
Thank you.
Benjamin Moses:
I really like the escalation of a term. That's fair, I mean, it's over-popularized. But it could have taken two seconds for you to answer this, Steve, before we got here. You could have looked up the etymology of "that's a wrap."
Stephen LaMarca:
But wasn't this better for the camera?
Benjamin Moses:
Fine, Chris.
Ramia Lloyd:
And that's a wrap. LinkedIn person.
Elissa Davis:
I mean, they do say, like the "one for the books." That is one that, I do kind of... I use that, and that's partially because-
Stephen LaMarca:
But you're actually writing a book. I feel like you're allowed to say that.
Elissa Davis:
What if you journal? Does that count?
Ramia Lloyd:
Yeah, I would count-
Elissa Davis:
Because I talked about IMDS a lot, in my journal. So does that count?
Benjamin Moses:
Next level is-
Stephen LaMarca:
You're allowed. I would give you a pass. Thanks.
Benjamin Moses:
... if they start putting the acronym, the acronym for, "That's one for the books," that would throw me over the edge.
Stephen LaMarca:
T-O-F-
Elissa Davis:
F-T-B?
Stephen LaMarca:
... T-B. That actually, that sounds better.
Ramia Lloyd:
TOFTB?
Stephen LaMarca:
TOFTB.
Ramia Lloyd:
TOTFB.
Benjamin Moses:
So Steve, you also mentioned something, words we should stop using about robots?
Stephen LaMarca:
Oh. A word that we don't have to stop using yet, because it's just become a bad word, but it's not cancelable yet. Is the derogatory term for robots.
Benjamin Moses:
Tell me.
Stephen LaMarca:
I'm not going to say it.
Benjamin Moses:
Are you waiting for me to say it?
Stephen LaMarca:
Somebody else say it.
Benjamin Moses:
Clankers.
Stephen LaMarca:
Yeah. "Send them back to the factory."
Ramia Lloyd:
Why would you make him say that?
Benjamin Moses:
How did that come up? Why did you ever run cross that?
Stephen LaMarca:
It's something, I saw it on social media.
Benjamin Moses:
Okay.
Ramia Lloyd:
Send them back to the factory, it's [inaudible 00:15:58].
Elissa Davis:
I mean-
Benjamin Moses:
The idea, so I do have an interesting concept I want to talk to you guys about, about humanizing robots, in general. Of humanizing technology, in general. So, giving a tool a human name, I don't think that's useful in the future. How do you guys feel about that? I think we should just call it-
Stephen LaMarca:
It's amazing that you're bringing that up now, because my house, we just watched the first Ant-Man movie.
Benjamin Moses:
Sure.
Elissa Davis:
Oh my God.
Benjamin Moses:
And Michael Douglas's character, I forget the name. Professor Pym, or something like that.
Ramia Lloyd:
Yeah. Hank.
Elissa Davis:
Hank.
Stephen LaMarca:
Hank Pym. He's controlling the ants. He can communicate with the ants. And he's like, "They don't have names, they have numbers." Do you even realize how many ants there are? They can't all have a name. They have numbers. It was like-
Ramia Lloyd:
Ant-thony said, "It's a wrap."
Stephen LaMarca:
... "This is very insightful." The Ant-Man-
Benjamin Moses:
But-
Stephen LaMarca:
... that was an insightful movie. Robots shouldn't have names.
Benjamin Moses:
... the counterpoint was, the Ant-Man that, I forgot his name. Paul Rudd's character, does start naming the characters.
Ramia Lloyd:
Ant-thony.
Benjamin Moses:
Yeah.
Ramia Lloyd:
His name is Ant-thony.
Stephen LaMarca:
And he's amazing.
Benjamin Moses:
But to your point, I mean, if you look at, their intellectual, kind in between of machinery, versus something that could be anthropomorphized, but when you look at a piece of [inaudible 00:17:15]-
Stephen LaMarca:
High ranking robots are allowed to have names.
Benjamin Moses:
High ranking?
Stephen LaMarca:
Yeah. Ant-thony was just, like the one ant that they constantly went back to.
Elissa Davis:
So my question is then, at IMTS we had Apollo, from Apptroniks. His name was Apollo.
Stephen LaMarca:
Was that robot's name Apollo, or was that the model?
Elissa Davis:
That's a great question. I don't know. I think-
Stephen LaMarca:
Because I would like to know Apollo's serial number, and then I'm calling him by the last four digits. If it's a him. They, them.
Benjamin Moses:
While we're talking about robots, I do want to about Tesla. They released a bunch of stuff recently, so we'll start a little further upstream. Michael Mark brought this up, about Tesla, obviously working with humanoid robots. To your point, we saw Apollo at IMTS, but there's a big influence in big resurgence of humanoid robots. And I feel like there's a parallel of lessons learned from AMRs, and their influence, to where we are with humanoid robots. Where BMW is experimenting with them, in their factory. There's still a lot of experimentation, like pre-innovation space still in the hype cycle, I think. What is your guys' hot takes on humanoid robots?
Elissa Davis:
So, first I just want to say, I saw something on Twitter where it literally compared shots from I, Robot to the things that Tesla released. And they're exactly the same. And I just feel like that's important to note, and just shows further that Elon Musk is trying to destroy this world. Anyway.
Stephen LaMarca:
He needs a reason to get to Mars, if he destroys this.
Ramia Lloyd:
I love your long-standing beef with Elon Musk.
Elissa Davis:
I really-
Ramia Lloyd:
Because I'm supportive.
Elissa Davis:
... I've never liked him.
Ramia Lloyd:
I'm so supportive.
Elissa Davis:
Even when people were like, "Ooh, he's changing the world."
Ramia Lloyd:
No.
Elissa Davis:
I'm like, "But he could change it for the better, and he's not."
Ramia Lloyd:
He's [inaudible 00:18:54].
Elissa Davis:
Like, he has the money to do it.
Benjamin Moses:
All right, back to humanoid robots.
Elissa Davis:
Anyway.
Ramia Lloyd:
[inaudible 00:18:56] that man.
Elissa Davis:
Humanoid robots. I mean, I don't know. I thought Apollo was just really cool, at IMTS.
Benjamin Moses:
Yeah, yeah. It was.
Elissa Davis:
And how they're teaching him, through modeling, how to do everything. So he has that kind of human aspect, but it's not scary human.
Benjamin Moses:
Sure.
Elissa Davis:
But he has a very non-threatening face, where Elon's robots have no face.
Benjamin Moses:
Okay.
Elissa Davis:
And that makes me-
Benjamin Moses:
Did Apollo not, did it have a face?
Elissa Davis:
He had a face.
Benjamin Moses:
It's subtle. It's very subtle.
Elissa Davis:
Yes.
Benjamin Moses:
Because he had the cameras built in, so he had the stereo vision.
Elissa Davis:
I don't he had a mouth, he had like, eyes.
Stephen LaMarca:
But they were built, I wouldn't say they were built in, they were modular cameras. Because that was an Intel RealSense up front, and there was also an Intel RealSense camera in the back of his head. So it was like, "This robot has eyes in the back of his head. I don't trust this robot. He literally has eyes in the back of his head." That's not true. Apollo is fine. It's a piece of equipment. But my favorite comment on Apollo was Asa, Asa Gurney was like, "Yo, Apollo doesn't skip leg day. You've seen that guy and his quads?" It was like, "Excellent. Astute observation."
Benjamin Moses:
Because it was under the table, so it's a good observation.
Ramia Lloyd:
I'm just peeking under the table at Apollo's legs. I also want go back really quick, to the Ant and Ant-Man. The other one's name is Ant-tonio Banderas. I love that. So there's-
Stephen LaMarca:
I totally missed that one.
Ramia Lloyd:
... two of them with names.
Benjamin Moses:
That was good.
Ramia Lloyd:
Thank you.
Benjamin Moses:
So many good puns. Paul Rudd's good for puns.
Ramia Lloyd:
Anything Marvel, I got you. I was like, "I know they have names." Ant-thony.
Benjamin Moses:
The last thing I want to hit on with Tesla is, they just released their robotaxis, which I think is funny, because there's like 30 different names for the robotaxi, that they aren't official yet.
Elissa Davis:
Once again, looks just like I, Robot.
Benjamin Moses:
I, Robot just picks you up, and carries her around. But what they're releasing is a car that has no drivability, it completely drives itself.
Stephen LaMarca:
I love that.
Benjamin Moses:
It's literally, you order it, and it pulls up. There's two seats in this one, that they unveiled. No steering wheel, no pedals. There's no buttons for a door, the doors open by themselves, and you have to use the touchscreen to open the doors. And supposedly priced pretty low, but as people that travel quite a bit, I want you to get your hot take about ordering a... It is a driverless car, because there's no ability to drive it at all. What's your confidence, as in right now, to order a driverless car?
Elissa Davis:
See, I'm split on that.
Stephen LaMarca:
Same one.
Elissa Davis:
Because one, I'm less likely to be human trafficked in a driverless car.
Stephen LaMarca:
Excellent point.
Elissa Davis:
However, if someone's controlling the robotaxi, then that's a whole other story. But I think, I've just heard too many horror stories about the self-driving cars, and stuff like that. That I would, unless it was on a track, I would get nervous.
Benjamin Moses:
Okay.
Ramia Lloyd:
Yes.
Elissa Davis:
I trust things on a track, like a trolley, or like-
Ramia Lloyd:
Yes. I need it to be on like, a trolley track.
Stephen LaMarca:
Yeah.
Elissa Davis:
... a cable car-
Ramia Lloyd:
I agree.
Elissa Davis:
... anything like that, I'd be like, "Okay, fine. It's got a direction to go in." But if it's just on the street?
Ramia Lloyd:
Free, free roaming.
Elissa Davis:
I've seen humans who are absolutely insane drivers. I wouldn't trust a car that doesn't have a driver, to not be insane.
Ramia Lloyd:
It's kind of cool, I think, though.
Stephen LaMarca:
I don't know, it seems that cleans up some of the flaw, of humans. I pass humans on the way to work every day, that are supposed to be driving their car, that has controls in their vehicle. And they are literally in their right hand, watching YouTube, on their phone.
Elissa Davis:
Yeah, I've seen that, too.
Stephen LaMarca:
And it's like, "You know what? Can you just not have, if you don't want to drive, you shouldn't have a car that can be driven. You should be driven in a car." And not everybody has the money for a chauffeur, so I think it's great.
Also, once the algorithm, I'm just using that as a buzzword, I don't know if it's an algorithm or not. But once the programming is ironed out, to where it's much safer, because there have been problems with it. Teslas, their autopilot mode, which they tell you you shouldn't be using. Like, "It's just a fun thing to have. You should experiment with it every now and then."
Ramia Lloyd:
And what's the point, you're on it.
Stephen LaMarca:
"You should not trust it." It's freedom. But it's been known to kill motorcyclists, especially on HOV lanes, because it sees cruiser motorcycles like Harleys, and it sees the two little tail lights in the back, down low by the rear wheel. The vision systems are not advanced enough to tell that that is a motorcycle that's very close, it thinks it's a car very far away, so it speeds up and runs over a motorcyclist. It's happened. They're covering up a lot of motorcyclist deaths. That's something I'm very, not cool, about Tesla.
Elissa Davis:
Well, they've been recalled in China a bazillion times, for the self-driving stuff.
Stephen LaMarca:
Also, do they care if people die?
But the other thing is, that I do think is cool, is if the navigation in the Tesla autopilots has to be rerouted, the programming knows the capability of the vehicle. And when people were first playing with the autopilot, and they had to reroute mid-driving, mid-transit, the car would literally do an e-brake turn in the middle of the road, and change directions. And it was freaking out the passengers. It was like, "Oh, humans have a G tolerance. We can't be doing J turns in the middle of traffic, regardless of how safe it may be because nobody else is around. Not a cool thing to do." And the driver still gets pulled over. It wasn't the driver that did it, but the driver, they're still the captain of the ship. And they're the ones who are in control.
I think it's great, because as somebody that still likes to drive, if 70% of the cars around me are autonomous, and following a guideline, or following guidelines that... Hey, there are traffic laws in place now that people don't follow, because they're on YouTube. If the program, which can be trusted to follow the guidelines better than humans can, then think of all of the autonomous vehicles that will be on the road. Let's say it's 70% of them. Think about how many cars you can cut off, and they will put on the brakes, and you can speed away. And they won't think anything of it, because the person who should be driving, who isn't driving, isn't even paying attention anyway. It would reduce the stress of your commute by so much. I have no figures on anything.
Benjamin Moses:
There's two takeaways, back to, basically autonomous machines in general. We started the conversation on human robots, and obviously with Tesla's robotaxis. The key thing is, and I think it gets into a lot of technology adoption issues, there's a lack of trust. There's a lot we don't know about it, and the observable influences versus what the company's pushing. To your point of accidents with motorcycles, that's a fairly big issue, but no one's talking about that. But the idea of, we know there's flaws in the system, and there are companies like in this scenario, where they're willing to push as much as they can, learn as they're going into production. And then feed that back into the overall confidence model.
It's an interesting point, about the trust in the technology, versus the actual use case. To your point, you hit on a scenario where... In the Northern Virginia area, everyone's on their phone.
Elissa Davis:
Oh my God, it's so scary.
Benjamin Moses:
Deepa yells at me all the time, because I give people two seconds at the light. If they aren't moving, I honk, because I know they're texting.
Elissa Davis:
That's true.
Stephen LaMarca:
Your horn is not optional equipment, it's standard equipment. You should use it.
Benjamin Moses:
So the problem statement of, "distracted humans needs to be solved with autonomous driving," that is one solution. I mean, there's other ways to solve the problem of distracted driving, too. So back to the technology issue, where we're all interested in is, there's still a problem with building trust with autonomous equipment, because there's a lot we don't know. There's the black box that we're unsure of, and that's where I think the next evolution is about, with the autonomous vehicles. And autonomous technology is, how do you build trust with that equipment?
Elissa Davis:
If I got into a taxi, and there was no one driving, I'd be really scared. Like if I got into a regular taxi, and then it just started moving, and no one was driving?
Benjamin Moses:
Yeah. That's throw you off.
Elissa Davis:
I'd think I was dreaming. I'd think I'd be like, "This is my nightmare."
Benjamin Moses:
Like if you jumped in a vehicle and it was like, what was it, Waymo. That's retrofitting a lot of cars. You didn't see the branding, and just jump in, it's like, no cars. And then it takes away. That would be alarming.
Elissa Davis:
That'd be terrifying.
Stephen LaMarca:
Every other Lyft, Uber, taxi that I get into, has a tire pressure light, maintenance required light-
Ramia Lloyd:
Facts.
Stephen LaMarca:
... or a check engine light.
Elissa Davis:
The tire light, I let go.
Stephen LaMarca:
I keep... The dry... No. I'm not going to let that go.
Ramia Lloyd:
No.
Elissa Davis:
Mine is-
Stephen LaMarca:
That's what killed Paul Walker.
Elissa Davis:
... mine is so sensitive, though. None of my tires are low, if my low tire light is on.
Benjamin Moses:
Well, that's a different problem statement.
Stephen LaMarca:
Yeah, well-
Elissa Davis:
Yeah, that's true.
Stephen LaMarca:
... you need better sensors.
Benjamin Moses:
Let's move to today's sponsor, Ramia.
Ramia Lloyd:
Tune in for Modern Machine Shop's Made in the USA podcast to explore manufacturing issues faced by companies making an intentional choice to Manufacturing in the US. Featuring commentary from OEM leaders, Made in the USA blends its nearly 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, Ramia.
Ramia Lloyd:
Reading is hard.
Stephen LaMarca:
I think you did a great job with that.
Elissa Davis:
You did wonderful.
Ramia Lloyd:
Thank you.
Benjamin Moses:
I do want to get back to our conversation about computing.
Stephen LaMarca:
Computers.
Benjamin Moses:
Also, we came in pretty hot this morning to talk about computers, but you found an article about NPUs.
Stephen LaMarca:
NPUs. In the last week or two, I forget when I shared that. I mean, I shared it this morning, but I shared it in industry news a week or two ago. I didn't know, when I saw this video on YouTube, which was very informative. I had no idea what an NPU was. When it comes to computers, I know the two processing units, and then the other three letter word, the PSU. Power Supply Unit, which is not a computer. I mean, maybe, there's probably a chip in there somewhere.
Elissa Davis:
Hey, what's NPU stand for?
Ramia Lloyd:
Thank you.
Stephen LaMarca:
We're getting there.
Elissa Davis:
Oh, okay.
Stephen LaMarca:
So you got the CPU, Central Processing Unit, which is your main CPU. That's the main brain of a computer. It's a very powerful processor. It has a limited number of cores, well, a smaller number of cores, than something like the GPU. But it does very complex calculations, it gets a very precise result, and it's flexible. It can do many different types of calculations.
Then you have your GPU, your Graphics Processor Unit, which has many, many cores, a whole lot of cores, and can do a bunch of very simple calculations. And really only one kind of calculation. It calculates what each individual, there is a calculation, or multiple calculations to be done, for each pixel on your screen. And it just calculates what color that pixel needs to be, if it needs to be on or off, and how bright it needs to be.
And then the new one that is being highlighted, is NPU, Neural Processor Unit, and this, we think of neural networks. And a lot of people go to AI right away. Like myself, I'm just going to speak for myself. I go right away to AI, LLMs. And it's like, well, LLMs. If I use ChatGPT, that's not being done on my computer. I'm connecting to a website, that is connecting to somebody's massive mainframe server, that's doing the LLM work for me. And then it's just spitting back, through a internet connection, what the result is.
But the NPU, the neural processor unit, which is as of right now, is not a dedicated, standalone processor in a computer. It's more common in mobile devices, even though it's making its way to desktop, laptop, workstation computers. The neural processor unit is the dedicated processor cores, specifically to do predictive text entries, and stuff like when you take a photo on your camera, or your phone's camera. And when it processes the image, after it snaps it, snaps a raw picture. You can have that raw picture if you want, but most phones today will add some processing to it, which is basically, it's the computer doing Photoshop for you real quick, just to make sure that you're getting what you want, by taking the target in your picture, putting it in perfect focus, and putting everything in the background that may be standing out over your target. Blurs it a little bit. That's done by a neural processor unit.
And the video's really cool, because it's asking the question, "Are NPUs going to be the next computer component that a lot of people buy for their computers?" And it goes into the current state, and the current state with a lot of powerful computers, and it focuses on Apple's computers. The NPU, in terms of the surface area, or the square area of the silicon die. How much space it takes up in the silicon die is actually consistent across the board, regardless of which tier processor you get.
Benjamin Moses:
Interesting.
Stephen LaMarca:
Like whatever Apple's equivalent of an Intel i5, i7, or i9, the NPU area of that die, even though the die gets larger as you pay for a bigger chip, the NPU stays constant. Because most computers, desktop computers, are using the same amount of neural processing. But where it's the biggest, where NPUs are the biggest, are in mobile devices.
Benjamin Moses:
Right.
Stephen LaMarca:
You're using the most neural processing with your photos, with your text generation, your text prediction.
Elissa Davis:
That predictive text is a lifesaver, when I'm texting one-handed. Like if I have stuff in one hand, and I'm trying to text with the other, that predictive text. Sometimes it's way off, but a lot of times I'm just like, "Oh, yes." I type like two letters, and it's like, "This is what you want." I'm like, "Yes, it's exactly what I want. Thank you."
Stephen LaMarca:
You might be a candidate for somebody that would want to invest in a larger NPU for your next phone, or computer.
Benjamin Moses:
And you did bring up a couple of interesting points where, like the importance of edge computing. Mobile devices are the true edge computing, where you're trying to figure out, should the device be computed at the point of information gathering, or at the cloud? And that gets back into our conversation, back to the computer that we're building. That scenario is like, should this... And that's a conversation we had, about, "Should we do this on the server room, or do we need a device locally?" And for us, in our scenario, it made sense to do something locally because we still want to keep it within the concept of the test bed. And then, understand, how do we scale it from here? Does scaling up mean two workstations, or does it go to a server instance, where we build something in server? Or in most cases, in the industrial case, you're paying for server time.
NPUs are a very interesting example where, you see the technology not quite being diverse, but being more specific and or narrow. So you have the CPU, you do not have a GPU which still powers the screen itself, but now you're pulling calculations away from those two devices, and having a separate device. And then, at some point in the next couple of years... And we've seen that over and over again on other devices, where it segregates, because you can control that device or manufacture that piece of equipment. But then it gets combined into a larger scenario. Where at some point, maybe it's a bigger portion die of the CPU, or it's a separate card, right?
Stephen LaMarca:
Right.
Benjamin Moses:
And it gets back to how edge computing devices are made, but I think the biggest important thing that this plays the role in, is spatial computing. When you start looking at the influence of AR applications, that is the core of why AR is important.
Stephen LaMarca:
Right.
Benjamin Moses:
And the need for NPUs would probably continue growing, based on the influence of spatial computing.
Stephen LaMarca:
Yeah, I agree. I don't think it's going to get to a point where you're going to have, for example on a desktop computer build, a dedicated card for neural processing. I think it's going to go more the way of, like NVIDIA GPUs have dedicated cores, specifically for ray tracing. And then there's the other GPU cores that do everything else, that isn't ray tracing. I think CPUs are going to go the way of... The video also mentioned that CPUs aren't really getting that, we've maxed out CPUs at 16 cores. You really don't need more than 16 cores. But what we could see in the future is a 24 core CPU, that has 16 cores dedicated for your flexible CPU usage, and then the rest are specifically for powerful NPU processing.
Benjamin Moses:
Yeah, the biggest takeaway is, understand what your workload is. That's the biggest takeaway is like, "What do you need to do?" Especially in the works. I mean, at your home or whatever, you're just going to do whatever you're going to do. But in terms of the workstation is, reverse engineer is, a reverse engineering solution. Because it is getting more and more important about, the hardware is driving your capability. And we're going to see that when we build our workstations, we're going to hit a capability limit and we're going to realize, if we need to do more, then we need to expand that capability.
I have one on automated defect recognition. It's actually back to our conversation about computing power. So this article I found. From Quality Mag, talks about quality or metrology for batteries. It was interesting where, we're not doing too much in batteries, but the battery show just exited, was a wrap a couple of weeks ago.
Elissa Davis:
There's a battery show?
Benjamin Moses:
There's a battery show.
Stephen LaMarca:
There is a battery show.
Benjamin Moses:
It's actually really fascinating because-
Stephen LaMarca:
Did you go?
Benjamin Moses:
No.
Stephen LaMarca:
Oh.
Benjamin Moses:
A lot of our automation committee members are there, because automation is a big portion of battery creation, but the battery show is interesting where it talks about the entire life cycle of a battery. So from material sciences, to manufacturing, to end usage, the entire ecosystem of batteries. Not just, it's skewed more towards mobile batteries of course. But it's a fascinating place where, on the manufacturing side, a lot of our automation companies had a presence there.
But in this case they're talking about leveraging X-rays, specifically computed tomography, and machine learning and artificial intelligence, to do automated defect detection. So previously, you would either physically test it, at the end. So this also brings up the concept of where you can inspect the end component, or the end device. Where traditional manufacturing processes you'd design, build, and then test. And then if you find something, a failure at test, you've scrapped all that value-add. Where this one, they're looking at doing stuff early in the process, where you can inject a rework cycle if you do have a failure, or you do find a defect.
So it's a very fascinating approach, about using CT scanning and X-rays. Basically, non-destructive testing, earlier in the process of your manufacturing. So I thought it was a very good article about the future of other applications, where you have a lot of subsurface assemblies. And that's the biggest takeaway for me is, subsurface assemblies, how do you know they work until you physically test it? Push that upstream, to hopefully prevent errors.
Stephen LaMarca:
I have a dumb question. Is CT scanning considered, could it be considered, optical metrology? I'm confident it's not considered, but could it be considered optical metrology?
Benjamin Moses:
We'll find out, on next episode. And the last takeaway they had was faster decision makings. Back to, "Should this part move on, or should it not?" I thought it was a good article.
Elissa Davis:
Yeah.
Benjamin Moses:
Elissa?
Elissa Davis:
Yeah.
Benjamin Moses:
I want to end on 3D printing.
Elissa Davis:
Yeah. And I'm always good for ending on 3D printing. So, Yahoo Finance put out a list of the 15 most advanced countries in 3D printing.
Benjamin Moses:
That's fascinating.
Elissa Davis:
Yes. And I was like, "Cool, and it's trustworthy, because it's from Yahoo Finance." So, there we go. So in terms of how, this is their methodology, so they ranked the most advanced countries in 3D printing technology, based on Insider Monkey scores. Which is, that's who wrote the article, is Insider Monkey. So they incorporated several weighted criteria, including startups seen in the 3D printing space, and population-adjusted patents, an H-index for research papers published related to 3D printing and additive manufacturing.
Benjamin Moses:
Aside from the name, the process is pretty robust.
Elissa Davis:
Yeah. And they're like, "At Insider Monkey, we are obsessed with the stocks that the hedge funds pile into." So I guess that's also a consideration. So number 15 is Denmark.
Benjamin Moses:
Interesting.
Elissa Davis:
Yeah, so Denmark is number 15.
Benjamin Moses:
Denmark does have a big importation scene, because of their coastal alignment, so they bring in a bunch of stuff into Denmark and then export it back out to the rest of Europe.
Elissa Davis:
Yeah, and it says, "They have a robust manufacturing industry, with its focus on precision engineering and high-quality production, has embraced 3D printing technologies to enhance production processes, reduce costs, and improve product customization."
Benjamin Moses:
Nice.
Elissa Davis:
Number 14 is Belgium. So you're going to notice some patterns, that a lot of these are European countries. Belgium. It can be largely attributed to, so their place on there, "can be largely attributed to the strong collaboration between universities, in Belgium, and their research institutions and private companies." So they're putting out, probably... Their score is probably more weighted towards the research, and the research papers that they're doing.
And number 13 is Sweden. "It consistently ranks as the EU's most innovative country, providing fertile ground for 3D printing advancements."
Benjamin Moses:
My favorite country in the EU.
Stephen LaMarca:
Which one?
Benjamin Moses:
Sweden.
Stephen LaMarca:
Oh.
Ramia Lloyd:
Swede.
Benjamin Moses:
Swedes.
Elissa Davis:
A couple that they highlighted was GKN Aerospace, and Siemens Industrial Turbomachinery. So those are a couple in there. Oh, and Sandvik, which is... I think they're a member, MT[inaudible 00:42:14]-
Ramia Lloyd:
Mm-hmm.
Elissa Davis:
Yeah.
Benjamin Moses:
Oh, yeah.
Elissa Davis:
And then, Spain is number 12. "They host world-class research facilities dedicated to 3D printing. The 3D Printing and Digital Manufacturing Center of Excellence is in Barcelona."
Benjamin Moses:
Nice.
Elissa Davis:
"So it stands out as one of the largest and most advanced 3D printing research and developments centers, globally." It's over 150,000 square feet, so that was pretty big.
11 is Switzerland. "It leads Europe in 3D printing patent applications per capita, and in economic performance. Between 2001 and 2020, Swiss companies and universities filed approximately 8% of the patent applications related to 3D printing technologies in Europe." So, it's driven by strong research institutions like EPFL, CSEM, and the iPrint Institute.
Number 10 is Japan.
Benjamin Moses:
Interesting.
Stephen LaMarca:
Wow.
Elissa Davis:
Yeah. So, "They have a long-standing reputation of precision manufacturing and quality control. It's translated well in the 3D printing sector." A couple of companies they highlighted is Kurmioto and, "IHI Aerospace partnered with 3DO, to enhance the adoption of AM technologies in the aerospace sector. Their researchers have made big strides in bioprinting," as well. And then, "It was the team at Osaka University who successfully 3D printed Wagyu beef." So, that was-
Benjamin Moses:
Delicious.
Stephen LaMarca:
Big-
Benjamin Moses:
Hopefully it drives the cost down.
Elissa Davis:
Yeah. Number nine is Italy. It's ninth on the list, and it's the most advanced country in 3D printing technology. So even though it's not number one, they're considering it the most advanced. It has 3D printing printing companies like WASP, CRP technology, which is a leader in high performance 3D printing for motor sports, and DWS Systems.
Benjamin Moses:
Are they 3D printing pepperonis?
Elissa Davis:
That'd be really cool.
Benjamin Moses:
That'd be fun.
Stephen LaMarca:
I'm still wait for a 3D printed pizza. I think you mentioned earlier, though, that WASP was 3D printing with clay?
Benjamin Moses:
Yeah.
Elissa Davis:
Yes. They focus on large scale printers, using local materials like clay.
Stephen LaMarca:
Okay. You mentioned something else, I'm sorry.
Elissa Davis:
The motor sports, the CRP technology.
Stephen LaMarca:
That's right, the-
Elissa Davis:
"A leader in high performance 3D printing for motor sports." That's also in Italy.
Number eight is China.
Benjamin Moses:
Interesting.
Elissa Davis:
So China is number eight, on this list. "Boosted 3D printing through national development goals, and the Made in China 2025 initiative, which highlights its role in industrial transformation."
Seven is Netherlands. "The Netherlands 3D printing market was valued at $13.3 billion US dollars, in 2023, and is projected to grow a compound annual growth rate of 20.9% from 2023 to 2030."
Benjamin Moses:
Hm, a lot of growth.
Elissa Davis:
Number six is actually Australia. That kind of surprised me.
Benjamin Moses:
I could see that. Those guys are wild.
Elissa Davis:
And that's, again, a large part due to research that they're doing down in Australia.
Benjamin Moses:
I wonder how connected some of the research is with some of the defense doings.
Stephen LaMarca:
Oh, yeah.
Benjamin Moses:
Because we can talk about how much money they're just pouring into defense, for the past 10 years. And it's been a lot.
Elissa Davis:
Yeah, they say they're focused on, "fused deposition modeling is the most commonly used 3D printing process in Australia, due to its simplicity and cost-effectiveness. And prices for 3D printers continue to decrease, starting at as little as $2000 Australian dollars," for a 3D printer that you can buy.
Number five is France.
Stephen LaMarca:
France.
Elissa Davis:
France.
Benjamin Moses:
Hey, we haven't talked about France, in a while. They don't come up much in manufacturing. They're pretty high up.
Elissa Davis:
Yeah.
Benjamin Moses:
That's interesting.
Elissa Davis:
Yeah, well they've got, let's see, companies like Dassault Systems, Safran and Airbus are investing heavily in AM technologies. And they're all based in France.
Stephen LaMarca:
That's right. Dassault is huge. Dassault and Airbus, that's the only reason they're up there.
Elissa Davis:
And, "the French Navy has recently validated the use of Meltio's metal 3D printing technology for manufacturing and repair process."
Stephen LaMarca:
Meltio is huge. They came out of nowhere.
Benjamin Moses:
Yeah.
Elissa Davis:
Number four is Canada. "They invested over 20.8 million in funding in 2023, to help businesses access 3D printing technology and create jobs."
Benjamin Moses:
I like the fact that, it sounds like they put it right to the end users, as opposed to earlier in the process. So, further adoption.
Elissa Davis:
Number three is Germany. So, "According to Maximize market research, the German 3D printing medical device market alone is valued at $1.24.81 million US dollars in 2023, and is expected to grow 13.9% from 2023 to 2030." So that would mean it reaches $310.4 million dollars.
Benjamin Moses:
Fascinating.
Elissa Davis:
So I guess it's growing a lot faster than we realize. They've got, "laser beam melting technology accounts for 42% of the country's 3D printed medical device market." So it seems like medical device is a very big one, in Germany.
Number two is the United Kingdom. That also kind of surprised me. "Over one third of surveyed UK professionals have adopted 3D printing technology, with two thirds believing it will revolutionize their industry. And nearly 90% see 3D printing as highly beneficial for realizing new concept and design ideas."
Benjamin Moses:
Fascinating.
Elissa Davis:
And then number one, this is going to be shocking for everybody, is the United States.
Stephen LaMarca:
Yay.
Benjamin Moses:
Yay, yay.
Ramia Lloyd:
Whoo. Be like bald eagles, like flying. "Caw, caw."
Stephen LaMarca:
USA. USA.
Ramia Lloyd:
Okay, I'm good.
Benjamin Moses:
A bald eagle, carrying a 3D printer.
Elissa Davis:
So we dominate the 3D printing market-
Ramia Lloyd:
Let's go, America.
Elissa Davis:
... with a market value of $3.1 billion.
Stephen LaMarca:
You heard?
Elissa Davis:
That represents 22% of the worldwide market share. And, "it's predicted to strengthen further, with projections indicating that the US 3D printing market will reach an estimated..." Okay, I'm not sure how this number translates. It says 33,782.4 million. So I think there's either a typo there, or it's a very large number, by 2032. And, "It's largely used across many sectors. The automotive segment dominated in 2023, with over 61% of the market share, in the United States."
Stephen LaMarca:
Holy cow.
Benjamin Moses:
Yeah.
Elissa Davis:
And in healthcare, it's used for artificial tissues, customized implants, and medical devices. And, "the aerospace and defense industry has leveraged 3D printing for complex parts and prototypes, while prototyping held a significant market position with over 54% of the market share in 2023." So, it's a lot.
Benjamin Moses:
I do like... Yeah, it's a lot. The list is very interesting, because within the US, I've been thinking about regional expertise. So if you look at, okay, besides San Francisco and the whole West Coast, if you look at who's doing interesting things in artificial intelligence. You could pick out geographical areas, that are Boston, Pittsburgh. If you look at automation, Pittsburgh is a big one.
And I think there are countries, each country may have a different expertise in terms of 3D printing. Except for the US, because it's such a big market. So if you take the US out, where Italy... Or, no. One of the previous countries is looking at using clay, the other country's looking heavy in, or Germany's doing stuff a lot to medical. And then automotive is kind of spliced in there, a little bit. But again, back to the US consumption market, it's so high. But I do think there's a lot of country specialization that the article talks about.
Stephen LaMarca:
I really like that the UK is coming in strong, and second. That actually didn't surprise me too much, having watched enough Top Gear. I know that the UK automotive market, and not just Top Gear, but watching a lot of Jay Leno's Garage. The UK is amazing at making some of the most beautiful, intricate, high performance products, but only making one or two of them.
Elissa Davis:
Yeah.
Stephen LaMarca:
Don't ask them to mass produce anything. And it's like, 3D printing is perfect for them.
Elissa Davis:
I mean, I think it also, in the UK... A lot of the European countries, I think too, while a lot of companies are based there, there's a lot of universities overseas that are doing a ton of research into 3D printing, and revolutionizing it. Especially when it comes to medical, and it comes to things that are very cutting edge, I think a lot of those universities overseas.
Stephen LaMarca:
Absolutely.
Elissa Davis:
I mean, they're doing it here in the United States too, but I think in some ways they have more funding to do it overseas, than we do here. So I think that's also, and they said that research part is part of the weighted part of it. And I think that the UK definitely, they've got the university, Oxford University, I think that does a lot of research. Things like that.
Stephen LaMarca:
Well, if you think about, if you have the funding for manufacturing, but have no infrastructure, 3D printing closes that gap.
Elissa Davis:
Yeah.
Benjamin Moses:
Yeah, yeah.
Elissa Davis:
Yeah.
Benjamin Moses:
We're seeing that quite a bit, the point of use of 3D printing, adding to that benefit. But that was a great list, Elissa, thank you. And I do really appreciate the deep dive into the countries.
Stephen LaMarca:
That was cool. Yeah.
Benjamin Moses:
Ramia, where can they find more info about us?
Ramia Lloyd:
amtonline.org/resources. Like, share, subscribe.
Stephen LaMarca:
Bing bong.
Benjamin Moses:
Bye.
