“Creating a physical representation of what used to be an idea.”

Jade McAdams
At the intersection of ideas and action, this is Longitude Sound Bytes, where we bring innovative insights from around the world directly to you.

​​I’m Jade McAdams, Longitude fellow from Rice University. In today’s episode we will be featuring highlights from a conversation I had with Don Scott, Chief Technology Officer of MarineAI and the Director of Engineering at Submergence Group LLC. He is also part of the Mayflower Autonomous Ship project, which we explored for our Imagination series to learn about the individuals involved and the experiences that turned an idea into reality.

As a mechanical engineer, I was interested to hear about Don’s experiences with the design aspects of the project and delve into the engineering mindset behind this AI ship. Being able to take an idea from the brainstorming phase all the way to a concrete product can be a difficult process, and I wanted to learn more about the hurdles that can arise and how they can be overcome.  We started our conversation with his description of the Mayflower Autonomous Ship project, otherwise known as MAS400.

.

Don Scott
Basically, it’s a vehicle of exploration and discovery. The goal is to design and develop a completely autonomous vehicle—in other words, so nobody’s on board—to head out and explore the world’s oceans, collect scientific data about the ocean to increase our knowledge and understanding of this incredible resource that we have.

Jade
So can you talk a little bit about what your role as CTO entails?

Don
Sure, well, this particular project, a lot of the work that’s involved has been managing the teams and focusing our energies on how to develop these new and innovative capabilities. And the key for us is not just developing the ideas, but developing them so that they’ll work in the real world. So deployment. In a development cycle, you have conceptualization, R&D (research and development), and then finally deployment. We follow sort of an applied research and development approach where- we call it full contact engineering, where you’re not just sitting around doing thought experiments on how you would do something, you’re actually building something and putting it out into the environment, and seeing how it works and then improving on that. That’s mainly my job. My job is basically managing all of those different projects and keeping them focused on the direction we want to move in.

Jade
So following off of that, how do you come to know everything that you need to account for when you’re designing this brand new technology, essentially? And then make it happen? Can you go a little bit more in depth on the problem-solving process?

Don
Sure. Just to step back just a little bit first, though, I mean, we’re not just a couple of guys who were sitting around our living room and said, hey, let’s build an autonomous ship, right. My partner and I, we’ve been working together for about 30 years in ocean engineering, and for the last 15 building manned and unmanned submarines for various navies. We have a lot of experience in building these kinds of things. What we’re doing for Mayflower is really just sort of a natural extension of that. So I lost your original question, but it’s just sort of giving you background about how we got into this.

Jade
Yeah, no, you’re totally fine. That’s definitely important. I was just asking a little bit more about the problem-solving process, like ideating, for this?

Don
Yeah. It’s a tough question because we have a pretty clear understanding of where we want to go. What’s missing is the roadmap on how to how to get there. And I think that’s a big stumbling block for a lot of this type of development, like new innovative capability. You have this vague, sort of nebulous idea of what you think things should be and how it should act, but you really don’t know how to get there, right? And I think that stops a lot of people, and they don’t just make that first step. Okay, let’s really try to break this down into these little steps that we can accomplish. You know, there’s a saying, basically, if you’re trying to describe a project, if you can’t break it down into steps that take four or five hours, you don’t really understand what you’re doing. It’s the same on these larger scale projects, you know, the timeline’s not four or five hours, it’s like one or two weeks, right, or a month. So I think that’s what we do. And we have this expression, Done is better than perfect, right? That idea of actually getting something out on the water that works well enough and then building on that. And that actually also creates an energy too. I’m a big fan of not just creating a story of what you’re developing, but actually building the object itself, because it creates this growing energy amongst the people that are working on it to say, Oh, look, it’s out of the water. It’s doing something, it’s not working all that great, but it’s actually out there doing something, and now I really can have a tangible understanding of what I’m doing and when I’m working.

Jade
Yeah, that was an awesome answer. You said that sometimes done is better than perfect and that sometimes it’s better to just get something out there. So what would you say has been the biggest hurdle that you’ve had to overcome to get this ship on the water? Or a second part to this question, and you can answer both, would be something that you guys have had to sacrifice in the engineering of this project that you originally would have loved to have, but at some point, it just became impractical.

Don
When I say done is better than perfect, like, you do have to meet a threshold of goodness. So we’re there. The idea of things that we had to sacrifice, I guess, in the grand scheme, especially of defense-oriented ships, we’re a small company. Like the MarineAI side, we’re only about 10 people. And the MSubs side, which is our parent company, is about 70 people. We’re no Lockheed Martin or Electric Boat, right? So for us, budget is always a concern that we have to be sort of- an expectation that it’s going to take a little more time, because a lot of the people that are working on it are also working on other projects, right. So when we started doing this, there weren’t a lot of people doing it. There’s a lot more people doing it now. And so we have had to sacrifice- speed, let’s say, we’ve been able to be fairly quick and innovative, I know we can act quicker than some of the larger companies but what we can’t do is throw massive amounts of money and resources on a problem.

Jade
Right. I mean, budgets, open schedule, always go hand in hand. Trimmed down budget. The other two also have to be sacrificed a little bit.

Don
Yeah. Our engineers have been given a fair bit of agency in order to pursue their problems, not a lot of handholding. There are extra high expectations of performance, and some people are great with that, some people aren’t. You’re asking, like, things we’ve had to sacrifice and what were the biggest hurdles? To be honest, I think some of the biggest hurdles were regulatory. This is pretty disruptive technology in a fairly conservative environment. Like, the marine domain is not really known for its acceptance of change, right. But the issue you have is you have an established set of operators, master mariners, ship captains always operate a certain way, then you have the regulatory agencies, which are overseeing. I’m not trying to dismiss them at all, because their conservatism and resistance to change is well warranted, right, because an accident at sea means loss of life and major damages. So they’re naturally reticent for this type of technology to be introduced. They have these very significant safety concerns. That said, it creates a very difficult environment in order to introduce new technology. So the onus is on us to develop trust in the systems. So we do everything we can to develop trust, and it’s a difficult relationship to navigate for sure.

Jade
You know, you’re working to build trust with these regulatory agencies, and obviously you’re putting in place a project that is very innovative and new. So how do you navigate that field? What are the things you do to share your vision with someone who may not originally appreciate what you’re trying to do?

Don
Yeah, this is a kind of a classic “you’re afraid of what you don’t understand” type of scenario. So we’re very open about what we’re doing, very transparent. You know, these are sort of keywords when you’re talking about AI-based systems: transparency, explainability, trust, all these sort of things. And also approaching it as, how does it assist the industry. Where a lot of companies make the mistake, maybe they’ve done it with autonomous cars, when they first talked about it, like, we’re gonna have completely autonomous cars, you’re gonna be able to hop into this taxi and it’s going to, without a person, gonna drive you to the airport, right? They’re not saying that anymore. Now they’re saying, we’re not gonna have driverless cars, what we’re gonna have are- we’re not calling it artificial intelligence anymore. We’re calling it augmented intelligence, right. We’re gonna do all these things that are going to enhance your ability to operate safely. So that’s sort of where we are. The purpose of Mayflower is to develop the- it’s sort of like a forcing function, to develop the technology that can be introduced and used in the maritime industry to augment their capability. We call it sharing the cognitive burden, the computer vision and the collision avoidance, we call it collision regulations, recommendations, have all of that stuff working in the background because machines are very good at doing that and providing advice to the human who is now freed up to actually do what they’re better at, which is strategic navigation, things like that.

Jade
Just to paraphrase to make sure that I understand everything you’re saying, when it comes to getting people on board for something like this, a lot of it has to do with showing how it can benefit either them or other people further down the line. Where do you see the future of this project going? And how do you see, as people become more open to the idea of artificial or augmented intelligence, how far along do you see a project like the Mayflower going? And then how do you see it leading into new initiatives?

Don
Basically, creating a platform for ocean science and research, right, having an autonomous ship that could sail around the world’s oceans and collect data, creating this autonomous capability for ships. I think Mayflower is really in an interesting position. Because I think a lot of people are,  “scared of AI,” right? Like everyone watched the Terminator, and all that sort of stuff, and it’s never good when the machines “take over.” But I think you see Mayflower and it’s sort of like, Okay, this is an AI-based system, and look, it’s actually doing something good, right? It’s actually enhancing our lives, or we hope that it does, in this way. It’s essentially- we think of it as tech for good, right? It’s providing a platform for not only actually learning about the ocean, and the need for preservation, but it’s also engaging the public as well. It’s a platform. And so I think, in terms of the future, where Mayflower is positioned, it can be used almost as a demonstration that, like, it’s not all bad. AI can be used for good things.

Jade
Yeah. Going along with that, actually, what do you think people’s general biggest fears regarding AI are? And what would you say as an engineer, someone who’s actually making this happen, to rebut that?

Don
I think a lot of people’s concept of AI is informed by hyperbole, this idea of generalized AI. Do you know what I mean when I say general AI?

Jade
No, actually.

Don
Okay, so general AI is that idea of innate intelligence, the idea of which I think of as really human capability, which is making conceptual leaps, having one set of information and being able to transfer that to another situation. But it’s about the idea that, you know, if you have an AI system used to build pencils, that AI system isn’t going to take over the world. All it’s going to do is build a really good pencil, right? Like, it’s not going to be able to take all its knowledge about building pencils and all of a sudden take over the world. That is generalized AI, right, what we have is very limited. And AI is such a difficult term too. Like, it’s basically statistics. And we’re using neural networks and things like that to do things really quickly that we couldn’t do before. Technology probably did itself a disservice by adopting this word AI. And we tend to use the term “deep learning” and “machine learning.” It seems to not have the same reaction in the general public when we use those terms because it’s like, oh, okay, it’s not this AI system that’s going to take over, you’re just using machine learning and deep learning to perform tasks that are otherwise kind of mundane.

Jade
When did you first develop your interest in this field? And how have you taken that interest and ended up on this project in the first place?

Don
So my dad was a geology professor. He was diving down to the bottom of the ocean in deep sea submersibles. That really piqued my interest in this environment, in the ocean environment and working in that space. So I pretty much worked towards forging a career in the ocean environment. I worked down at Woods Hole, at the Woods Hole Oceanographic Institute, with the deep submergence lab guys, for a year between my second and third year university. And that pretty much solidified my interest in working in this space. I’ve always been focused on the R&D side of things, working on new things. I know myself well enough that I like working on new things all the time. I get kind of bogged down and bored if we sort of spend- I’m not a production guy, right. I don’t need to see the end of a project, much to the chagrin of some of my business partners. I like white paper design where you’re gonna have a blank piece of paper or a blank whiteboard behind you, and you have a concept, you have this idea of where we want to be. I find a lot of satisfaction in making that roadmap and carving away through all the clutter and actually realizing those concepts and making them happen and creating a physical representation of what used to be an idea and seeing that come to fruition.

Jade
How has your role changed from the beginning of the project in this conception, making that first roadmap, till now?

Don
So Brett, who is my partner in this, the guy that I’ve been working with for like 30 years, he was in the city of Plymouth at a meeting about Mayflower. And they were talking about how are they going to commemorate the anniversary. And someone suggested that they make a replica ship. And he said, well, that’s not a great idea. There are tons of replica ships, right? And he said, Well, we’ll build an autonomous ship. He got out of that meeting. And his first call was to me, and he said, Don,  can we build an autonomous ship that will sail across the Atlantic? I said, sure. So that was the concept. To actually make it happen, we went through- I went through a lot of fits and starts. Back to your question of how did it happen? Or how does it happen? I think it takes a really diverse group of people. So that’s why I think Brett and I are actually a really good partnership. He’s definitely the idea person. And then I’m the one who likes to take those ideas and pursue them and make them happen. Like, basically where we’re both sort of standing on the edge of the cliff. And he says, oh, wouldn’t it be a great idea to jump? I’ll say, yeah, sure, maybe. And then he basically pushes me off the cliff and then jumps right after me. I do see, that’s how our partnership works. And Mayflower is a really prime example of that. He had this idea, you know, this concept. It was in his head, and he was able to express it quite clearly but also motivate others to do that. It’s great to have an idea, but you need to build a team to actually make it happen. You need mechanical engineers, you need electrical engineers, you need software people and stuff like that. So we were lucky in that we already had the mechanical, electrical, software people, right? We just needed to focus them on a different project.

.

Jade
To me, the most fascinating aspect of Don’s experience with the Mayflower Autonomous Ship project is that success at every phase seemed to tie back to developing a strong understanding of the end goal, in this case, creating an autonomous vehicle of exploration and discovery.

Whether he was breaking down this vast idea into feasible steps, focusing team members on a new vision, iterating on a less than perfect design attempt, or convincing others of the future benefits of augmented intelligence, taking the time to understand the unknown and share this understanding with those around him opened the doors for innovation.

As Don says, people fear what they don’t understand, and so taking a risk like that of the MAS400 requires being willing to work to discover what you don’t know, and then help educate others on the value of your ideas. As I graduate with my mechanical engineering degree, I plan to take these insights from my conversation with Don and use them to drive my own engineering design process.

We hope you enjoyed today’s segment. Please feel free to share your thoughts over social media and visit Longitude.site for the episode transcript.  Join us next time for more unique insights on Longitude Sound Bytes.