Sea Control - Sea Control 372 - The Problem with Autonomous Ships

Episode Date: August 18, 2022

Links:1.  Feedback Loops and Fundamental Flaws in Autonomous Warships, by Jonathan Panter and Johnathan Falcone, War on the Rocks, June 24, 2022. 2. The Unplanned Costs of an Unmanned Fleet, by Jonathan Panter and Johnathan Falcone, War on the Rocks, December 28, 2021. 3. RIMPAC Testing Will Inform the Fate of Medium Unmanned Surface Vessel, by Mallory Shelbourne, USNI News, August 1, 2022.4. Sea Control 301 - Task Force 59 with John "Fozzie" Miller & Ari Cicurel, by Jared Samuelson, CIMSEC, December 12, 2021. 

Hey folks, it's Jared. My guests today are Jonathan Falcone and Jonathan Panter, and we'll be discussing their War on the Rocks article on autonomous warships. If you haven't been following along, the U.S. Navy tested some of its autonomous vessels as part of the recently concluded Rim of the Pacific exercise in Hawaii, so this is particularly pertinent. This episode was edited and produced by Nathan Miller. I'd like to pause here to highlight our local chapters, whether you're in South Korea, Egypt, Singapore, France, New York, India, or the Caribbean, chances are there's a SimSec local chapter near you. You can find a full listing of local chapters and contact information on our website at simsec.org,

so if you're interested, please reach out and get involved. Finally, I want to take the opportunity to recommend our partners in the SimSec Podcast Network, The Bilge Pumps. You can find Alex, Jamie, Drack, and a pot of iron brew bottles wherever you download your podcasts. With that, Kimber's Men. You're listening to Sea Control, hosted by the Center for International Maritime aloha shipmates and welcome back aboard sea control my guests today are jonathan panter and jonathan falcone we'll be discussing their article from war on the rocks entitled feedback loops and fundamental flaws in autonomous warships the gentlemen welcome jonathan panter

would you start by introducing yourself to the audience please sure and thanks jared for inviting me to the podcast. I'm currently a PhD candidate at Columbia University studying political science. My research focuses broadly on issues of command and control in military operations and civil military relations. Specifically, I'm looking right now at how high levels of administrative control in the U.S. Navy, often those that have been imposed as a result of accidents, collisions, or liberty incidents abroad, can have adverse impacts on the Navy's long-term goals of crisis management or combat operations, in other words, doing the nation's business. And prior to coming to Columbia, I served as a surface warfare officer in the U.S.

Navy, and I graduated from Cornell University in 2012. Thanks. And Jonathan Falcone, how about you? Yeah, so Jared, again, thanks for having us on the show. It's an awesome chance for us to kind of dig in on something that Jonathan and I are both super excited about, is like our Navy's future. Currently serving as chief engineer on board USS Manchester, and then kind of a winding road to get here. So I graduated Yale in 2011, worked on Wall Street as an investment banker for a little bit. Then I went to OCS, commissioned, served on board the Gonzales and John S. McCain. Then for my shore tour, I was picked up for the Fleet Scholar Program. I went to Princeton where I concentrated on science and technology policy and international affairs, and now I'm here,

beautiful San Diego. Great, and as a reminder to the listeners, all opinions are our own and not reflective of any institution with which we might be otherwise associated, so I'm going to go immediately off script, and are you two now going to brand yourself as the Jonathans and attempt to rise to overthrow the Bryans, McGrath and Clark, as the rising generation of navalists? That sounds like a fun plan. I think Brian McGrath was one of our first editors on our initial War on the Rocks article. So, yeah, you know what? Why not? We'll throw them up there. Excellent. OK, this is the second article that you two have authored on uncrewed ships. So what got you interested in this topic?

So for me, I'm just generally interested, intrigued by innovation. I love reading Wired and The Drive, and kind of one of my favorite things about that is a lot of the writers themselves, they're intelligent, and they have some type of experience and vocabulary with some of these really super high-tech topics, but they're not technologists, they're not computer scientists themselves. So I really admire that type of writing and reporting. But in addition to that, just as a naval officer and as an engineer, uncrewed ships are fascinating. First, because philosophically, you know, what does leadership look like in a hybrid setting? How do you lead something that's unmanned? And how does it impact your sailors on board your ship? You know, what type of biases, heuristics, what are we trying to overcome? And then second, kind of on the nerdier side, as I mentioned, being the Chang on an LCS, you know, we're minimally manned. We're that kind of first step, I would say, towards uncrewed vessels. And I constantly see how data monitoring and data limitations impact our ability to most efficiently run our engineering plants. So those are the limitations I'm seeing. And I'm curious on how we're overcoming that in the USB world.

And then besides that, it just gives Jonathan and I have been friends for about a decade. So this is a good opportunity for us to kind of take those conversations over whiskey and put it on paper. And yeah, I have to immediately lift off where John left off, which is that it's just fun to geek out about these things. sometimes if you're in this profession of arms or if you're adjacent to it, like I am now as a former Naval officer, but it's even more gratifying to do that with a friend. So what got me interested in these topics originally is, as I said, in grad school, I've been looking at issues of command and control. You might say, well, you're studying political science. What does command and control have to do with that? That really sounds like military science, but it's actually a really

intensely political issue because where we locate authority in a chain of command can have political consequences down the line. A great example of this is in nuclear deterrence, where we want to ensure in order to deter the adversary that the adversary understands that if he attacks us first or attempts a decapitating strike, we're still able to respond, that we have a secure second strike capacity. And one of the ways that is done is by devolving authority to lower levels of chain of command such that if a decapitation occurs, you can still have a response. Now, the problem with that is the lower you place the authority to, let's say, release nuclear weapons in the chain of command, the more potential adverse consequences you have in terms of accidents or

in terms of unwanted escalation. So these are intensely political questions where we locate authority in a military organization. And naturally, people are the most essential piece of this puzzle. This is a story about people. And I think discussions about automation often treat people in a bit of a disparaging manner. There are some aphorisms you might have heard in discussions about automation. One of them is machines never get tired or quote unquote machines don't make mistakes or machines don't get hungry. Perhaps that's all true. And there's a lot we can learn from that because there are advantages of automation. But on the flip side of that equation, so far machines can't put their heads together. They can't innovate. They can't

encourage each other to exceed their limits. They can't dream big. And there are many things that people with the current state of technology that people still bring to military operations. And so I'm interested in dynamic environments in crises or in combat or in international signaling. What happens when you start to remove people from those situations? And the last point I put on all this about my interest in these subjects is that I'd like to make clear, because this is a very technical article, that I'm not an expert in industrial control systems. And I had the opportunity to workshop this article at the Naval War College at the Cyber and Innovation Policy Institute a few weeks back, I realized that I knew the least of anybody in the room. And that's

actually been a great discovery because there are entire fields out there of research and development that John and I are both learning about through writing this article. And that's been very gratifying. So what happened to IBM's Mayflower Autonomous Ship and why did you two opt to open your article with that vignette? IBM's, basically, the Mayflower is an autonomous vessel that has been developed by a consortium of independent researchers with some help from IBM. And it recently conducted a transatlantic crossing to Greyfant there. And it was an enormous success, particularly the software, the technological, the AI components of that whole project. But there were a couple hiccups along the way. And what's interesting is the hiccups themselves

were actually mechanical. They weren't in the software and the AI systems. So two things happened. First of all, in its first attempted crossing, which was last year, there was a diminishment in the power available below parameters. And so the vessel communicated back to shore with its engineer handlers. They determined that they had to return it to shore and they brought it back and conducted repairs. This year, just a couple of months ago, they reattempted that crossing. And again, you had a set of mechanical failures. The first one was an isolation switch for a generator. It failed. And then they sent the vessel back to a port. They repaired it again. A few weeks later, they made another attempted crossing, and there was

a casualty to the charging circuit for the generator starter batteries. Once again, same deal, communication off-ship, return to port, repairs completed by human engineers. Now, the purpose of this anecdote is to show that while mechanical kinks are being ironed out as we introduce new technologies. Human intervention is still often needed, and the autonomous vessel to get that human intervention has to, so to speak, call home. That's totally fine in a research project, and it's certainly not a knock on the extraordinary work that the Mayflower engineers have completed. But in a Navy that faces an adversary, this can be a very deadly phenomenon because the adversary can track those electromagnetic emissions, those off-ship

communications. Another good example, if we bring it to the military world, is the Sail Drone, which is currently a small autonomous platform that's being tested in Fifth Fleet. The platform itself actually has already been purchased by NOAA, the National Oceanic Atmospheric Agency, to collect weather data and do things that NOAA does. And it's performing very admirably in that role. But again, that's an unopposed mission. NOAA doesn't have to worry about adversarial targeting and collection of emissions of the sail drone. So the overall point here is that everything from software bugs to mechanical failures to communication is a whole different ballgame in the military world than in the civilian one. And what we're trying to get

across is that we're not anti-automation whatsoever. We're anti-hype. We believe that these technologies do represent the future, but what's important is integrating them into a military context judiciously. So what are the Navy's current lines of effort for USV development? How do these systems interact with one another? Yeah, so I'm going to, I'll jump on that question. So I kind of want to help frame both the conversation for us here and then also for our listeners. Start with kind of defining kind of these terms a little bit more broadly. That way we're all stepping off from the same stone. And then we'll move into the Navy's high-level efforts and then kind of at the more system level.

First, like we've already been throwing out the words unmanned, autonomous, and everybody kind of has different ideas in their head of what this could mean. In the Navy sense, right, what kind of USV div who is kind of the new squadron that's been stood up and taking responsibility of these, the word unmanned systems is simply to suggest that there's no human on the platform itself, meaning that control of an unmanned system can be remote. It can be supervised by humans, or the machine can be making near-autonomous decisions. Kind of maybe change the Terminator idea of unmanned systems and grounding it a little bit more.

So now, autonomous vessels. So when we use the term autonomous vessels, what we're saying is it's a vessel that can operate by itself, by sensing its environment and performing the necessary functions it needs to do to execute what the human determined mission is all without human intervention. Autonomous vessel overall is an aspiration, right? And that's what we're kind of working to, right? So with those two kind of definitions in mind, kind of the simple way to put it is autonomy, unmanned systems really exist on a spectrum from remote operations to near autonomous operations. So more specific to the Navy's current efforts, I think it's helpful to kind of break it up between what's happening

in the Pacific, what's happening in the 5th Fleet. So in the Pacific, you just saw on Twitter, USNI articles, we have four USVs headed to participate in RIMPAC, which is so awesome to see. You have Seahawk, Sea Hunter, Nomad, and Ranger. Seahawk and Sea Hunter, they're medium displacement USVs, and Nomad and Ranger are classified as OUSVs, or Optionally Unmanned Surface Vessels. So that's out in the kind of the Pacific theater. In Fifth Fleet, you have the much smaller USVs that Jonathan alluded to, including Sail Drone. The other one that you might have seen in recent articles on is Sail Drone and Mantis, whether it's their 24-foot, so T-24, or their 38-foot versions T-38. So obviously all this stuff is going to be in flux as program of records actually

get funded, et cetera. But right now, the larger, the MUSVs, LUSV concepts are residing in the Pacific and then some of the smaller sensor platforms, sail drone Mantis operating in fifth fleet. The Navy, in an effort to iterate and improve upon these prototypes, they're focusing on a couple of major enabling lines of effort. Above all, kind of, well, maybe not above all, but reliable whole mechanical and electrical systems, sensory perception and autonomy, and payload packages. If you actually look at some of the documentation coming out of PMS 406 prior to this year, reliable whole mechanical and electrical systems didn't actually make any appearances in their

kind of initial slide decks, right? That's a new feature of a lot of their materials. So I think, you know, where our article comes in and kind of the control systems that Jonathan discussed that the CIPI at the Naval War College really is starting to ramp up in terms of visibility. Driving kind of more at the heart of your question, right? How are these systems interacting with each other? we'll just look at these enabling lines of effort. So, reliable hull mechanical electrical systems, sensory perception and autonomy, and the payloads. So, simple take, reliable hull mechanical electrical systems provides the stable propulsion, power, and cooling that's needed to operate the sensors, support autonomous processes, and to deliver payloads. So, whether that's just getting to the location to perform a mission or powering fire control, launch systems, radars, whatever

the sensors might be, reliable home mechanical electrical systems is critical to that. Sensory perception will provide kind of the control systems with the necessary data to manage these home mechanical systems and monitor them so that either engineers on the shore side or stationed on a DDG or some type of logistics platform is able to observe what's going on onboard these plants. And at the same time for payloads, sensory systems will provide the data that is necessary for fire control or for the towed sensors to perform its operations. So you can see very much, right? These things start to feed each other. And then when I was kind of thinking about it, you know, payloads kind of just like in the manned world, they're more just

customer consumers. It makes a lot of demands on the rest of the ship rather than providing anything to anybody. That is a Chang's perspective. I just want to say for all the operations and combat systems officers out there, that is very specifically a chief engineer's perspective of that. But please go ahead. No, that's why I have 51% of the vote, as I like to say. So like I said, very much so. You kind of get that cyclical relationship, the engineering systems providing what the combat systems sensory perceptions needs. At the same time, those monitoring functions provide the necessary information and data to make sure that the plant is operating appropriately. Sorry for the long-winded answer on that one, but I hope that kind of like frames out the discussion,

both from how we're using the terms autonomous and unmanned, some of kind of what the big headline programs that the navies are, and then kind of those specific enabling lines of effort. Now, that's a great argument for the listeners. Be sure you attend PV4M and give valuable inputs there so that the engineers know what's going on. I will say for the listeners, we'll have links to the USV's head of RIMPAC. And then if you want to hear more about CTF-59, we did interview Vice Admiral Fozzie Miller, retired Fifth Fleet Commander in Sea Control 301.

So if you want to find out what Task Force 59 is up to or sort of the thinking behind it, he had a lot of interesting thoughts in that episode. How does the interconnectivity of these systems make uncrewed vessels more vulnerable? I think this kind of gets at one of our primary objectives when we set out to write this article. And something that we were wrestling with, continued to wrestle with during the writing, and I'm sure we'll continue to wrestle with. So the maritime environment, inherently dynamic. And I think it's safe to say that mechanical systems are going to degrade as a function of time at sea, right? And because crews are absent to break any chains of cascading failures, the systems on board across kind of those enabling lines that I mentioned before are reliant upon each other.

So this kind of creates inevitable vulnerabilities that just cannot be eliminated. And at the same time, there is a reduction in our ability to mitigate it because of the lack of people. I think an example is kind of helpful to illustrate this point. So with a focus on those medium displacement USBs like Seahawk and Seahunter, its power is provided by a diesel generator. Two of them are on board. Only one is required for ship operations, but those engines don't have any remote restart ability. So if you lose that engine, that's it. You need to get somebody on board that vessel in order to restart it. These generators provide the stable power that's necessary to run the sensors, power the cooling to the spaces that house what I imagine are incredibly high heat generating equipment, communications gear, etc. At the same time, whether you are operating in remote mode or conducting autonomous operations, these generator outputs are also providing power to the control systems that are managing the distribution of power, monitoring equipment for any alarm states.

now like as a caveat all right i'm not privy to exactly how these these plants operate right and i am sure there are plenty of redundancies built in uh such as uninterrupted power supply backed up batteries etc but anytime you even have a shift in power there are potentials there are there's potential for failure there's there's potential for things to drop off And you start to see this picture about how there's like almost a bit of circularity that the engineering equipment is responsible for powering the very control and monitoring systems necessary to ensure the proper functioning of these systems, all happening in the absence of a human watch stander. As kind of Jonathan mentioned, also mentioned earlier, we're not anti-USB at all, right? We want to make that very clear. I think one of the things in the article that we wish we had kind of more space and time to get to was, what do we do about these vulnerabilities? How do we improve upon the current system? And how do we get to a place where we can successfully operate these vessels?

And really one of the glaring things to me, as somebody who operates on a minimally manned vessel right now, is just our data limitations. Whether it is the data rate, the rate at which information is collected off of these systems, whether it's storage, or whether it's simply identifying what parameters are most important for us to be logging, all of this seems like the opportunity for us to create more resilient unmanned systems. And as anybody with kind of even just like a little bit of education and machine learning and autonomy will tell you, data is the lifeblood of these systems. The more data you have to train them, the more data you collect, the more opportunities it has to operate. That's where is going to become more of an efficient system. And so speaking more directly to what we kind of see in the fleet right now versus where we need to get to with some of these vessels is data

collection can't just be status and analog points. One of the challenges is that we don't really collect in situ trend data. As a result, we're kind of looking at focus points in time. Some of reasons that we are challenged by that is simply you can't just have a ship that's full of data storage, full of servers, et cetera. Overcoming this data collection problem is going to be key to us solving kind of what USVs are capable of. And then kind of like the last point I'll leave off of is very much with data, we are looking at system set limits, high and low alarms, et cetera. and kind of where the in-situ trend data is going to be helpful is, okay, something is still operating within parameters, but we're getting a negative trend. And so for instance, we were

recently operating a hydraulic system and one of the circulating pumps dropped offline, right? And so as a result, we saw the hydraulic oil slowly climbing. The person who was in control, the engineering duty officer noticed that this was climbing up the seawater temperature still within parameters no alarms whatsoever went back kind of send the watch standard down to the space noticed that that circulating pump had tripped offline right we were able to get that pump back online and control the rise of that hydraulic oil right now as designed without hitting any of those limits right you're not necessarily going to get the watch standard response and the watch and watchstander reactions that is going to be necessary to keep these vessels safe and

operating and on mission. So when we're talking about the disaggregation of capabilities, is that a U.S. phenomenon in that way of thinking about the disaggregation of capabilities? And is it not something that we can train out of? Because I would argue with you that our NATO allies, particularly those in the North, if you look at what the Danes used to operate, what the Norwegians still operate, the Swedes operate to a certain extent. I get that they're not our NATO allies yet. What the Germans used to operate, they used to operate vessels who were primarily focused on conducting anti-service warfare that by and large would not sense the target themselves, but they would go to a certain location, they would fire where they were told to fire, and then they would

scamper back to their fjords. So is it maybe that we need to approach the unmanned vessels and the command and control over them as a single system and that command and control piece, that person who's telling them to fire wherever that individual happens to be sitting, that should be viewed as one system rather than viewing it as separate capabilities and not, you know, just because they're not co-located, they should be viewed differently. Well, I see that as two questions. One is, is this oncoming disaggregation of capabilities that John and I mentioned in the article, is that a uniquely U.S. phenomenon? So that's the first question. I'm not really familiar enough with foreign navies to say definitively one way or the other. But what

I would imagine is that the problem sets that a lot of these partner navies that you mentioned face are very different than the U.S. is because they're not operating, they're not really navies that are designed to operate forward at great distance from the homeland. And they're not optimized, therefore, for power projection. So the U.S. fleet, which is optimized for power projection, has at its core some high value units, carriers, and those require a great deal of multi-mission platforms arrayed around them to protect the carrier from any number of surface threats. And that is probably what has contributed to the aggregation of capabilities onto given platforms. I assume that the Chinese Navy, especially with, as we all know from this week,

the most recent launch of its now third carrier, is going to be structured on similar grounds. The second part of the question, just because unique capabilities are disaggregated on separate units, if they have unified command and control, why are we considering those disaggregated capabilities? If it's being controlled, let's say, from a mock or something shoreside, one can even think about how German submarines in the Second World War with Wolfpack attacks were coordinated from the shoreside. It's an interesting question. I think from a systems perspective, we can definitely see those as unified capabilities. One of the things I would caution, and this brings us back to our article, has to do with the interruption of communications, which we believe

to be a fundamental problem here. So if we're correct that you have these feedback loops between your underlying engineering systems providing the basic enabling functions, such as electrical power and cooling, and their ability to communicate off-ship, and we start to see those failures, then perhaps even if you're right, Jared, that this can be understood to be a unified system, it rapidly becomes a disaggregated system once those problems come to the fore. So what do you two see as the future role for uncrewed assets and what's the potential impact to the force? As we spoke about earlier, if the Navy's plan doesn't change, if what we see in the current planning documents comes to fruition, then the Navy intends to send unmanned vessels or apply unmanned vessels to the so-called, quote unquote, dull, dirty and dangerous problems. And those, as currently outlined, are things like mine laying, oceanography, sonar surveillance, bathymetry, surface intelligence surveillance and reconnaissance, remote sensing, and perhaps adjunct missile magazines.

As far as the potential impact of putting USVs in all those roles, it can really fall on a spectrum. On the perfect side of the spectrum, the technology might be so sophisticated and mature that everything John and I said is wrong. and we have a fantastic future fleet capable of competing with China and Indo-Pacom for the next high-end fight. On the other side of the spectrum, we have a world where everybody buys the hype, where these programs, these new vessels and their technologies are not integrated judiciously, which is what we spoke about right at the beginning of our discussion. And you start to have rapid failure of these platforms in a crisis or a combat scenario. If that happens, If John and I are correct, in other words, then we'll start to see things like ISR sensing and certain self-defense capabilities like anti-submarine warfare fail earlier than the fleet's capability to generate fires, particularly long range fires.

And that will lead to what we discussed in the last question, which was an acceleration of decision making. In other words, commanders facing pressure to employ weapon systems before the sensor data that enables them stops arriving. Another phenomenon we also spoke about in the article that we could see is a maintenance and sustainment issue, which is even in non-wartime conditions, we have to work out if these vessels start failing in large numbers, how to repair them in theater and how to station all the repair parts and the expertise and so on that we need, whether that be on forward deployed platforms or on the shore side. And then we have to work out how we're going to do that in combat operations, which is going to be significantly more difficult. And then if we accept the fact

that a certain number of these are going to fail in combat operations and not be able to be employed on scene, we have to work out how to stop them from falling into enemy hands. Because under this unified fleet concept, distributed maritime operations, networked information streams, coordination of fires, these vessels are going to be highly networked. Their primary role is the provision of data. That data has to get back to the fleet. And so if one of these things falls in adversaries' hands. That's an avenue of exploitation that's potentially system-wide very dangerous. I agree. I think it's a safe assumption. These assets will be more and more part of our present fleet and part of our future force. I've even had the opportunity in some of

my time underway during this chief engineer tour is to see some of these vessels remotely controlled from the Manchester. And so we know it's going to be a part of our fleet. And I think Jonathan does a good job summarizing their role as dull, dirty, and dangerous. I'm personally a fan of alliteration. So a couple more, there could be scouts, screens, and sweeps, right? Depending on our ability to get over some of the anti-boarding force protection and engineering concerns, right? If they're able to become shooters as well. Ranger, one of the vessels participating in RIMPAC this year. They launched an SM6 last fall. Again, stuff you can look up in USNI articles, et cetera. We do see what these roles, how they're going to be.

I think what I am most kind of excited and nervous about from an integration standpoint is kind of what I alluded to in my opening comments. Why am I interested in this? How do future commanding officers, strike group commanders, et cetera, manage and lead hybrid teams? There are going to be biases, even with automation. That automation is written in computer code by humans. Those humans are going to be making assumption decisions that are going to be baked into these systems, right? So the biases will exist. How are we going to be able to identify what those biases are ahead of time or during operations remains to be seen. There will be decisions of when to employ these systems, when to leave them on the bench, based on how we perceive their reliability

and resiliency to be at sea. Another integration point is what is the training going to be for onboard personnel? One of the things we saw in the LCS program was this, I think, kind of not a great belief that we were going to be able to contract everything out, right? If we had a system failure or if, you know, equipment dropped off, right, we'd be able to identify quickly the contractors to come on board, meet us in port, and fix it. And the sailors would be kind of operators only, right? How are we going to prepare our sailors to recognize that when you're in the maritime environment, it's just you and the assets you have available to you out there, right? The provision of contractors might not be available to you. So how are we providing these training to our

sailors in order to operate legacy equipment with kind of the challenges they present? And at the same time, how do you operate on demand assets with the challenges that are kind of yet to be discovered? I think besides some of these like tactical and operational impacts that Jonathan discussed, like these integration pieces at very much the human level and that operator level is going to be a unique challenge that we still have to grapple with. And I think that might be what's most exciting for me. Well, unfortunately, that's all the time that we have for today. I'd like to thank my guests, Jonathan Falcone and Jonathan Panter. Jonathan Falcone, where can we find you online? What are you working on next? Best places to find me online, LinkedIn. I have a Twitter

account. Those are the best places to find me. As you can probably tell during the course of this conversation. Next project for me is on data collection, but I'm working on a more kind of technical piece, which is why I'm still kind of forming my thoughts around it. So if there are any listeners out there with ideas on data compression, data monitoring systems, or the like, please reach out. I'd love to have a conversation and learn more, but I look forward to kind of putting out a more technical piece on that. Jonathan Panter, where can we find you and what's your next project? Best place to find me is on LinkedIn. You can just search my name. My major ongoing project is, of course, my dissertation, which is about civilian oversight

in response to accidents and collisions and other incidents and whether that has adverse impacts on the U.S. Navy's performance in crisis and combat. John and I are also going to turn the Warren the Rocks article that this podcast is about into a book chapter for a forthcoming edited volume in USNI Press that's ongoing this year. And I'm also working on a draft article on fleet escalation dynamics in a future Taiwan Straits or South China Sea scenario. Well, thank you both again for joining us. To listeners, thanks for tuning in. We'll see you next time. © BF-WATCH TV 2021 We'll be right back.