Setting Course, an ABS Podcast

The Humanoid Robotics Reimagining Shipyards with Persona AI

• American Bureau of Shipping • Season 1 • Episode 30

Share episode

Imagine a future where humanoid robots work side by side with shipbuilders, enhancing efficiency and safety in the maritime industry.

In this episode, Persona AI CEO and Co-Founder Nic Radford joined Patrick Ryan, ABS Senior Vice President and CTO, to discuss the potential for humanoid robotics in shipyards. Hosted by Brad Cox, the episode was recorded at the Ion Innovation District in Houston, TX.

Join us as we discover how advances in artificial intelligence and shifting social acceptance of robots are enabling a radical change to how ships are constructed.

Don’t forget to share this episode on social media, leave a review on your favorite podcast platform or send feedback to podcast@eagle.org. Explore how ABS is helping advance the maritime industry at www.eagle.org. 

Key Takeaways

  • Environmental challenges in shipyards hinder traditional automation.
  • The robotic hand allows for flexibility in using existing tools.
  • Humanoids can work alongside humans without needing facility redesign.
  • Automation can supplement the workforce rather than replace it.
  • Safety benefits include reducing human exposure to hazardous environments.
  • AI has made significant strides in processing and vision capabilities.

Guests

Nicolaus Radford is the Co-Founder and CEO of Persona AI, with more than 25 years of experience designing, developing, and commercializing advanced robotics technologies for space, defense, and industry. Previously, he founded Nauticus Robotics, a NASDAQ listed maritime robotics company, which developed and deployed autonomous underwater vehicles and autonomous surface ships. He also founded Jacobi Motors, an electric motor company serving traction motor markets. He also founded and serves as Managing Partner of Rad Capital Ventures, a successful hedge fund exposed to trading electricity.

Prior to founding all his technology ventures, he spent 15 years at NASA’s Johnson Space Center, where he led several of the agency’s robotics initiatives for future spaceflight missions and International Space Station experiments, including the Valkyrie and Robonaut programs. He holds B.S. and M.S. degrees in Electrical and Computer Engineering from Purdue University and is currently pursuing a Doctorate in Business Administration at the University of Houston.


Patrick Ryan is Senior Vice President and Chief Technology Officer for ABS. He is responsible for ABS’ technology research and development, digital product development, engineering software organizations, and eight Global Technology Centers. For 21 years prior to joining ABS, he was in various ship design, program management and engineering leadership roles at the largest nuclear shipyard in the U.S.

He graduated from Virginia Tech with a bachelor’s degree in aerospace and ocean engineering, and master’s in ocean engineering. He currently sits on the Texas A&M Engineering Experiment Station (TEES) Advisory Board, University of Michigan’s College of Engineering Leadership Advisory Board (LAB), as well as Virginia Tech’s Kevin T. Crofton Department of Aerospace and Ocean Engineering Advisory Board, in addition to multiple Houston-area, non-profit volunteer boards.

Brad Cox (00:07)
Welcome to Setting Course, an ABS Podcast, where we're charting the future of the marine and offshore industries. I'm your host, Brad Cox, and today we're talking about a technology that could completely change the way ships, offshore assets, and really any large structures are built. 

I'm very excited to be recording at The Ion in downtown Houston at the headquarters of Persona AI, a startup that's developing humanoid robotics for some of the world's most challenging worksites. I have a couple of guests with me who will help us understand what that's going to mean for shipyards and really the broader maritime industry. 

First, it's a pleasure to welcome Persona AI CEO and Co-Founder Nic Radford to the show. And I really want to mention some of Nic's background. He founded the maritime robotics company Nauticus and is a former NASA robotics engineer. Thanks for hosting us, Nic. 

Nic Radford (00:51)
Love being here. Thank you very much.

Brad Cox (00:53)
And it's also great to have ABS Senior Vice President and CTO Patrick Ryan with us. In a different life, Patrick spent more than two decades in various roles at the largest nuclear shipyard in the U.S. Thanks for being on the show, Patrick. 

Patrick Ryan (01:05)
Thank you, Brad. Great to be here. 

Brad Cox (01:06)
So, let's jump right in here. While industrial robotics aren't especially new, including in maritime, you don't really see them in shipyards. So, Nic, what's been the limiting factor for shipyard robotics before and what's driving that development now?

Nic Radford (01:20)
Well, I think there is a surprising amount of automation in the shipyard. In fact, that's one of the things that we got to see firsthand with HD Hyundai. They do have a judicious use of some automation, but really what is preventing or constraining that is really the complexities, the environmental challenges, the dynamic aspects of the shipyard. 

I’ve really been fortunate enough to now visit probably not nearly as much as Patrick has, but in the last 12 months for sure, I’ve spent a lot of time walking around the shipyard and it's one, a fascinating place, but two, there's a lot going on and fixed installations of the way we've traditionally done automation just really are a challenging thing. And so, this idea of having a robot encompass the same workspaces that humans are able to do now, handling the vagaries of the locomotion aspects and the challenges that comes with. I mean getting inside the ship and being able to work overhead or constrained, where there's something above you, I think really sets the stage well for the kinds of technologies we're deploying. But I think it's really just the environmental challenges that the shipyard presents that has really been the difficult aspect of putting in more traditional automation.

Brad Cox (02:36)
And Patrick, obviously you have a lot of experience in shipyards. So, what is your perspective on that? And why are we seeing robotics really kind of emerge now? 

Patrick Ryan (02:44)
Yeah, so good questions. And, you know, as Nic points out rightly, panel lines and robotic panel lines have been around for a long time. But once the structure is erected, that's where the skilled craftsman comes to play. And that's where a bulk of the welding really takes place, once the ship starts to really come together and the panel line is no longer an option. 

Why do we see it today? That’s a good question. I think a lot of that has to do with the workforce and the areas of interest, of course, are how can we supplement a workforce that is hard to keep interested or hard to attract into shipbuilding out of schools when there so many other tech options or tech-driven options today. And I think also something really important has happened in the recent couple of years from a technology perspective that makes this interesting, and I think that's really the advent of AI, which I'm sure we'll talk about in some more detail. But really using the confluence of technologies, oftentimes will really kind of accelerate things. 

When I was in the shipyard, in the shipbuilding business prior to joining ABS, a lot of my focus was on work packages and digitalization and getting mobile and it was really the confluence of CAD and mobility that made the difference, right? We had 3D before, but we didn't have a way to get it into the hands of a skilled craftsman, but phones and iPads made the difference and I see the same type of thing happening with AI and with humanoid robotics.

Brad Cox (04:01)
Yeah, so obviously part of the reason we're here today is to talk about Persona. So, Nic, can you expand a little bit on the technology that you guys are developing? I know a lot of it focuses on that dexterous hand as opposed to maybe tool attachments or something. Why go that particular route? And why is that really going to matter in that shipyard setting? 

Nic Radford (04:21)
Yeah, Persona is really fortunate enough to have just an incredible team that we've been able to put together so far.

Relatively new company for sure, but we've been punching above our weight, raised the largest amount of money in the world for a company of our stage in our pre-seed round. We've been really going quickly and hiring insane talent. So, our staff and the primary co-founders have an incredible amount of germane background and experience in developing humanoids specifically. I've got 25 years in it. My co-founder Jerry Pratt's got 30 years in it. I mean, we understand humanoid morphology and all the problems that they present exceptionally well. 

That said, we're not the first mover in this moment to commercialize embodied AI in the context of a humanoid. So, when we surveyed the industry and the markets, we thought long and hard. We talked to investors about what they saw. We talked to clients about what they saw. And it wasn't obvious to us initially what first markets were going to be the best or the first adoption places. 

The antithesis of our strategy would be the home. We are not big believers that in the short term, the home is a credible market. I think there's a lot of unresolved challenges there that it might be easy to demonstrate something in isolation, in lab setting, show them moving some laundry around, but actually attacking the business model, understanding the purchasing power of the consumer, the security, the safety implications that it's going to present. Those are huge, unresolved aspects right now. So, we did a lot of investigation of what we thought the right market to pursue and to project our capabilities into. 

And driven a lot by the business case where you rule out warehousing really quickly, you rule out large automation for shipping parcels or sorting parcels. These all have really weak ROIs. So, then you say, okay, well, where do we have an acute shortage? I mean, an absolutely massive shortage of qualified individuals that are preventing production from happening. Where do you have the purchasing power? Where do you have the willingness to adopt? And when we looked at all those factors, we found that the heavy industry was suffering. And also at the same time, not a market that traditional humanoid companies were even concerned with. 

The closest you got to that were some moving sheet metal in an automotive assembly plant, which again has, I don't know, you've had a lot of experience with this, but automotive companies tend to think in terms of decades and not days. And so commercializing that, we just didn't want — we're looking at it, but we were also very upfront with ourselves about the true timelines that could occur there. And so instead of, instead of embarking on this general purpose aspect of humanoids, we said, what if we attack a general input process and do that the best we can in a place that clearly has the need and welding came to the top of the list. 

And we were really fortunate enough to hit this market right when shipbuilding in general just became trumpeted at the White House all the way worldwide with all the companies and countries that have national and strategic initiatives behind this. So, we got really lucky from that aspect, but that's really what differentiates Persona in our pursuit is building an industrial-rated humanoid that is going to be able to handle the complexities of the shipyard, doing one of the most important input processes of welding.

Brad Cox (07:50)
You know, it's interesting that you mentioned the safety and security aspect with relation to the home, but talking about the benefits of the robotics themselves, Patrick, what are those safety and productivity benefits expected from integrating these kinds of robots in the shipyard? 

Patrick Ryan (08:05)
Yeah, well, a lot of that remains to be proved out, but that's why we're excited to be working with Persona. Listening to Nic mentioned the automotive example is kind of an interesting use case considered and put aside. It's the high variability, low production rate manufacturing that makes humanoids differentiated versus something that's production line, automotive manufacturing, that type of thing.

I would like to circle around and come back to the robotic hand, if you don't mind. And I'm going to tell a little story. I haven't told Nic this story before so I don't know maybe you'll kick me under the table for it. But when I was first driving from the ABS headquarters here to The Ion to meet with Nic for the first time back in June, I was riding with another executive who was responsible for our survey operations. 

We were talking about what we were about to see and the robotic hand and how it was all going to fit. And I observed that it's like, wow, it seems to me that you would put the torch on the robotic arm or the grinder on the robotic arm. Why the hand? And I left that meeting, three hours later, completely convinced that the hand was a huge differentiating factor that makes everything important. And the reason is, Nic, I don't want to steal any of your thunder, maybe you'll go down this pathway, but just to introduce it. 

Nic Radford (09:17)
I'm learning a lot, keep going. 

Patrick Ryan (09:20)
The idea that, for the near term, humans and humanoid robotics would be working in the same work center at the same time completely makes sense to me. And that today on a particular work package, you might have a human and tomorrow you might have a robot, and the next day you might have a human again, and the tooling and the infrastructure and the capital investment in the shop doesn't need to change by just employing a robot. The light bulb went off. I'm like, yeah, that makes a ton of sense. And the hand, the robotic hand makes that possible. And I believe in that now. 

So now how do we go and explore that and how do we use that? ABS’ perspective is, okay we have this new automation technology in the shipyards. What more can it do? What sensors are involved? What pictures can be taken? Can I measure data through the torch? Now it's a traditional torch. And how do I aggregate all that and use that for the advancement of safety? Getting back to your actual question and the application of class. That's what we're interested in, of course. From a class perspective, can we do a better job, with technology, understanding the quality of the welds or the quality of the shipbuilding process or the compliance against the rules and the alignment of the production to the approved drawings. For us, that's where the rubber meets road and having a mobile, autonomous, intelligent, robotic sensor, if you will, out in the shipyard is super, super exciting for us.

Brad Cox (10:41)
I don't know if you'd mentioned this, but my understanding is the hand stems from a NASA project, right? 

Nic Radford (10:45)
Yeah, I was really, really lucky and fortunate to be around some of the most impressive, best and brightest people, engineers, roboticists, scientists that I've ever been in. And it was an incredible way to start a career. And I was involved at the beginning when we started really looking at a robotic astronaut. 

I got to the Dexterous Robotics Lab at JSC when the current prototype had one arm and a head and was hanging from a wall and actually had a really early version of the hand that we're using now. And so this sales pitch has been beat into me for the last 25 years. You can imagine building a robotic astronaut at the home of the astronauts at JSC. And so we went in and out of popularity. But they came around to it. 

We actually ended up flying the second generation of Robonaut to the space station we called Robonaut 2. But the sales pitch for Robonaut was really simple. Qualifying space tools for the astronauts is an enormous expense. You don't want to have to then create a whole new set of tools that are robotically compatible to be doing work encompassing the same space as the astronaut corps. And it was actually a better approach to build a robotic hand capable of interfacing with the human-rated tools because it was cost prohibitive. 

And so that's what we set out to achieve. And oddly enough, a lot of these space tools look like common tools, right? If I showed you the whole catalog of space tool infrastructure, you'd look, that looks like a DeWalt drill, that looks like a hammer, that looks — right? So, they're all the same. They just, you know, can survive in vacuum and the like. And so, one of the very first publications ever written on Robonaut 2's hand was a hand designed to use tools. 

And what, again, I had a eureka light bulb moment as well when we were walking around the shipyard in Korea. And although the worker was welding, that worker had to use four other tools involved in that process. And so, you don't want to get caught up in the psychology of overfitting because that's the problem with industrial robotics is that they're really good at one specific task, right? 

Humanoids are not task optimal, but they're set optimal. So, if you want to do two or three or four things, you start to derive the anthropomorphic and anthropometric form because you're typically dealing with the same spaces and the same implements that humans use. And so, Patrick brings up an incredible point, which is we want the deployment of these robots, the job to be invariant to whether you're using a human on Tuesday or a humanoid on Wednesday. Because one of the promises is we're not going to have to make a huge capital investment to redo your facility, because you're not going to, to make it robotically compatible. So, let's make the robots compatible with the facilities. 

Brad Cox (13:26)
You talk about the human one day, humanoid the next day. How do you see the role of humans changing in shipyards or in maybe not just shipyards, but in general where you have these situations where they're having to interact with each other? And what lessons can the industry take from other industries where robotics have become more common, like you mentioned the automotive aspect? 

Nic Radford (13:46)
Well, I think we look to history to help guide us on what major technological change might do and as they transform industry. And technology tends to be inflationary to employment, oddly enough. 

If you look at one of the most unsung moments in an industry where we had automation replace people was the advent of the ATM machine in 1972 or thereabouts. You used to have to, I know you're very young, so you probably don't remember this, but you used to have to go to a bank. 

Patrick Ryan (16:36)
I remember. 

Nic Radford (16:38)
To get cash. You just couldn't walk up to, you know, you couldn't put in a little plastic thing in this little robot looking cash register machine and punch a couple numbers and get, you know, your $20 bills out. You had to go talk to somebody to do that. So, you had to go talk to a teller. But the ATM came out and displaced that whole concept. 

So, what do banks do? Banks moved tellers into higher, more abstract positions that were more profitable for the bank, like loan processing. And so, it is true that tellers per bank location reduced, but banks opened more locations because they were more profitable in general. And since 2000, the teller has outpaced the national average of employment. So just think about that for a second, right?

So, we implemented a piece of technology that was going to displace a certain area of worker, and we've actually created more need for that worker overall. It's a fascinating concept. I can't remember, I wish I had the law written or remembering it, but there's an effect, and it was actually first discovered in coal, that when you make the process so efficient, the demand for that aspect goes up. So, it doesn't remove it, it actually makes the need for more of it.

And so, it's anyone's guess how the shipyard is going to transform with this usage and deployment of humanoids across the shipyard. But, I'm not a believer in that we're going to decimate the entire workforce in some capacity. Right now, we're supplementing the workforce, right? The need is we need to add resources to the shipyard because we can't even meet the production schedules. 

One of the big light bulbs that went off with me when I was speaking with HD Hyundai about their challenges was just the open positions that they had. I don't want to give exact numbers, but let's just say a lot. And they're struggling to meet production schedules because they can't get the right folks. The training and the turnover in the shipyard is a big deal. And they were looking for automation to help fortify that. So, I think there's going to be some mobility between jobs, but history has proven that those jobs actually become more desirable. They go into higher cognitive functions at the human level and then the operations become more profitable in general. So, I think it's going to be great for all involved. 

Brad Cox (16:26)
We've had several episodes talking about that interaction between human and robotic (autonomous). Patrick, I'd like your perspective on that and, from the class perspective, of that human oversight and how that factors into the puzzle of making these things a success. 

Patrick Ryan (16:40)
Yeah, well, I mentioned a few advantages to understanding the safety of the ship relative to quality and filler metal and those types of things. But there's also a big safety benefit to the shipyard worker. Will the robot be the first to go into a confined space to do hot work rather than a human? I think that's a huge advantage. 

When you think about what shipbuilders need to do every day, it's hard, long, sometimes outdoors in heat or in cold and weather, where the robot will be able to function in those environments without having detrimental effects. And not to mention the length of the workday can be transformed by a robot. Can a robot work more hours per day than a human can? Obviously we've got battery things and Nic will solve those problems for the industry.

Nic Radford (17:26)
Oh, we're not going to solve them, we're just going to plug it in.

Patrick Ryan (17:29)
But I think there's a lot of advantages to just production rate. Because right now, slips are full across the globe, and orderbooks are full, and there's capacity problems. Not to mention, if you start thinking from a domestic U.S. perspective of what politically and from a kind of a national infrastructure, national security position America wants to take to advance and grow shipbuilding, technology is a natural fit for the U.S. and I think Korea also. 

A lot of great opportunities to really kind of supplement the workforce, make the shipbuilder safer, put them in less risky positions and let the robot take some of those tasks on. We think and we'll be able to demonstrate and record that the quality of the vessel and the quality of the construction goes up as well. So, lots of positive benefits. 

Brad Cox (18:15)
Patrick, you mentioned the AI boom earlier, and I do want to kind of revisit that. So, Nic, we talked about this a little bit before recording. So, how has that growth in AI in really the last three, four years been a driver for all of this? 

Nic Radford (18:30)
Well, it's multi-stage. So, AI has been around a long time. I can't remember the very first paper written on it, but let's just call it like 1960 something. So, theorizing on machines that have elements of thinking and capacities of how humans think. And then you go 70s and 80s and 90s and we started getting into the neural networking aspects and trying to process data similar to how a human mind processes data. That's why we call it neural nets, right? And convolutional neural nets. And then that gave way a little bit, and more recently, to more abstract versions of it, that are a little bit more black box-ish. What has enabled all that? Processing. 

So, the human mind is pretty fascinating. It's actually slower clock rates than, say, a typical computer is, but it's massively parallel. It does amazing things when it comes to imagery and computers have been typically bad at that. But more recently, we have basically solved vision processing. Even five, seven years ago, if you were to take a camera into a room and try to identify the objects and try to tell the robot how it might approach an object or grasp an object, it would take seconds, I mean, absolutely seconds to process that. I mean, robot videos and demos back in the day were so painful to watch because it was like watching paint dry on a wall. Robot staring at a door handle wondering if it's a butterfly or a door handle, right? It's just, it is so hard. Now you take a camera into a room, you can identify all the objects, where they're at, what they might be, trajectories unto approach and grasp, and you're doing it at frame rate. 

So, once you have that front end figured out, and now you couple that to the LLM aspects of what the modern AI revolution has been about, and now you're able to orchestrate behaviors that look to be pretty robust to uncertainty in the environment. We've developed training techniques of how to teach the robots and replicate human behavior through demonstration learning or human behavior cloning. We're wrapping that whole concept in reinforcement learning techniques, which again, motivated by the fact that we can process so well. GPUs have become so much more affordable than they used to be and you're getting AI chips that are just tailored to processing the algorithms effectively. And now it's — now you're having a Cambrian explosion of capability. Every time I turn around there's a new humanoid company that's raised a quadrillion dollars and going after what we think is the largest TAM of our lifetime.

But I would also mention, if you permit me, there’s two other components to this besides the technology that's really moving this. Markets move with really three forces: technology component, an economic component, and a social component. And we've had all three of them tip at the same time. What I mentioned with technology, processing vision, algorithms have gotten better, et cetera. But socially it's become much more acceptable.

I was really fortunate at NASA, when I was working with the Robonaut 2 team and I was the chief engineer then became number two civil servant on the program, General Motors came to the facility and we co-developed Robonaut 2 with General Motors. And we weren't allowed to talk about it with anyone. It was super secretive. We were hidden away in some facility at NASA because they were freaked out if anyone were to find out that General Motors — this is my opinion — if General Motors were working with NASA to build a humanoid. What would that mean? All this kind of stuff. 

Now, contrast that with today that you have humanoid companies and car companies, logos on the same video on YouTube, walking through the facility, moving sheet metal and punching buttons to make machines go down. It's incredible. Like what a social shift has occurred. And now socially acceptable. We're talking about robots in the home. We're talking about robots being out among us every day, right? You can walk down the street and see a self-driving car now, right? So, we've really become socially accepted to that. And so that's huge. 

But economically is the other — none of this matters if we can't do all these things at a price point. And processing, actuation, integration, bill of materials that aren't, you know, three or four million dollars. When we first built Robonaut 2, it was a multi-million-dollar machine. Now we're building these machines for a couple-hundred-thousand-dollar bill of  material and then production quantity is going to get to 50 grand and lower. So now you've had those three aspects are why you're seeing an explosion happen right now.

Brad Cox (22:52)
On that social point, we kind of talk in some ABS publications, we've talked about like trustworthy AI. Patrick, what's still needed to get this technology to a point where it would be accepted by the industry? 

Patrick Ryan (23:05)
Well, I think social acceptance has come a long way. This is one of the hottest topics that rings my phone every day already, and just within a few months. When we talk to other CTOs, this always comes up at the end. It used to be it was AI. Now AI is so yesterday and robotics is today or tomorrow. 

Nic Radford (23:23)
AI was so last summer. 

Patrick Ryan (23:26)
So, it's moving fast, super fast. Kind of getting back to how do we ensure that this all comes together? Well, that's the point of good R&D and good testing and good partnerships. 

We always say that ABS success is a team sport. ABS isn't going to build a robot. But could ABS use robots ourselves for survey purposes? I think so. Would we use other people's robots to collect information as an extension of the survey function and collect data? Yes, I think that's probably the bigger mover. And then can we be confident that robots performing structural work on a vessel in the yard will make that vessel safe? That's probably going to be first up for us because that's kind of the most important to ABS is to ensure vessel compliance with rules and ensure that process goes smoothly. 

That will be kind of the first work that we go and do together is how confident are we? What sensors are needed? How do we process the data? How do we collect the data? How do we ensure the data is cyber secure? What do we do with it and dispose of it once we're done making those assessments? And then how do you manage a fleet of these technologies operating at the same time on the same unit and kind of manage and understand all of that? 

For us, lots of moving parts that we really need to get our arms around. But that's why I'm excited to get up to go to work every morning, and that's why we've got great people working on this, and I am super optimistic for the future. 

Brad Cox (24:52)
So, Nic, you kind of alluded to it but humanoid robots are making headlines every day now. What is Persona's trajectory and y'all's near-term plans? And where do you see these things going in 5, 10, 20 years? 

Nic Radford (25:06)
Well, 20 years, that's impossible. That's like singularity type stuff. No one can see past that. We have a very ambitious program that we kicked off in January. So, we're 11 months into it. And we've got a 24-month timeline to deliver a welding humanoid robot to HD Hyundai in November of 2026.

We're building three generational prototypes, so Gen 1, Gen 2, Gen 3, every six months. We're rolling the revision. We got our first one done. It's walking around the lab. We've been a little quiet on YouTube because there's a lot of incredible companies out there producing and publishing videos that are absolutely fantastic. And although those companies went through their early learnings and it was a little bit more acceptable to publish, hey, this is what we did this week in Robot 101.

We're just playing it a little closer to the chest because we've got some really capable machines that are out there, highlighting to people what's possible. And so, because we're a relatively new company, we've been a little quieter about that, but rest assured our first robot is walking around. Our second one's in design. That one's going to come out next year around May. And then at the same time, we'll start designing the third gen based on the learnings of each of the first two and we'll deliver that to HD and we'll call that a pre-production prototype.

And then we'll go through another iteration as we're scaling towards manufacturability. Just hired a gentleman from Amazon. He spent seven years at Amazon scaling Amazon's production after the Kiva acquisition. So, he's laying out our entire production strategy. I was just reviewing plans with him this morning. He was telling me how much warehouse space he needs around here and I got to go find that. 

Your question is like, what's the short term, the medium term, long term look like? Well, short term is let's make sure that we make HD extremely happy with what we're doing. And not to mention we have several other partnerships that are in formation and signing. And again, as we've been talking about here, we're really honored to be partnered next to ABS and ingest all of their incredible historical aspects of the industry because we're newcomers a little bit. I've got some experience in marine, did some shipbuilding, got on some autonomous ships back in the day, but nothing at this scale. And so we're going to have to learn a lot a bit about that. 

So short term, let's just do our jobs. Medium term, how are we going to make more of them? And the longer term, you tell me, what do you think's going to happen? 

Brad Cox (27:15)
I don't know, Patrick, what's the future of shipyards look like?

Patrick Ryan (27:19)
Well, I think more automation. Let me maybe take a step back and zoom out. Do I think it will be all robots everywhere? Probably not in my lifetime. But will there be lots of robots? Yes, there will. 

And is the advances in computing and AI that are discovered by Persona in the application of, the development of the robot and use in shipbuilding applicable to other things like augmented reality or augmented humans wearables? I think all those things are going to all overlap each other. 

As I was listening to Nic talk about the camera walking in the room and identifying things and that was a big part of my life in developing augmented reality and simultaneous localization of mapping to put information in space for a human to consume and do those things build upon each other? I think they're going to because the convergence of those different type of technologies do overlap. And with the approach that Persona is taking by not recapitalizing the yard, but putting robots working together with humans, I think that's a great model. And that's why I'm excited about it. 

Will that allow us to build different types of ships, design them faster? Now you're talking about the digital thread and kind of the collection of 3D data and will it allow shipyards to design more quickly and turn around more quickly because the work package development isn't about production of 2D drawings and stapling them in a book and procedures and training to humans. All that stuff can be bundled and leveraged for a robot much more quickly than a human. 

So can you retool, I guess is the word that I'm looking for, to start producing different types of vessels in the same types of shops and slips that typically would be earmarked for one type of thing. And then again, in the U.S., does that mean military and commercial vessels can be built at the same place? And is this the technology that kind of solves that difficult challenge that has been traditional in our industry as a difficult thing to do for a long time? You know, maybe it is. So I’m, again, bursting with optimism. Super excited about where this is going and certainly there will be changes but changes are constant and change is good and all this is advancing us to being able to produce the kind of ships in a safe way for less cost that really will help the industry. 

Brad Cox (29:33)
Okay, cool. So, up against our time here, I wanted to give each of you a chance to maybe leave a closing thought for our listeners. A key takeaway or something you want people out there to know. So, Nic, you want to start with that one? 

Nic Radford (29:44)
Sure, I think we could take a look to history to help guide how we should think about things in the future. I think it was Gates that said people tend to underestimate the impact of technology in the long term and overestimate its impact in the short term. I don't think we're going to get away from that. I think we're all talking about how amazing all this is going to be, but let's not forget self-driving cars. We've been working on that for 15 years.

It took the automotive industry to transition fully from the horse and buggy in the early 1900s, it took them about 25 years to do that. Yes, arguably things are happening at a much quicker pace now, but it isn't going to happen overnight. And so what is top of mind for us are working with the people and partners that have a long-range view of where this is heading. Capital providers that have the patience to understand the transformation, the enormity of what's going on. 

But it would do us all well to keep in mind that although it sounds incredible, that there's going to be a few bumps along the way and that this is a long journey we're on. 

Brad Cox (30:41)
And Patrick, any closing thoughts for our listeners? 

Patrick Ryan (30:43)
I'll give a hear-hear to what I just heard. And maybe just one amplifying piece of that is it does take a lot of moving parts and it does take a team. And so, ABS is excited to work with Persona. Happy that Persona has already got shipyard clients lined up because without that in-yard experience, then it's all just a science project, right? We want to really make a difference. And I think bringing all those different people together to make the technology real is really important. And I'm happy with where this is going so far. And I think Nic would agree. We've got great people working on this hard problem. It's not just about technology. 

Brad Cox (31:21)
OK, great. Well this has been really fantastic. It was a great deep dive into robotics and AI and shipyards and what that can mean. So, really pleased we got the opportunity to come down here and visit The Ion and Persona AI. And Nic, thanks for hosting us. And Patrick, thanks for being on the episode. 

Patrick Ryan (31:37)
Yeah. Brad, thanks for prompting us. And I look forward to doing this again. 

Nic Radford (31:41)
Yeah. Prompted is just like a ChatGPT window. 

Brad Cox (31:45)
Yeah. You know, at work, they call me BradGPT.

And for everybody listening, thank you for joining us for another episode of Setting Course. Be sure to subscribe, leave a review, and share this episode. To keep up with the latest technologies reshaping the maritime industry, visit us at www.eagle.org. Thank you for listening.