Transcript
00:00:00 The following is a conversation with Keoki Jackson.
00:00:03 He’s the CTO of Lockheed Martin,
00:00:06 a company that through its long history
00:00:08 has created some of the most incredible engineering marvels
00:00:11 human beings have ever built,
00:00:13 including planes that fly fast and undetected,
00:00:17 defense systems that intersect nuclear threats
00:00:19 that can take the lives of millions,
00:00:22 and systems that venture out into space,
00:00:25 the moon, Mars, and beyond.
00:00:28 And these days, more and more artificial intelligence
00:00:31 has an assistive role to play in these systems.
00:00:34 I’ve read several books in preparation for this conversation.
00:00:38 It is a difficult one,
00:00:40 because in part Lockheed Martin builds military systems
00:00:43 that operate in a complicated world
00:00:45 that often does not have easy solutions
00:00:48 in the gray area between good and evil.
00:00:52 I hope one day this world will rid itself of war
00:00:56 in all its forms.
00:00:58 But the path to achieving that in a world
00:01:00 that does have evil is not obvious.
00:01:02 What is obvious is good engineering
00:01:05 and artificial intelligence research
00:01:07 has a role to play on the side of good.
00:01:11 Lockheed Martin and the rest of our community
00:01:13 are hard at work at exactly this task.
00:01:17 We talk about these and other important topics
00:01:19 in this conversation.
00:01:21 Also, most certainly, both Keoki and I
00:01:24 have a passion for space,
00:01:27 us humans venturing out toward the stars.
00:01:32 We talk about this exciting future as well.
00:01:35 This is the Artificial Intelligence Podcast.
00:01:38 If you enjoy it, subscribe on YouTube,
00:01:40 give it five stars on iTunes, support it on Patreon,
00:01:43 or simply connect with me on Twitter at Lex Friedman,
00:01:47 spelled F R I D M A N.
00:01:50 And now, here’s my conversation with Keoki Jackson.
00:01:55 I read several books on Lockheed Martin recently.
00:01:57 My favorite in particular is by Ben Rich,
00:02:00 Carlos Concord’s personal memoir.
00:02:03 It gets a little edgy at times.
00:02:05 But from that, I was reminded that the engineers
00:02:09 at Lockheed Martin have created some of the most
00:02:11 incredible engineering marvels human beings have ever built
00:02:15 throughout the 20th century and the 21st.
00:02:18 Do you remember a particular project or system at Lockheed
00:02:22 or before that at the Space Shuttle Columbia
00:02:25 that you were just in awe at the fact that us humans
00:02:29 could create something like this?
00:02:32 You know, that’s a great question.
00:02:34 There’s a lot of things that I could draw on there.
00:02:37 When you look at the Skunk Works and Ben Rich’s book
00:02:39 in particular, of course, it starts off with basically
00:02:42 the start of the jet age and the P 80.
00:02:44 And I had the opportunity to sit next to one
00:02:49 of the Apollo astronauts, Charlie Duke, recently at dinner.
00:02:53 And I said, hey, what’s your favorite aircraft?
00:02:56 And he said, well, it was by far the F 104 Starfighter,
00:02:59 which was another aircraft that came out of Lockheed there.
00:03:02 It was the first Mach 2 jet fighter aircraft.
00:03:08 They called it the missile with a man in it.
00:03:11 And so those are the kinds of things I grew up hearing
00:03:13 stories about.
00:03:15 You know, of course, the SR 71 is incomparable
00:03:19 as kind of the epitome of speed, altitude,
00:03:24 and just the coolest looking aircraft ever.
00:03:26 So there’s a reconnaissance, that’s a plane.
00:03:29 That’s a, yeah, intelligence surveillance
00:03:30 and reconnaissance aircraft that was designed
00:03:33 to be able to outrun, basically go faster
00:03:36 than any air defense system.
00:03:38 But, you know, I’ll tell you, I’m a space junkie.
00:03:42 That’s why I came to MIT.
00:03:44 That’s really what took me ultimately to Lockheed Martin.
00:03:49 And I grew up, and so Lockheed Martin, for example,
00:03:51 has been essentially at the heart of every planetary mission,
00:03:56 like all the Mars missions we’ve had a part in.
00:03:59 And we’ve talked a lot about the 50th anniversary
00:04:02 of Apollo here in the last couple of weeks, right?
00:04:04 But remember, 1976, July 20th, again, National Space Days,
00:04:09 the landing of the Viking lander on the surface of Mars,
00:04:15 just a huge accomplishment.
00:04:17 And when I was a young engineer at Lockheed Martin,
00:04:19 I got to meet engineers who had designed, you know,
00:04:23 various pieces of that mission as well.
00:04:25 So that’s what I grew up on is these planetary missions,
00:04:28 the start of the space shuttle era,
00:04:30 and ultimately had the opportunity
00:04:34 to see Lockheed Martin’s part.
00:04:38 Lockheed Martin’s part, and we can maybe talk about
00:04:40 some of these here, but Lockheed Martin’s part
00:04:42 in all of these space journeys over the years.
00:04:44 Do you dream, and I apologize for getting philosophical
00:04:47 at times, or sentimental.
00:04:49 I do romanticize the notion of space exploration.
00:04:53 So do you dream of the day when us humans colonize
00:04:56 another planet like Mars, or a man, a woman,
00:05:00 a human being steps on Mars?
00:05:03 Absolutely, and that’s a personal dream of mine.
00:05:06 I haven’t given up yet on my own opportunity
00:05:09 to fly into space, but as, you know,
00:05:12 from the Lockheed Martin perspective,
00:05:14 this is something that we’re working towards every day.
00:05:16 And of course, you know, we’re building
00:05:18 the Orion spacecraft, which is the most sophisticated
00:05:21 human rated spacecraft ever built.
00:05:23 And it’s really designed for these deep space journeys,
00:05:26 you know, starting with the moon,
00:05:28 but ultimately going to Mars and being the platform,
00:05:32 you know, from a design perspective,
00:05:34 we call the Mars base camp to be able to take humans
00:05:37 to the surface, and then after a mission
00:05:40 of a couple of weeks, bring them back up safely.
00:05:42 And so that is something I want to see happen
00:05:44 during my time at Lockheed Martin.
00:05:46 So I’m pretty excited about that.
00:05:49 And I think, you know, once we prove that’s possible,
00:05:52 you know, colonization might be a little bit further out,
00:05:57 but it’s something that I’d hope to see.
00:06:00 So maybe you can give a little bit of an overview
00:06:03 of, so Lockheed Martin has partnered with a few years ago
00:06:07 with Boeing to work with the DOD and NASA
00:06:09 to build launch systems and rockets with the ULA.
00:06:13 What’s beyond that?
00:06:15 What’s Lockheed’s mission timeline,
00:06:17 long term dream in terms of space?
00:06:19 You mentioned the moon, I’ve heard you talk about asteroids.
00:06:25 As Mars, what’s the timeline?
00:06:27 What’s the engineering challenges
00:06:29 and what’s the dream long term?
00:06:31 Yeah, I think the dream long term is to have
00:06:33 a permanent presence in space beyond low earth orbit,
00:06:37 ultimately with a long term presence on the moon
00:06:41 and then to the planets, to Mars.
00:06:43 And… Sorry to interrupt on that.
00:06:45 So long term presence means…
00:06:48 Sustained and sustainable presence in an economy,
00:06:51 a space economy that really goes alongside that.
00:06:54 With human beings and being able to launch perhaps
00:06:58 from those, so like hop?
00:07:02 You know, there’s a lot of energy
00:07:04 that goes in those hops, right?
00:07:06 So I think the first step is being able to get there
00:07:09 and to be able to establish sustained bases, right?
00:07:12 And build from there.
00:07:14 And a lot of that means getting, as you know,
00:07:18 things like the cost of launch down
00:07:21 and you mentioned United Launch Alliance.
00:07:23 And so I don’t wanna speak for ULA,
00:07:26 but obviously they’re working really hard
00:07:29 to on their next generation of launch vehicles
00:07:34 to maintain that incredible mission success record
00:07:39 that ULA has, but ultimately continue
00:07:41 to drive down the cost and make the flexibility,
00:07:43 the speed and the access ever greater.
00:07:46 So what’s the missions that are in the horizon
00:07:50 that you could talk to?
00:07:51 Is there a hope to get to the moon?
00:07:53 Absolutely, absolutely.
00:07:54 I mean, I think you know this, or you may know this,
00:07:58 there’s a lot of ways to accomplish some of these goals.
00:08:00 And so that’s a lot of what’s in discussion today.
00:08:03 But ultimately the goal is to be able to establish a base
00:08:09 essentially in cislunar space that would allow
00:08:12 for ready transfer from orbit to the lunar surface
00:08:19 and back again.
00:08:19 And so that’s sort of that near term,
00:08:21 I say near term in the next decade or so vision,
00:08:26 starting off with a stated objective by this administration
00:08:29 to get back to the moon in the 2024, 2025 timeframe,
00:08:34 which is right around the corner here.
00:08:37 How big of an engineering challenge is that?
00:08:41 I think the big challenge is not so much to go,
00:08:44 but to stay, right?
00:08:46 And so we demonstrated in the 60s
00:08:48 that you could send somebody up,
00:08:50 do a couple of days of mission
00:08:52 and bring them home again successfully.
00:08:55 Now we’re talking about doing that,
00:08:57 I’d say more to, I don’t wanna say an industrial scale,
00:08:59 but a sustained scale, right?
00:09:01 So permanent habitation, regular reuse of vehicles,
00:09:09 the infrastructure to get things like fuel, air,
00:09:15 consumables, replacement parts,
00:09:17 all the things that you need to sustain
00:09:18 that kind of infrastructure.
00:09:20 So those are certainly engineering challenges,
00:09:23 there are budgetary challenges,
00:09:26 and those are all things
00:09:28 that we’re gonna have to work through.
00:09:30 The other thing, and I shouldn’t,
00:09:33 I don’t wanna minimize this,
00:09:35 I mean, I’m excited about human exploration,
00:09:38 but the reality is our technology
00:09:40 and where we’ve come over the last 40 years essentially
00:09:44 has changed what we can do with robotic exploration as well.
00:09:48 And to me, it’s incredibly thrilling,
00:09:52 and this seems like old news now,
00:09:53 but the fact that we have rovers driving around
00:09:57 the surface of Mars and sending back data
00:10:00 is just incredible.
00:10:01 The fact that we have satellites in orbit around Mars
00:10:04 that are collecting weather,
00:10:06 they’re looking at the terrain, they’re mapping,
00:10:08 all of these kinds of things on a continuous basis,
00:10:11 that’s incredible.
00:10:12 And the fact that you got the time lag, of course,
00:10:15 going to the planets,
00:10:17 but you can effectively have virtual human presence there
00:10:22 in a way that we have never been able to do before.
00:10:25 And now with the advent of even greater processing power,
00:10:30 better AI systems, better cognitive systems
00:10:33 and decision systems,
00:10:35 you put that together with the human piece
00:10:38 and we’ve really opened up the solar system
00:10:41 in a whole different way.
00:10:42 And I’ll give you an example, we’ve got OSIRIS REx,
00:10:44 which is a mission to the asteroid Bennu.
00:10:47 So the spacecraft is out there right now
00:10:50 on basically a year mapping activity
00:10:54 to map the entire surface of that asteroid in great detail.
00:10:59 You know, all autonomously piloted, right?
00:11:02 But the idea then that, and this is not too far away,
00:11:04 it’s gonna go in,
00:11:05 it’s got a sort of fancy vacuum cleaner with a bucket,
00:11:09 it’s gonna collect the sample off the asteroid
00:11:12 and then send it back here to Earth.
00:11:14 And so, you know, we have gone from sort of those
00:11:18 tentative steps in the 70s, you know,
00:11:21 early landings, video of the solar system
00:11:23 to now we’ve sent spacecraft to Pluto,
00:11:27 we have gone to comets and brought and intercepted comets,
00:11:31 we’ve brought stardust, you know, material back.
00:11:37 So that’s, we’ve gone far
00:11:40 and there’s incredible opportunity to go even farther.
00:11:43 So it seems quite crazy that this is even possible,
00:11:47 that can you talk a little bit about
00:11:51 what it means to orbit an asteroid
00:11:54 and with a bucket to try to pick up some soil samples?
00:11:58 Yeah, so part of it is just kind of the, you know,
00:12:02 these are the same kinds of techniques we use here on Earth
00:12:05 for high speed, high accuracy imagery,
00:12:10 stitching these scenes together and creating
00:12:14 essentially high accuracy world maps, right?
00:12:17 And so that’s what we’re doing, obviously,
00:12:20 on a much smaller scale with an asteroid.
00:12:23 But the other thing that’s really interesting,
00:12:24 you put together sort of that neat control
00:12:28 and, you know, data and imagery problem.
00:12:33 But the stories around how we designed the collection,
00:12:36 I mean, as essentially, you know,
00:12:38 this is the sort of the human ingenuity element, right?
00:12:41 That, you know, essentially had an engineer who had a,
00:12:45 one day he’s like, oh, starts messing around with parts,
00:12:49 vacuum cleaner, bucket, you know,
00:12:51 maybe we could do something like this.
00:12:53 And that was what led to what we call
00:12:55 the pogo stick collection, right?
00:12:57 Where basically a thing comes down,
00:12:59 it’s only there for seconds, does that collection,
00:13:02 grabs the, essentially blows the regolith material
00:13:07 into the collection hopper and off it goes.
00:13:10 It doesn’t really land almost.
00:13:12 It’s a very short landing.
00:13:13 Wow, that’s incredible.
00:13:15 So what is, in those, we talked a little bit more
00:13:20 about space, what’s the role of the human in all of this?
00:13:24 What are the challenges?
00:13:25 What are the opportunities for humans
00:13:29 as they pilot these vehicles in space?
00:13:33 And for humans that may step foot
00:13:37 on either the moon or Mars?
00:13:41 Yeah, it’s a great question because, you know,
00:13:43 I just have been extolling the virtues of robotic
00:13:47 and, you know, rovers, autonomous systems,
00:13:50 and those absolutely have a role.
00:13:53 I think the thing that we don’t know how to replace today
00:13:57 is the ability to adapt on the fly to new information.
00:14:02 And I believe that will come, but we’re not there yet.
00:14:07 There’s a ways to go.
00:14:08 And so, you know, you think back to Apollo 13
00:14:13 and the ingenuity of the folks on the ground
00:14:15 and on the spacecraft essentially cobbled together
00:14:19 a way to get the carbon dioxide scrubbers to work.
00:14:23 Those are the kinds of things that ultimately, you know,
00:14:28 and I’d say not just from dealing with anomalies,
00:14:31 but, you know, dealing with new information.
00:14:33 You see something and rather than waiting
00:14:37 20 minutes or half an hour, an hour
00:14:39 to try to get information back and forth,
00:14:42 but be able to essentially revector on the fly,
00:14:45 collect, you know, different samples,
00:14:47 take a different approach,
00:14:49 choose different areas to explore.
00:14:52 Those are the kinds of things that human presence enables
00:14:56 that is still a ways ahead of us on the AI side.
00:15:00 Yeah, there’s some interesting stuff
00:15:01 we’ll talk about on the teaming side here on Earth.
00:15:04 That’s pretty cool to explore.
00:15:06 And in space, let’s not leave the space piece out.
00:15:08 So what does teaming, what does AI and humans
00:15:11 working together in space look like?
00:15:13 Yeah, one of the things we’re working on
00:15:15 is a system called Maya, which is,
00:15:18 you think of it, so it’s an AI assistant.
00:15:21 In space. In space, exactly.
00:15:24 And you think of it as the Alexa in space, right?
00:15:28 But this goes hand in hand with a lot of other developments.
00:15:31 And so today’s world, everything is essentially model based,
00:15:35 model based systems engineering
00:15:38 to the actual digital tapestry that goes through the design,
00:15:42 the build, the manufacture, the testing,
00:15:44 and ultimately the sustainment of these system.
00:15:47 And so our vision is really that, you know,
00:15:50 when our astronauts are there around Mars,
00:15:54 you’re gonna have that entire digital library
00:15:59 of the spacecraft, of its operations, all the test data,
00:16:04 all the test data and flight data from previous missions
00:16:08 to be able to look and see if there are anomalous conditions
00:16:11 and tell the humans and potentially deal with that
00:16:16 before it becomes a bad situation
00:16:20 and help the astronauts work through those kinds of things.
00:16:23 And it’s not just, you know,
00:16:25 dealing with problems as they come up,
00:16:26 but also offering up opportunities
00:16:29 for additional exploration capability, for example.
00:16:32 So that’s the vision is that, you know,
00:16:34 these are gonna take the best of the human
00:16:37 to respond to changing circumstances
00:16:41 and rely on the best of AI capabilities
00:16:44 to monitor these, you know,
00:16:46 this almost infinite number of data points
00:16:49 and correlations of data points
00:16:51 that humans frankly aren’t that good at.
00:16:54 So how do you develop systems in space like this,
00:16:56 whether it’s Alexa in space or in general,
00:17:01 any kind of control systems,
00:17:03 any kind of intelligent systems
00:17:04 when you can’t really test stuff too much out in space?
00:17:08 It’s very expensive to test stuff.
00:17:10 So how do you develop such systems?
00:17:14 Yeah, that’s the beauty of this digital twin, if you will.
00:17:19 And of course, with Lockheed Martin,
00:17:21 we’ve over the past, you know, five plus decades
00:17:24 been refining our knowledge of the space environment,
00:17:28 of how materials behave, dynamics,
00:17:32 the controls, the radiation environments,
00:17:35 all of these kinds of things.
00:17:37 So we’re able to create very sophisticated models.
00:17:39 They’re not perfect, but they’re very good.
00:17:43 And so you can actually do a lot.
00:17:46 I spent part of my career, you know,
00:17:49 simulating communication spacecraft,
00:17:53 you know, missile warning spacecraft, GPS spacecraft
00:17:56 in all kinds of scenarios and all kinds of environments.
00:17:59 So this is really just taking that to the next level.
00:18:01 The interesting thing is that now
00:18:03 you’re bringing into that loop
00:18:06 a system depending on how it’s developed
00:18:08 that may be non deterministic,
00:18:10 it may be learning as it goes.
00:18:13 And in fact, we anticipate
00:18:14 that it will be learning as it goes.
00:18:16 And so that brings a whole new level of interest,
00:18:21 I guess, into how do you do verification and validation
00:18:25 of these non deterministic learning systems
00:18:28 in scenarios that may go out of the bounds
00:18:31 or the envelope that you have initially designed them to.
00:18:35 So had this system and its intelligence
00:18:37 has the same complexity,
00:18:39 some of the same complexity human does
00:18:41 and learns over time, it’s unpredictable
00:18:43 in certain kinds of ways in the,
00:18:46 so you still, you also have to model that
00:18:49 when you’re thinking about it.
00:18:50 So in your thoughts, it’s possible
00:18:53 to model the majority of situations,
00:18:57 the important aspects of situations here on earth
00:18:59 and in space enough to test stuff?
00:19:02 Yeah, this is really an active area of research
00:19:05 and we’re actually funding university research
00:19:07 in a variety of places, including MIT.
00:19:10 This is in the realm of trust and verification
00:19:13 and validation of I’d say autonomous systems in general
00:19:18 and then as a subset of that autonomous systems
00:19:21 that incorporate artificial intelligence capabilities.
00:19:24 And this is not an easy problem.
00:19:27 We’re working with startup companies,
00:19:29 we’ve got internal R&D, but our conviction is
00:19:33 that autonomy and more and more AI enabled autonomy
00:19:39 is gonna be in everything that Lockheed Martin develops
00:19:42 and fields and it’s gonna be retrofitting it.
00:19:46 Autonomy and AI are gonna be retrofit
00:19:48 into existing systems, they’re gonna be part
00:19:50 of the design for all of our future systems.
00:19:54 And so maybe I should take a step back
00:19:56 and say the way we define autonomy.
00:19:58 So we talk about autonomy essentially a system
00:20:02 that composes, selects and then executes decisions
00:20:08 with varying levels of human intervention.
00:20:12 And so you could think of no autonomy.
00:20:15 So this is essentially the human doing the task.
00:20:18 You can think of effectively partial autonomy
00:20:23 where the human is in the loop.
00:20:25 So making decisions in every case
00:20:29 about what the autonomous system can do.
00:20:31 Either in the cockpit or remotely.
00:20:33 Or remotely, exactly, but still in that control loop.
00:20:36 And then there’s what you’d call supervisory autonomy.
00:20:39 So the autonomous system is doing most of the work,
00:20:42 the human can intervene to stop it
00:20:44 or to change the direction.
00:20:45 And then ultimately full autonomy
00:20:47 where the human is off the loop altogether.
00:20:50 And for different types of missions
00:20:52 wanna have different levels of autonomy.
00:20:55 So now take that spectrum and this conviction
00:20:58 that autonomy and more and more AI
00:21:01 are in everything that we develop.
00:21:05 The kinds of things that Lockheed Martin does,
00:21:07 a lot of times are safety of life critical kinds of missions.
00:21:12 You think about aircraft, for example.
00:21:15 And so we require and our customers require
00:21:20 an extremely high level of confidence.
00:21:23 One, that we’re gonna protect life.
00:21:26 Two, that these systems will behave
00:21:30 in ways that their operators can understand.
00:21:33 And so this gets into that whole field.
00:21:36 Again, being able to verify and validate
00:21:40 that the systems have been and that they will operate
00:21:44 the way they’re designed and the way they’re expected.
00:21:48 And furthermore, that they will do that
00:21:50 in ways that can be explained and understood.
00:21:55 And that is an extremely difficult challenge.
00:21:58 Yeah, so here’s a difficult question.
00:22:00 I don’t mean to bring this up,
00:22:04 but I think it’s a good case study
00:22:05 that people are familiar with the Boeing 737 Max
00:22:10 commercial airplane has had two recent crashes
00:22:13 where their flight control software system failed
00:22:15 and it’s software.
00:22:17 So I don’t mean to speak about Boeing,
00:22:19 but broadly speaking, we have this
00:22:21 in the autonomous vehicle space too, semi autonomous.
00:22:24 We have millions of lines of code software making decisions.
00:22:30 There is a little bit of a clash of cultures
00:22:32 because software engineers don’t have the same culture
00:22:37 of safety often that people who build systems
00:22:41 like at Lockheed Martin do where it has to be
00:22:45 exceptionally safe, you have to test this on.
00:22:48 So how do we get this right when software
00:22:50 is making so many decisions?
00:22:53 Yeah, and there’s a lot of things that have to happen.
00:22:57 And by and large, I think it starts with the culture,
00:23:01 which is not necessarily something that A,
00:23:04 is taught in school or B is something that would come,
00:23:07 depending on what kind of software you’re developing,
00:23:10 it may not be relevant, right?
00:23:13 If you’re targeting ads or something like that.
00:23:15 So, and by and large, I’d say not just Lockheed Martin,
00:23:20 but certainly the aerospace industry as a whole
00:23:23 has developed a culture that does focus on safety,
00:23:27 safety of life, operational safety, mission success.
00:23:31 But as you note, these systems
00:23:34 have gotten incredibly complex.
00:23:36 And so they’re to the point where it’s almost impossible,
00:23:40 you know, state spaces become so huge
00:23:42 that it’s impossible to, or very difficult
00:23:45 to do a systematic verification across the entire set
00:23:50 of potential ways that an aircraft could be flown,
00:23:53 all the conditions that could happen,
00:23:55 all the potential failure scenarios.
00:23:59 Now, maybe that’s soluble one day,
00:24:01 maybe when we have our quantum computers
00:24:03 at our fingertips, we’ll be able to actually
00:24:06 simulate across an entire, you know,
00:24:09 almost infinite state space.
00:24:11 But today, you know, there’s a lot of work
00:24:16 to really try to bound the system,
00:24:20 to make sure that it behaves in predictable ways,
00:24:24 and then have this culture of continuous inquiry
00:24:29 and skepticism and questioning to say,
00:24:33 did we really consider the right realm of possibilities?
00:24:37 Have we done the right range of testing?
00:24:40 Do we really understand, you know, in this case,
00:24:42 you know, human and machine interactions,
00:24:44 the human decision process alongside the machine processes?
00:24:49 And so that’s that culture,
00:24:51 we call it the culture of mission success at Lockheed Martin
00:24:54 that really needs to be established.
00:24:56 And it’s not something, you know,
00:24:57 it’s something that people learn by living in it.
00:25:02 And it’s something that has to be promulgated, you know,
00:25:05 and it’s done, you know, from the highest levels
00:25:07 at a company of Lockheed Martin, like Lockheed Martin.
00:25:10 Yeah, and the same is being faced
00:25:12 at certain autonomous vehicle companies
00:25:14 where that culture is not there
00:25:15 because it started mostly by software engineers.
00:25:18 So that’s what they’re struggling with.
00:25:21 Is there lessons that you think we should learn
00:25:25 as an industry and a society from the Boeing 737 MAX crashes?
00:25:30 These crashes obviously are tremendous tragedies.
00:25:34 They’re tragedies for all of the people,
00:25:37 the crew, the families, the passengers,
00:25:41 the people on the ground involved.
00:25:44 And, you know, it’s also a huge business
00:25:47 and economic setback as well.
00:25:49 I mean, you know, we’ve seen that it’s impacting
00:25:51 essentially the trade balance of the US.
00:25:53 So these are important questions.
00:25:58 And these are the kinds that, you know,
00:26:00 we’ve seen similar kinds of questioning at times.
00:26:03 You know, you go back to the Challenger accident.
00:26:06 And it is, I think, always important to remind ourselves
00:26:10 that humans are fallible, that the systems we create,
00:26:14 as perfect as we strive to make them,
00:26:16 we can always make them better.
00:26:18 And so another element of that culture of mission success
00:26:21 is really that commitment to continuous improvement.
00:26:24 If there’s something that goes wrong,
00:26:27 a real commitment to root cause
00:26:31 and true root cause understanding,
00:26:33 to taking the corrective actions
00:26:35 and to making the future systems better.
00:26:38 And certainly we strive for, you know, no accidents.
00:26:45 And if you look at the record
00:26:47 of the commercial airline industry as a whole
00:26:50 and the commercial aircraft industry as a whole,
00:26:52 you know, there’s a very nice decaying exponential
00:26:57 to years now where we have
00:26:59 no commercial aircraft accidents at all, right?
00:27:02 Fatal accidents at all.
00:27:04 So that didn’t happen by accident.
00:27:08 It was through the regulatory agencies, FAA,
00:27:11 the airframe manufacturers really working on a system
00:27:16 to identify root causes and drive them out.
00:27:20 So maybe we can take a step back
00:27:23 and many people are familiar, but Lockheed Martin broadly,
00:27:28 what kind of categories of systems
00:27:32 are you involved in building?
00:27:34 You know, Lockheed Martin, we think of ourselves
00:27:36 as a company that solves hard mission problems.
00:27:39 And the output of that might be an airplane or a spacecraft
00:27:43 or a helicopter or a radar or something like that.
00:27:45 But ultimately we’re driven by these, you know,
00:27:48 what is our customer?
00:27:50 What is that mission that they need to achieve?
00:27:52 And so that’s what drove the SR71, right?
00:27:55 How do you get pictures of a place
00:27:59 where you’ve got sophisticated air defense systems
00:28:02 that are capable of handling any aircraft
00:28:05 that was out there at the time, right?
00:28:07 So that, you know, that’s what yielded an SR71.
00:28:10 Let’s build a nice flying camera.
00:28:12 Exactly.
00:28:13 And make sure it gets out and it gets back, right?
00:28:15 And that led ultimately to really the start
00:28:18 of the space program in the US as well.
00:28:22 So now take a step back to Lockheed Martin of today.
00:28:24 And we are, you know, on the order of 105 years old now
00:28:29 between Lockheed and Martin, the two big heritage companies.
00:28:32 Of course, we’re made up of a whole bunch
00:28:33 of other companies that came in as well.
00:28:36 General Dynamics, you know, kind of go down the list.
00:28:39 Today, you can think of us in this space
00:28:43 of solving mission problems.
00:28:44 So obviously on the aircraft side, tactical aircraft,
00:28:50 building the most advanced fighter aircraft
00:28:53 that the world has ever seen.
00:28:55 We’re up to now several hundred of those delivered,
00:28:57 building almost a hundred a year.
00:29:00 And of course, working on the things that come after that.
00:29:04 On the space side, we are engaged
00:29:06 in pretty much every venue of space utilization
00:29:12 and exploration you can imagine.
00:29:14 So I mentioned things like navigation and timing GPS,
00:29:18 communication satellites, missile warning satellites.
00:29:22 We’ve built commercial surveillance satellites.
00:29:24 We’ve built commercial communication satellites.
00:29:27 We do civil space.
00:29:29 So everything from human exploration
00:29:32 to the robotic exploration of the outer planets.
00:29:35 And keep going on the space front.
00:29:39 But a couple of other areas that I’d like to put out,
00:29:42 we’re heavily engaged in building
00:29:45 critical defensive systems.
00:29:47 And so a couple that I’ll mention, the Aegis Combat System.
00:29:51 This is basically the integrated air and missile defense
00:29:54 system for the US and allied fleets.
00:29:58 And so protects carrier strike groups, for example,
00:30:03 from incoming ballistic missile threats,
00:30:06 aircraft threats, cruise missile threats,
00:30:08 and kind of go down the list.
00:30:09 So the carriers, the fleet itself
00:30:13 is the thing that is being protected.
00:30:15 The carriers aren’t serving
00:30:17 as a protection for something else.
00:30:19 Well, that’s a little bit of a different application.
00:30:21 We’ve actually built the version called Aegis Ashore,
00:30:24 which is now deployed in a couple of places around the world.
00:30:27 So that same technology, I mean, basically can be used
00:30:32 to protect either an ocean going fleet
00:30:35 or a land based activity.
00:30:37 Another one, the THAAD program.
00:30:40 So THAAD, this is the Theater High Altitude Area Defense.
00:30:45 This is to protect relatively broad areas
00:30:49 against sophisticated ballistic missile threats.
00:30:53 And so now it’s deployed with a lot of US capabilities.
00:30:58 And now we have international customers
00:31:00 that are looking to buy that capability as well.
00:31:03 And so these are systems that defend,
00:31:05 not just defend militaries and military capabilities,
00:31:09 but defend population areas.
00:31:12 We saw maybe the first public use of these
00:31:15 back in the first Gulf War with the Patriot Systems.
00:31:20 And these are the kinds of things
00:31:22 that Lockheed Martin delivers.
00:31:25 And there’s a lot of stuff that goes into it.
00:31:28 A lot of stuff that goes with it.
00:31:29 So think about the radar systems and the sensing systems
00:31:33 that cue these, the command and control systems
00:31:36 that decide how you pair a weapon
00:31:39 against an incoming threat.
00:31:42 And then all the human and machine interfaces
00:31:45 to make sure that they can be operated successfully
00:31:48 in very strenuous environments.
00:31:51 Yeah, there’s some incredible engineering
00:31:54 that at every front, like you said.
00:31:57 So maybe if we just take a look at Lockheed history broadly,
00:32:03 maybe even looking at Skunk Works.
00:32:06 What are the biggest,
00:32:08 most impressive milestones of innovation?
00:32:11 So if you look at stealth, I would have called you crazy
00:32:14 if you said that’s possible at the time.
00:32:17 And supersonic and hypersonic.
00:32:21 So traveling at, first of all,
00:32:24 traveling at the speed of sound is pretty damn fast.
00:32:27 And supersonic and hypersonic,
00:32:29 three, four, five times the speed of sound.
00:32:32 That seems, I would also call you crazy
00:32:34 if you say you can do that.
00:32:35 So can you tell me how it’s possible
00:32:38 to do these kinds of things?
00:32:39 And is there other milestones and innovation
00:32:43 that’s going on that you can talk about?
00:32:45 Yeah.
00:32:45 Well, let me start on the Skunk Works saga.
00:32:49 And you kind of alluded to it in the beginning.
00:32:51 Skunk Works is as much an idea as a place.
00:32:54 And so it’s driven really by Kelly Johnson’s 14 principles.
00:32:59 And I’m not gonna list all 14 of them off,
00:33:01 but the idea, and this I’m sure will resonate
00:33:04 with any engineer who’s worked
00:33:06 on a highly motivated small team before.
00:33:09 The idea that if you can essentially have a small team
00:33:13 of very capable people who wanna work
00:33:17 on really hard problems, you can do almost anything.
00:33:20 Especially if you kind of shield them
00:33:23 from bureaucratic influences,
00:33:26 if you create very tight relationships with your customers
00:33:30 so that you have that team
00:33:32 and shared vision with the customer.
00:33:35 Those are the kinds of things that enable the Skunk Works
00:33:40 to do these incredible things.
00:33:42 And we listed off a number that you brought up stealth.
00:33:46 And I wish I could have seen Ben Rich with a ball bearing
00:33:51 rolling it across the desk to a general officer
00:33:54 and saying, would you like to have an aircraft
00:33:58 that has the radar cross section of this ball bearing?
00:34:01 Probably one of the least expensive
00:34:04 and most effective marketing campaigns
00:34:06 in the history of the industry.
00:34:08 So just for people that are not familiar,
00:34:10 the way you detect aircraft,
00:34:13 I’m sure there’s a lot of ways,
00:34:14 but radar for the longest time,
00:34:17 there’s a big blob that appears in the radar.
00:34:20 How do you make a plane disappear
00:34:22 so it looks as big as a ball bearing?
00:34:26 What’s involved in technology wise there?
00:34:28 What’s the broadly sort of the stuff you can speak about?
00:34:32 I’ll stick to what’s in Ben Rich’s book.
00:34:34 But obviously the geometry of how radar gets reflected
00:34:39 and the kinds of materials that either reflect or absorb
00:34:42 are kind of the couple of the critical elements there.
00:34:46 And it’s a cat and mouse game, right?
00:34:48 I mean, you know, radars get better,
00:34:51 stealth capabilities get better.
00:34:52 And so it’s a really a game
00:34:55 of continuous improvement and innovation there.
00:34:58 I’ll leave it at that.
00:35:00 Yeah, so the idea that something is essentially invisible
00:35:04 is quite fascinating.
00:35:06 But the other one is flying fast.
00:35:08 So speed of sound is 750, 60 miles an hour.
00:35:15 So supersonic is three, you know, Mach three,
00:35:18 something like that.
00:35:19 Yeah, we talk about the supersonic obviously,
00:35:21 and we kind of talk about that as that realm from Mach one
00:35:24 up through about Mach five and then hypersonic.
00:35:28 So, you know, high supersonic speeds would be past Mach five.
00:35:34 And you got to remember Lockheed Martin
00:35:37 and actually other companies have been involved
00:35:39 in hypersonic development since the late 60s.
00:35:42 You know, you think of everything from the X 15
00:35:45 to the space shuttle as examples of that.
00:35:50 I think the difference now is if you look around the world,
00:35:54 particularly the threat environment that we’re in today,
00:35:57 you’re starting to see, you know, publicly,
00:36:01 folks like the Russians and the Chinese
00:36:03 saying they have hypersonic weapons capability
00:36:08 that could threaten US and allied capabilities.
00:36:14 And also basically, you know, the claims are
00:36:17 these could get around defensive systems
00:36:19 that are out there today.
00:36:21 And so there’s a real sense of urgency.
00:36:24 You hear it from folks like the undersecretary of defense
00:36:28 for research and engineering, Dr. Mike Griffin,
00:36:30 and others in the department of defense that hypersonics
00:36:33 is something that’s really important to the nation
00:36:39 in terms of both parity, but also defensive capabilities.
00:36:43 And so that’s something that, you know, we’re pleased.
00:36:46 It’s something that Lockheed Martin’s, you know,
00:36:47 had a heritage in, we’ve invested R and D dollars
00:36:51 on our side for many years.
00:36:53 And we have a number of things going on
00:36:56 with various US government customers in that field today
00:36:59 that we’re very excited about.
00:37:01 So I would anticipate we’ll be hearing more about that
00:37:04 in the future from our customers.
00:37:06 And I’ve actually haven’t read much about this.
00:37:08 Probably you can’t talk about much of it at all,
00:37:10 but on the defensive side,
00:37:12 it’s a fascinating problem of perception
00:37:15 of trying to detect things that are really hard to see.
00:37:18 Can you comment on how hard that problem is
00:37:21 and how hard is it to stay ahead,
00:37:26 even if we go back a few decades,
00:37:29 stay ahead of the competition?
00:37:30 Well, maybe I’d, again, you gotta think of these
00:37:33 as ongoing capability development.
00:37:36 And so think back to the early days of missile defense.
00:37:40 So this would be in the 80s, the SDI program.
00:37:44 And in that timeframe, we proved and Lockheed Martin proved
00:37:47 that you could hit a bullet with a bullet, essentially,
00:37:50 and which is something that had never been done before
00:37:53 to take out an incoming ballistic missile.
00:37:56 And so that’s led to these incredible hit to kill
00:37:59 kinds of capabilities, PAC 3.
00:38:03 That’s the Patriot Advanced Capability Model 3
00:38:06 that Lockheed Martin builds,
00:38:08 the THAAD system that I talked about.
00:38:12 So now hypersonics, they’re different from ballistic systems.
00:38:17 And so we gotta take the next step in defensive capability.
00:38:21 I can, I’ll leave that there, but I can only imagine.
00:38:26 Now, let me just comment sort of as an engineer,
00:38:29 it’s sad to know that so much that Lockheed has done
00:38:33 in the past is classified or today,
00:38:38 and it’s shrouded in secrecy.
00:38:40 It has to be by the nature of the application.
00:38:46 So like what I do, so what we do here at MIT,
00:38:49 we would like to inspire young engineers, young scientists,
00:38:53 and yet in the Lockheed case,
00:38:56 some of that engineer has to stay quiet.
00:38:59 How do you think about that?
00:39:00 How does that make you feel?
00:39:02 Is there a future where more can be shown
00:39:08 or is it just the nature of this world
00:39:11 that it has to remain secret?
00:39:13 It’s a good question.
00:39:15 I think the public can see enough of,
00:39:21 and including students who may be in grade school,
00:39:25 high school, college today,
00:39:28 to understand the kinds of really hard problems
00:39:32 that we work on.
00:39:34 And I mean, look at the F35, right?
00:39:36 And obviously a lot of the detailed performance levels
00:39:41 are sensitive and controlled.
00:39:43 But we can talk about what an incredible aircraft this is,
00:39:48 supersonic, super cruise, kind of a fighter,
00:39:53 stealth capabilities.
00:39:55 It’s a flying information system in the sky
00:39:58 with data fusion, sensor fusion capabilities
00:40:02 that have never been seen before.
00:40:03 So these are the kinds of things that I believe,
00:40:06 these are the kinds of things that got me excited
00:40:08 when I was a student.
00:40:08 I think these still inspire students today.
00:40:12 And the other thing I’d say,
00:40:14 I mean, people are inspired by space.
00:40:17 People are inspired by aircraft.
00:40:21 Our employees are also inspired by that sense of mission.
00:40:25 And I’ll just give you an example.
00:40:27 I had the privilege to work
00:40:30 and lead our GPS programs for some time.
00:40:34 And that was a case where I actually worked on a program
00:40:39 that touches billions of people every day.
00:40:41 And so when I said, I worked on GPS,
00:40:43 everybody knew what I was talking about,
00:40:45 even though they didn’t maybe appreciate
00:40:46 the technical challenges that went into that.
00:40:51 But I’ll tell you, I got a briefing one time
00:40:55 from a major in the Air Force.
00:40:57 And he said, I go by callsign GIMP, GPS is my passion.
00:41:04 I love GPS.
00:41:05 And he was involved in the operational test of the system.
00:41:09 And he said, I was out in Iraq,
00:41:11 and I was on a helicopter, Blackhawk helicopter,
00:41:17 and I was bringing back a sergeant
00:41:20 and a handful of troops from a deployed location.
00:41:23 And he said, my job is GPS.
00:41:26 So I asked that sergeant,
00:41:27 and he’s beaten down and kind of half asleep.
00:41:31 And I said, what do you think about GPS?
00:41:34 And he brightened up, his eyes lit up,
00:41:36 and he said, well, GPS,
00:41:37 that brings me and my troops home every day.
00:41:40 I love GPS.
00:41:41 And that’s the kind of story where it’s like,
00:41:43 okay, I’m really making a difference here
00:41:45 in the kind of work.
00:41:46 So that mission piece is really important.
00:41:49 The last thing I’ll say is,
00:41:51 and this gets to some of these questions
00:41:53 around advanced technologies.
00:41:56 It’s not, they’re not just airplanes
00:41:58 and spacecraft anymore.
00:42:00 For people who are excited
00:42:01 about advanced software capabilities,
00:42:03 about AI, about bringing machine learning,
00:42:06 these are the things that we’re doing
00:42:08 to exponentially increase the mission capabilities
00:42:13 that go on those platforms.
00:42:14 And those are the kinds of things
00:42:15 that I think are more and more visible to the public.
00:42:18 Yeah, I think autonomy, especially in flight,
00:42:21 is super exciting.
00:42:23 Do you see a day, here we go, back into philosophy,
00:42:28 future when most fighter jets
00:42:30 will be highly autonomous to a degree
00:42:35 where a human doesn’t need to be in the cockpit
00:42:38 in almost all cases?
00:42:40 Well, I mean, that’s a world
00:42:42 that to a certain extent we’re in today.
00:42:44 Now these are remotely piloted aircraft, to be sure.
00:42:47 But we have hundreds of thousands of flight hours a year now
00:42:53 in remotely piloted aircraft.
00:42:55 And then if you take the F35,
00:42:58 there are huge layers, I guess,
00:43:03 in levels of autonomy built into that aircraft
00:43:06 so that the pilot is essentially more of a mission manager
00:43:11 rather than doing the data,
00:43:13 the second to second elements of flying the aircraft.
00:43:17 So in some ways it’s the easiest aircraft
00:43:19 in the world to fly.
00:43:20 And kind of a funny story on that.
00:43:22 So I don’t know if you know
00:43:23 how aircraft carrier landings work,
00:43:27 but basically there’s what’s called a tail hook
00:43:30 and it catches wires on the deck of the carrier.
00:43:33 And that’s what brings the aircraft
00:43:37 to a screeching halt, right?
00:43:39 And there’s typically three of these wires.
00:43:41 So if you miss the first, the second one,
00:43:43 you catch the next one, right?
00:43:45 And we got a little criticism.
00:43:49 I don’t know how true this story is,
00:43:50 but we got a little criticism.
00:43:52 The F35 is so perfect, it always gets the second wires.
00:43:56 We’re wearing out the wire because it always hits that one.
00:44:00 But that’s the kind of autonomy that just makes these,
00:44:04 essentially up levels what the human is doing
00:44:06 to more of that mission manager.
00:44:08 So much of that landing by the F35 is autonomous.
00:44:12 Well, it’s just, the control systems are such
00:44:14 that you really have dialed out the variability
00:44:18 that comes with all the environmental conditions.
00:44:20 You’re wearing it out.
00:44:20 So my point is to a certain extent,
00:44:24 that world is here today.
00:44:27 Do I think that we’re gonna see a day anytime soon
00:44:30 when there are no humans in the cockpit?
00:44:31 I don’t believe that.
00:44:33 But I do think we’re gonna see much more
00:44:35 human machine teaming, and we’re gonna see that much more
00:44:38 at the tactical edge.
00:44:40 And we did a demo, and you asked about
00:44:42 what the Skunk Works is doing these days.
00:44:43 And so this is something I can talk about,
00:44:46 but we did a demo with the Air Force Research Laboratory.
00:44:51 We called it Have Raider.
00:44:52 And so using an F16 as an autonomous wingman,
00:44:59 and we demonstrated all kinds of maneuvers
00:45:02 and various mission scenarios with the autonomous F16
00:45:06 being that so called loyal or trusted wingman.
00:45:09 And so those are the kinds of things that,
00:45:12 we’ve shown what is possible now.
00:45:15 Given that you’ve up leveled that pilot
00:45:17 to be a mission manager, now they can control
00:45:20 multiple other aircraft.
00:45:22 Think of them almost as extensions of your own aircraft
00:45:25 flying alongside with you.
00:45:27 So that’s another example of how this is really
00:45:30 coming to fruition.
00:45:31 And then I mentioned the landings,
00:45:35 but think about just the implications for humans
00:45:39 and flight safety, and this goes a little bit back
00:45:41 to the discussion we were having about
00:45:43 how do you continuously improve the level of safety
00:45:47 through automation while working through the complexities
00:45:51 that automation introduces.
00:45:53 So one of the challenges that you have
00:45:54 in high performance fighter aircraft is what’s called G lock.
00:45:57 So this is G induced loss of consciousness.
00:46:00 So you pull nine Gs, you’re wearing a pressure suit,
00:46:02 that’s not enough to keep the blood going to your brain,
00:46:05 you black out.
00:46:07 And of course that’s bad if you happen to be flying low,
00:46:12 near the deck and in an obstacle or terrain environment.
00:46:17 And so we developed a system in our aeronautics division
00:46:22 called Auto Gcast, so autonomous ground collision
00:46:25 avoidance system.
00:46:27 And we built that into the F16.
00:46:30 It’s actually saved seven aircraft, eight pilots already
00:46:33 in a relatively short time it’s been deployed.
00:46:35 It was so successful that the Air Force said,
00:46:39 hey, we need to have this in the F35 right away.
00:46:41 So we’ve actually done testing of that now on the F35.
00:46:45 And we’ve also integrated an autonomous
00:46:49 air collision avoidance system.
00:46:51 So think the air to air problem.
00:46:52 So now it’s the integrated collision avoidance system.
00:46:56 But these are the kinds of capabilities,
00:46:58 I wouldn’t call them AI.
00:46:59 I mean, they’re very sophisticated models
00:47:04 of the aircraft dynamics coupled with the terrain models
00:47:08 to be able to predict when essentially the pilot
00:47:12 is doing something that is gonna take the aircraft
00:47:14 or the pilot’s not doing something in this case.
00:47:18 But it just gives you an example of how autonomy
00:47:22 can be really a lifesaver in today’s world.
00:47:25 It’s like a autonomous automated emergency braking in cars.
00:47:30 But is there any exploration of perception of,
00:47:34 for example, detecting a G lock that the pilot is out?
00:47:39 So as opposed to perceiving the external environment
00:47:43 to infer that the pilot is out,
00:47:44 but actually perceiving the pilot directly.
00:47:47 Yeah, this is one of those cases
00:47:48 where you’d like to not take action
00:47:50 if you think the pilot’s there.
00:47:52 And it’s almost like systems that try to detect
00:47:54 if a driver’s falling asleep on the road, right?
00:47:57 With limited success.
00:48:00 So, I mean, this is what I call
00:48:02 the system of last resort, right?
00:48:03 Where if the aircraft has determined
00:48:06 that it’s going into the terrain, get it out of there.
00:48:10 And this is not something that we’re just doing
00:48:13 in the aircraft world.
00:48:15 And I wanted to highlight,
00:48:16 we have a technology we call Matrix,
00:48:18 but this is developed at Sikorsky Innovations.
00:48:22 The whole idea there is what we call optimal piloting.
00:48:26 So not optional piloting or unpiloted, but optimal piloting.
00:48:32 So an FAA certified system.
00:48:35 So you have a high degree of confidence.
00:48:37 It’s generally pretty deterministic.
00:48:40 So we know that it’ll do in different situations,
00:48:43 but effectively be able to fly a mission
00:48:48 with two pilots, one pilot, no pilots.
00:48:51 And you can think of it almost as like a dial
00:48:56 of the level of autonomy that you want,
00:48:58 but able, so it’s running in the background at all times
00:49:01 and able to pick up tasks,
00:49:03 whether it’s sort of autopilot kinds of tasks
00:49:05 or more sophisticated path planning kinds of activities
00:49:12 to be able to do things like, for example,
00:49:14 land on an oil rig in the North Sea
00:49:16 in bad weather, zero, zero conditions.
00:49:19 And you can imagine, of course,
00:49:20 there’s a lot of military utility to capability like that.
00:49:24 You could have an aircraft that you want to send out
00:49:27 for a crewed mission, but then at night,
00:49:29 if you want to use it to deliver supplies
00:49:31 in an unmanned mode, that could be done as well.
00:49:35 And so there’s clear advantages there.
00:49:40 But think about on the commercial side,
00:49:41 if you’re an aircraft taken,
00:49:44 you’re gonna fly out to this oil rig.
00:49:46 If you get out there and you can’t land,
00:49:48 then you gotta bring all those people back,
00:49:50 reschedule another flight,
00:49:51 pay the overtime for the crew that you just brought back
00:49:54 because they didn’t get where they were going,
00:49:55 pay for the overtime for the folks
00:49:57 that are out there in the oil rig.
00:49:58 This is real economic,
00:50:00 these are dollars and cents kinds of advantages
00:50:03 we’re bringing in the commercial world as well.
00:50:06 So here’s a difficult question from the AI space
00:50:09 that I would love it if you’re able to comment.
00:50:11 So a lot of this autonomy in AI you’ve mentioned just now
00:50:15 has this empowering effect.
00:50:17 One is the last resort, it keeps you safe.
00:50:20 The other is there’s a, with the teaming
00:50:22 and in general, assistive AI.
00:50:29 And I think there’s always a race.
00:50:33 So the world is full of, the world is complex.
00:50:36 It’s full of bad actors.
00:50:41 So there’s often a race to make sure
00:50:43 that we keep this country safe, right?
00:50:48 But with AI, there is a concern
00:50:52 that it’s a slightly different race.
00:50:55 Though there’s a lot of people in the AI space
00:50:56 that are concerned about the AI arms race.
00:50:59 That as opposed to the United States becoming,
00:51:04 having the best technology and therefore keeping us safe,
00:51:07 even we lose ability to keep control of it.
00:51:11 So this, the AI arms race getting away
00:51:14 from all of us humans.
00:51:16 So do you share this worry?
00:51:19 Do you share this concern
00:51:20 when we’re talking about military applications
00:51:23 that too much control and decision making capabilities
00:51:27 giving to software or AI?
00:51:31 Well, I don’t see it happening today.
00:51:34 And in fact, this is something from a policy perspective,
00:51:38 it’s obviously a very dynamic space,
00:51:40 but the Department of Defense has put quite a bit
00:51:42 of thought into that.
00:51:44 And maybe before talking about the policy,
00:51:46 I’ll just talk about some of the why.
00:51:48 And you alluded to it being a sort of a complicated
00:51:52 and a little bit scary world out there,
00:51:54 but there’s some big things happening today.
00:51:57 You hear a lot of talk now about a return
00:51:59 to great powers competition,
00:52:01 particularly around China and Russia with the US,
00:52:05 but there are some other big players out there as well.
00:52:10 And what we’ve seen is the deployment of some very,
00:52:16 I’d say concerning new weapon systems,
00:52:20 particularly with Russia and breaching
00:52:23 some of the IRBM,
00:52:24 Intermediate Range Ballistic Missile Treaties,
00:52:26 that’s been in the news a lot.
00:52:29 The building of islands, artificial islands
00:52:33 in the South China Sea by the Chinese
00:52:35 and then arming those islands.
00:52:38 The annexation of Crimea by Russia,
00:52:42 the invasion of Ukraine.
00:52:44 So there’s some pretty scary things.
00:52:47 And then you add on top of that,
00:52:49 the North Korean threat has certainly not gone away.
00:52:53 There’s a lot going on in the Middle East
00:52:54 with Iran in particular.
00:52:56 And we see this global terrorism threat has not abated.
00:53:02 So there are a lot of reasons to look for technology
00:53:06 to assist with those problems,
00:53:07 whether it’s AI or other technologies like hypersonics,
00:53:11 which we discussed.
00:53:12 So now let me give just a couple of hypotheticals.
00:53:17 So people react sort of in the second timeframe, right?
00:53:24 Photon hitting your eye to movement
00:53:27 is on the order of a few tenths of a second
00:53:30 kinds of processing time.
00:53:34 Roughly speaking,
00:53:36 computers are operating in the nanosecond timescale, right?
00:53:41 So just to bring home what that means,
00:53:44 a nanosecond to a second is like a second to 32 years.
00:53:50 So seconds on the battlefield,
00:53:53 in that sense, literally are lifetimes.
00:53:56 And so if you can bring an autonomous
00:54:01 or AI enabled capability
00:54:03 that will enable the human to shrink,
00:54:05 maybe you’ve heard the term the OODA loop.
00:54:07 So this whole idea that a typical battlefield decision
00:54:12 is characterized by observe.
00:54:15 So information comes in, orient.
00:54:18 How does that, what does that mean in the context?
00:54:21 Decide, what do I do about it?
00:54:23 And then act, take that action.
00:54:25 If you can use these capabilities to compress that OODA loop
00:54:29 to stay inside what your adversary is doing,
00:54:32 that’s an incredible powerful force on the battlefield.
00:54:37 That’s a really nice way to put it,
00:54:39 that the role of AI and computing in general
00:54:41 has a lot to benefit from just decreasing
00:54:45 from 32 years to one second,
00:54:47 as opposed to on the scale of seconds and minutes and hours
00:54:50 making decisions that humans are better at making.
00:54:53 And it actually goes the other way too.
00:54:54 So that’s on the short timescale.
00:54:57 So humans kind of work in the one second,
00:54:59 two seconds to eight hours.
00:55:01 After eight hours, you get tired,
00:55:04 you gotta go to the bathroom, whatever the case might be.
00:55:07 So there’s this whole range of other things.
00:55:09 Think about surveillance and guarding facilities.
00:55:16 Think about moving material, logistics, sustainment.
00:55:20 A lot of these, what they call dull, dirty
00:55:22 and dangerous things that you need
00:55:24 to have sustained activity,
00:55:26 but it’s sort of beyond the length of time
00:55:28 that a human can practically do as well.
00:55:30 So there’s this range of things that are critical
00:55:36 in military and defense applications
00:55:39 that AI and autonomy are particularly well suited to.
00:55:43 Now, the interesting question that you brought up is,
00:55:46 okay, how do you make sure that stays within human control?
00:55:49 So that was the context for now the policy.
00:55:52 And so there is a DOD directive called 3000.09
00:55:56 because that’s the way we name stuff in this world.
00:56:00 But I’d say it’s well worth reading.
00:56:04 It’s only a couple of pages long,
00:56:05 but it makes some key points.
00:56:07 And it’s really around making sure
00:56:09 that there’s human agency and control
00:56:12 over use of semi autonomous and autonomous weapons systems,
00:56:20 making sure that these systems are tested,
00:56:23 verified and evaluated in realistic,
00:56:25 real world type scenarios,
00:56:28 making sure that the people are actually trained
00:56:30 on how to use them,
00:56:31 making sure that the systems have human machine interfaces
00:56:36 that can show what state they’re in
00:56:38 and what kinds of decisions they’re making,
00:56:40 making sure that you’ve established doctrine
00:56:42 and tactics and techniques and procedures
00:56:45 for the use of these kinds of systems.
00:56:48 And so, and by the way, I mean, none of this is easy,
00:56:52 but I’m just trying to lay kind of the picture
00:56:56 of how the US has said,
00:56:58 this is the way we’re gonna treat AI and autonomous systems,
00:57:02 that it’s not a free for all.
00:57:04 And like there are rules of war and rules of engagement
00:57:08 with other kinds of systems,
00:57:09 think chemical weapons, biological weapons,
00:57:12 we need to think about the same sorts of implications.
00:57:15 And this is something that’s really important
00:57:17 for Lockheed Martin.
00:57:18 I mean, obviously we are a hundred percent complying
00:57:21 with our customer and the policies and regulations,
00:57:26 but I mean, AI is an incredible enabler,
00:57:30 say within the walls of Lockheed Martin
00:57:32 in terms of improving production efficiency,
00:57:35 doing helping engineers, doing generative design,
00:57:38 improving logistics, driving down energy costs.
00:57:42 I mean, there are so many applications,
00:57:44 but we’re also very interested in some of the elements
00:57:48 of ethical application within Lockheed Martin.
00:57:51 So we need to make sure that things like privacy
00:57:54 is taken care of, that we do everything we can
00:57:58 to drive out bias in AI enabled kinds of systems,
00:58:03 that we make sure that humans are involved in decisions,
00:58:06 that we’re not just delegating accountability to algorithms.
00:58:10 And so for us, it all comes back,
00:58:13 I talked about culture before,
00:58:14 and it comes back to sort of the Lockheed Martin culture
00:58:17 and our core values.
00:58:19 And so it’s pretty simple for us and do what’s right,
00:58:21 respect others, perform with excellence.
00:58:24 And now how do we tie that back to the ethical principles
00:58:27 will govern how AI is used within Lockheed Martin.
00:58:31 And we actually have a world, pretty,
00:58:34 so you might not know this,
00:58:35 but there are actually awards for ethics programs.
00:58:37 Lockheed Martin’s had a recognized ethics program
00:58:41 for many years.
00:58:42 And this is one of the things that our ethics team
00:58:44 is working with our engineering team on.
00:58:47 One of the miracles to me, perhaps a layman,
00:58:51 again, I was born in the Soviet Union.
00:58:53 So I have echoes, at least in my family history
00:58:57 of World War II and the Cold War.
00:59:00 Do you have a sense of why human civilization
00:59:04 has not destroyed itself through nuclear war,
00:59:07 so nuclear deterrence?
00:59:09 And thinking about the future,
00:59:11 does this technology have a role to play here?
00:59:14 And what is the long term future
00:59:17 of nuclear deterrence look like?
00:59:20 Yeah, this is one of those hard, hard questions.
00:59:24 And I should note that Lockheed Martin is both proud
00:59:29 and privileged to play a part in multiple legs
00:59:31 of our nuclear and strategic deterrent systems
00:59:35 like the Trident submarine launch ballistic missiles.
00:59:42 You talk about, is there still a possibility
00:59:47 that the human race could destroy itself?
00:59:49 I’d say that possibility is real.
00:59:50 But interestingly, in some sense,
00:59:55 I think the strategic deterrence have prevented
01:00:00 the kinds of incredibly destructive world wars
01:00:03 that we saw in the first half of the 20th century.
01:00:07 Now, things have gotten more complicated since that time
01:00:10 and since the Cold War.
01:00:12 It is more of a multipolar great powers world today.
01:00:16 Just to give you an example, back then,
01:00:19 there were, in the Cold War timeframe,
01:00:21 just a handful of nations
01:00:23 that had ballistic missile capability by last count.
01:00:27 And this is a few years old.
01:00:28 There’s over 70 nations today that have that.
01:00:31 Similar kinds of numbers
01:00:33 in terms of space based capabilities.
01:00:38 So the world has gotten more complex and more challenging
01:00:42 and the threats, I think, have proliferated
01:00:45 in ways that we didn’t expect.
01:00:49 The nation today is in the middle of a recapitalization
01:00:53 of our strategic deterrent.
01:00:55 I look at that as one of the most important things
01:00:58 that our nation can do.
01:01:00 What is involved in deterrence?
01:01:01 Is it being ready to attack
01:01:08 or is it the defensive systems that catch attacks?
01:01:11 A little bit of both.
01:01:12 And so it’s a complicated game theoretical kind of program.
01:01:16 But ultimately,
01:01:20 we are trying to prevent the use of any of these weapons.
01:01:24 And the theory behind prevention is that
01:01:29 even if an adversary uses a weapon against you,
01:01:33 you have the capability to essentially strike back
01:01:37 and do harm to them that’s unacceptable.
01:01:40 And so that will deter them from making use
01:01:44 of these weapons systems.
01:01:48 The deterrence calculus has changed, of course,
01:01:50 with more nations now having these kinds of weapons.
01:01:56 But I think from my perspective, it’s very important
01:02:02 to maintain a strategic deterrent.
01:02:05 You have to have systems that you know will work
01:02:08 when they’re required to work.
01:02:11 Now you know that they have to be adaptable
01:02:13 to a variety of different scenarios in today’s world.
01:02:17 And so that’s what this recapitalization of systems
01:02:20 that were built over previous decades,
01:02:23 making sure that they are appropriate, not just for today,
01:02:26 but for the decades to come.
01:02:29 So the other thing I’d really like to note
01:02:32 is strategic deterrence has a very different
01:02:35 character today.
01:02:40 We used to think of weapons of mass destruction
01:02:42 in terms of nuclear, chemical, biological.
01:02:45 And today we have a cyber threat.
01:02:48 We’ve seen examples of the use of cyber weaponry.
01:02:54 And if you think about the possibilities
01:02:58 of using cyber capabilities or an adversary attacking the US
01:03:03 to take out things like critical infrastructure,
01:03:07 electrical grids, water systems,
01:03:11 those are scenarios that are strategic in nature
01:03:16 to the survival of a nation as well.
01:03:19 So that is the kind of world that we live in today.
01:03:23 And part of my hope on this is one that we can also develop
01:03:28 technical or technological systems,
01:03:30 perhaps enabled by AI and autonomy,
01:03:33 that will allow us to contain and to fight back
01:03:38 against these kinds of new threats
01:03:41 that were not conceived when we first developed
01:03:44 our strategic deterrence.
01:03:46 Yeah, I know that Lockheed is involved in cyber,
01:03:48 so I saw that you mentioned that.
01:03:52 It’s an incredibly, nuclear almost seems easier than cyber
01:03:57 because there’s so many attack,
01:03:58 there’s so many ways that cyber can evolve
01:04:01 in such an uncertain future.
01:04:03 But talking about engineering with a mission,
01:04:05 I mean, in this case that you’re engineering systems
01:04:09 that basically save the world.
01:04:13 Well, like I said, we’re privileged to work
01:04:18 on some very challenging problems
01:04:20 for very critical customers here in the US
01:04:23 and with our allies abroad as well.
01:04:25 Lockheed builds both military and nonmilitary systems.
01:04:30 And perhaps the future of Lockheed
01:04:32 may be more in nonmilitary applications
01:04:35 if you talk about space and beyond.
01:04:38 I say that as a preface to a difficult question.
01:04:41 So President Eisenhower in 1961 in his farewell address
01:04:46 talked about the military industrial complex
01:04:48 and that it shouldn’t grow beyond what is needed.
01:04:51 So what are your thoughts on those words,
01:04:55 on the military industrial complex,
01:04:58 on the concern of growth of their developments
01:05:04 beyond what may be needed?
01:05:06 That where it may be needed is a critical phrase, of course.
01:05:12 And I think it is worth pointing out, as you noted,
01:05:15 that Lockheed Martin,
01:05:16 we are in a number of commercial businesses
01:05:19 from energy to space to commercial aircraft.
01:05:24 And so I wouldn’t neglect the importance
01:05:28 of those parts of our business as well.
01:05:32 I think the world is dynamic and there was a time,
01:05:36 and it doesn’t seem that long ago to me,
01:05:38 it was while I was a graduate student here at MIT
01:05:41 and we were talking about the peace dividend
01:05:43 at the end of the Cold War.
01:05:45 If you look at expenditure on military systems
01:05:49 as a fraction of GDP,
01:05:51 we’re far below peak levels of the past.
01:05:55 And to me, at least, it looks like a time
01:05:59 where you’re seeing global threats changing in a way
01:06:02 that would warrant relevant investments
01:06:06 in defensive capabilities.
01:06:10 The other thing I’d note,
01:06:14 for military and defensive systems,
01:06:17 it’s not quite a free market, right?
01:06:21 We don’t sell to people on the street.
01:06:25 And that warrants a very close partnership
01:06:29 between, I’d say, the customers and the people
01:06:33 that design, build, and maintain these systems
01:06:38 because of the very unique nature,
01:06:42 the very difficult requirements,
01:06:44 the very great importance on safety
01:06:49 and on operating the way they’re intended every time.
01:06:54 And so that does create,
01:06:56 and frankly, it’s one of Lockheed Martin’s great strengths
01:06:59 is that we have this expertise built up over many years
01:07:03 in partnership with our customers
01:07:05 to be able to design and build these systems
01:07:08 that meet these very unique mission needs.
01:07:11 Yeah, because building those systems is very costly,
01:07:14 there’s very little room for mistake.
01:07:16 I mean, it’s, yeah, just Ben Rich’s book and so on
01:07:18 just tells the story.
01:07:20 It’s nerve wracking just reading it.
01:07:22 If you’re an engineer, it reads like a thriller.
01:07:24 Okay, let me, let’s go back to space for a second.
01:07:29 I guess.
01:07:30 I’m always happy to go back to space.
01:07:33 So a few quick, maybe out there,
01:07:36 maybe fun questions, maybe a little provocative.
01:07:40 What are your thoughts on the efforts
01:07:43 of the new folks, SpaceX and Elon Musk?
01:07:48 What are your thoughts about what Elon is doing?
01:07:50 Do you see him as competition?
01:07:52 Do you enjoy competition?
01:07:55 What are your thoughts?
01:07:56 Yeah, first of all, certainly Elon,
01:08:00 I’d say SpaceX and some of his other ventures
01:08:03 are definitely a competitive force in the space industry.
01:08:08 And do we like competition?
01:08:09 Yeah, we do.
01:08:11 And we think we’re very strong competitors.
01:08:15 I think it’s, you know, competition is what the US
01:08:19 is founded on in a lot of ways
01:08:22 and always coming up with a better way.
01:08:24 And I think it’s really important
01:08:27 to continue to have fresh eyes coming in, new innovation.
01:08:33 I do think it’s important to have level playing fields.
01:08:35 And so you wanna make sure
01:08:37 that you’re not giving different requirements
01:08:41 to different players.
01:08:42 But, you know, I tell people, you know,
01:08:45 I spent a lot of time at places like MIT.
01:08:47 I’m gonna be at the MIT Beaverwork Summer Institute
01:08:50 over the weekend here.
01:08:52 And I tell people, this is the most exciting time
01:08:55 to be in the space business in my entire life.
01:08:58 And it is this explosion of new capabilities
01:09:03 that have been driven by things like the, you know,
01:09:06 the massive increase in computing power,
01:09:08 things like the massive increase in comms capabilities,
01:09:13 advanced and additive manufacturing
01:09:15 are really bringing down the barriers to entry in this field
01:09:19 and it’s driving just incredible innovation.
01:09:21 And it’s happening at startups,
01:09:23 but it’s also happening at Lockheed Martin.
01:09:25 You may not realize this, but Lockheed Martin,
01:09:27 working with Stanford actually built the first CubeSat
01:09:31 that was launched here out of the US
01:09:33 that was called QuakeSat.
01:09:35 And we did that with Stellar Solutions.
01:09:37 This was right around just after 2000, I guess.
01:09:41 And so we’ve been in that, you know,
01:09:43 from the very beginning.
01:09:45 And, you know, I talked about some of these,
01:09:48 like, you know, Maya and Orion,
01:09:50 but, you know, we’re in the middle of what we call smartsats
01:09:54 and software defined satellites
01:09:55 that can essentially restructure and remap their purpose,
01:10:00 their mission on orbit to give you almost, you know,
01:10:04 unlimited flexibility for these satellites
01:10:06 over their lifetimes.
01:10:08 So those are just a couple of examples,
01:10:10 but yeah, this is a great time to be in space.
01:10:13 Absolutely.
01:10:14 So Wright Brothers flew for the first time 116 years ago.
01:10:20 So now we have supersonic stealth planes
01:10:23 and all the technology we’ve talked about.
01:10:25 What innovations, obviously you can’t predict the future,
01:10:29 but do you see Lockheed in the next 100 years?
01:10:32 If you take that same leap,
01:10:34 how will the world of technology and engineering change?
01:10:37 I know it’s an impossible question,
01:10:39 but nobody could have predicted
01:10:42 that we could even fly 120 years ago.
01:10:45 So what do you think is the edge of possibility
01:10:50 that we’re going to be exploring in the next 100 years?
01:10:52 I don’t know that there is an edge.
01:10:54 I, you know, we’ve been around
01:10:56 for almost that entire time, right?
01:10:58 The Lockheed brothers and Glen L. Martin
01:11:03 starting their companies in the basement of a church
01:11:08 and an old service station.
01:11:11 We’re very different companies today
01:11:14 than we were back then, right?
01:11:15 And that’s because we’ve continuously reinvented ourselves
01:11:19 over all of those decades.
01:11:21 I think it’s fair to say, I know this for sure,
01:11:24 the world of the future, it’s gonna move faster,
01:11:27 it’s gonna be more connected,
01:11:29 it’s gonna be more autonomous,
01:11:31 and it’s gonna be more complex than it is today.
01:11:36 And so this is the world, you know,
01:11:38 as a CTO at Lockheed Martin that I think about,
01:11:40 what are the technologies that we have to invest in?
01:11:42 Whether it’s things like AI and autonomy,
01:11:45 you know, you can think about quantum computing,
01:11:47 which is an area that we’ve invested in
01:11:49 to try to stay ahead of these technological changes,
01:11:53 and frankly, some of the threats that are out there.
01:11:56 I believe that we’re gonna be out there in the solar system,
01:11:59 that we’re gonna be defending and defending well
01:12:02 against probably, you know, military threats
01:12:04 that nobody has even thought about today.
01:12:08 We are going to be, we’re gonna use these capabilities
01:12:12 to have far greater knowledge of our own planet,
01:12:15 the depths of the oceans, you know,
01:12:17 all the way to the upper reaches of the atmosphere
01:12:20 and everything out to the sun
01:12:21 and to the edge of the solar system.
01:12:23 So that’s what I look forward to,
01:12:26 and I’m excited, I mean, just looking ahead
01:12:30 in the next decade or so to the steps
01:12:33 that I see ahead of us in that time.
01:12:36 I don’t think there’s a better place to end,
01:12:38 Keoki, thank you so much.
01:12:39 Lex, it’s been a real pleasure,
01:12:41 and sorry it took so long to get up here,
01:12:43 but I’m glad we were able to make it happen.