Pushkin. If you go to the front line of the war in Ukraine and you look up, there's a very good chance you'll see a drone. Could be a friendly drone, could be an enemy drone, could be a reconnaissance drone, or it could be a drone that is trying to kill you. The war in Ukraine is the first war in history where huge numbers of cheap drones powered largely by commercial grade hardware and open source software, are playing
a central role. This technological change has profound implications. It gives smaller countries like Ukraine new ways to fight, and it means the militaries of big traditional powers military is built on twentieth century technologies like tanks and aircraft carriers, need to radically change just to keep up. I'm Jacob Goldstein, and this is what's your problem. My guest today is
Christopher Kirkoff. Nine years ago, Christopher helped to start the Defense Innovation Unit, a Defense Department office in Silicon Valley. He currently works as an advisor there. Christopher says, in many ways, the US military has failed to keep up with the way military technology has changed. That fundamentally is the problem. DIU is trying to solve how can the giant bureaucracy that is the US military catch up with
the latest military technology. Christopher is a civilian. He had a bunch of national security related jobs in the Obama administration, and he recently co wrote a book about his work at the Defense Innovation Unit, the DIIU. Christopher and I talk about the book and the DIIU later in the interview, but to start, we talked about the weapons and tech being used in the war in Ukraine. So, I mean,
it's in some ways like the war in Ukraine. It is some version of the war or the kind of war you've been thinking about for a long time, years and years.
Well, I've always looked at Ukraine, you know, at sort of two levels. The first, of course, is that it's just a very tragic and frankly needless war. You know, there's there's very few reasons right ever to go to war, and this is a particularly stupid war that's just going to cost a lot of young men and women their lives. So it's tragic, and I'm terribly sorry for the suffering of the Ukrainian people, and frankly the suffering of the young conscripted Russian soldiers as well, So let me just
say that to start. But the second way that I look at Ukraine as you know, essentially a laboratory of war. We saw in you know, a decade ago already the rise of commercial technologies that it was it was obvious
could be weaponized and weaponized at scale. And I mean I remember sitting on the e ring of the Pentagon watching YouTube videos of drones, you know, first of all, quad copters flying around, you know, first sinkly, then in pairs, then doing dances, and it was really easy to imagine that at some point in the future those drones could
themselves also be carrying payloads of various kinds. And not only that, but I mean these are drones that you know, you could order on Amazon, right, they were a few hundred dollars, so they were incredibly inexpensive to begin experimenting with. So all the ingredients were there for a complete revolution in the tools with which which war could be fought, and they came together kind of all at once in a way when Russian armored columns began and advancing in Ukraine. Right.
I mean, it's it's kind of easy to forget now, but it at the time when Russia invaded Ukraine, seems like the basic conventional wisdom was, oh, giant country with a huge army invading, They're going to win in a week. And that didn't happen. I mean, what what role did this sort of new set of technologies play in allowing Ukraine to defend itself.
Well, you know that the technologies actually came into play even before the invasion commenced. And one of the most visible ways that came into play is I mean, this is the first war in a way where there's so many private imagery companies that have their own satellite constellations that are beaming back to Earth high resolution imagery, resolution of the order of what would have been, you know,
a Cold War spy satellite just a few years ago. Uh. And this this importantly, this, these images weren't just going to you know, the US government or to other militaries. They were available to anybody that wanted to see them. And so when Putin and the Kremlin was issuing statements to the effect that no, they weren't going to invade Ukraine, in fact, they were demobilizing the Russian military and preparing
to ship material back to Russia. By rail. The Biden administration was able to release commercial imagery that showed no, in fact, there's armored columns lining up on the border ready to go.
Uh huh. I mean the fact that it was commercial means you didn't even need the Biden administration, right like anybody who paid for a subscription could see that the Russian tanks were there.
Right exactly, and so and this carried on. Then so the invasion begins and the Ukrainians not only have eyes and ears from this commercial data fees, not just imagery, but also radio frequency intelligence, and that gets paired quickly with starlink at a moment when you know, the Russians are using very advanced cyber attacks to attempt to take down not only the Internet in Ukraine, but also the command and control structure of the Ukrainian military. And what
then happened next was really quite unique and amazing. The Ukrainian government had become quite adept at offering digital services to its citizens, both through online websites and through a smartphone app that the government built. In fact, Zelensky actually campaigned as a somebody who would bring in a new generation of digital services, whether it was taxes or you know,
the ability to incorporate online. So the Ukrainian government very quickly reconfigure a smartphone app that Ukrainians used every day in their ordinary lives to allow Ukrainians to be able to send in spot reports of you know, I just saw a tank column moving past my town.
You basically just pull out your phone and open your app and instead of whatever, calling an uber, you say, I just saw a tank.
Well, yeah, so the nickname for this became Uber for artillery. It literally was possible for citizens to file a spot report that went right away to a command center in the Ministry of Defense, and artillery and other weapons were literally dispatched on that information.
So, okay, So that's the sort of initial state of play. And then there's this evolution, right it seems like it's three years from now, a long time. There has been this really intense yeah, evolution, innovation right on both sides. Talk to me about that. How has how has technology driven this sort of rapid evolution of war fighting over the last few years.
Well, Ukraine has been described the Ukrainian military has been described, you know, as a as a military that already came with a silicon valley wing attached to it because Ukraine over the last two decades has redefined itself as a
country of programmers. It has a large endemic IT industry that was there before the war, and the Ukrainian government was really smart in that it mobilized, of course, military age males to serve It really is a citizen army in that respect, but it took anybody that had engineering or programming skills and it allowed them either to work for a private company developing technology or to work for
the military developing technology. And so from the first weeks of the war you had an incredible number of engineers spitballing on how to develop indigenous technology that could take
on the Russians. And this mattered because as soon as the Russians got their advanced electronic warfare systems to the front and move them forward and turn them on even advanced Western drones that either Ukraine already had in this inventory or rushed in and the opening weeks of the war became essentially ineffective because they were not programmed nor did they have the sensors on them to be able to defeat the extremely advanced Russian countermeasures that were immediately rolled out.
So the Russians were essentially jamming the drones in some way. They had some kind of technology that rendered sort of off the shelf or even military grade drones ineffective.
The jamming was so powerful that if you were in Kiev and you say you were on the third or fourth form of a hotel and you tried to order an uber your location on the map in the Uber app, which show up in the middle of the Indian Ocean. Because the GPS jamming was so effective that even in cities, it wasn't possible to pick up a reliable signal that you know, cities that were hundred miles from the front.
So at the front, the electromagnetic spectrum was just flooded with energy and that caused right away, it touched off this technology race to invent drones that could essentially fly
on their own, even through incredibly intense jamming. And so the Ukrainians and it was remarkable to go see when my co author Radshaw and I got a chance to visit Ukraine a year and a half ago, we went to a number of drone companies and they have essentially the Ukrainians in a matter of months, recreated the entire American Cold War precision strike complex that we first got to know so well in the original Gulf War, where we all of a sudden were dropping bombs down chimneys.
So the Ukrainians have an essence, used a combination of open source software and low cost hardware to recreate everything from the function of our a wax planes. These are the airline looking airplanes with the giant raidardomes on top that provide a kind of God's eye view of the battlefield. So they literally have drones whose sole purpose is to keep track of large parts of the battlefield to other smaller surveillance drones that will go take pictures of a
specific area that the Ukrainians want to target. They'll relay that imagery back kind of almost like miniature YouTube spy planes flying and dropping film canisters and whatnot, and they'll quickly be able to then load the images of Russian targets that they want to hit up onto killer drones that will then use image recognition and terrain following software
to go hunt and find the target. It's with fully you know autonomy on fully autonomous routes right that don't require phoning home for them either to navigate or have permission to engage. On top of that, you actually have the Ukrainians using open source software to fuse different sensor streams, so you can have a Ukrainian commander that on a tablet computer in front of them has data feeds from surveillance drones that that are relaying video in real time
to maps of the electronic warfare spectrum. You know. So it's really incredible what they've been able to do for essentially pennies on the dollar of what the American military is spent to recreate the same technological stack.
So you went to Ukraine a year and a half ago or so. You mentioned a couple of the places you visited. One of them is sort of a had been the Ukrainian version of like best Buy right and had been transfer warmed. Tell me about that.
Yeah, it was wild running around the city of Kiev visiting a number of the places that the government had
commandeered for the war effort. And so one of these places was an all literal you know, Ukraine's version of Best Buy, a big box store essentially on the wall were still signed saying you know, televisions, appliances, you know, there were ads for washing machines, you know, on the walls, like walking to the Best Buy, but instead of there being a whole bunch of electronics for you to buy, there were benches, work benches laid out in an assembly
a line style rows and this particular electronics store was
being used to mass produce quad copter Kamakazi drones. So at one end of the assembly line you would start with the frame of the drone and by the time he got to the other end, you would have a completely working drone with communications software ready to be mated with either an anti personnel charge or a shaped charge to go through armor, you know, a couple of kilogram charge that would create these inexpensive mass produced quad copters into quite potent weapons.
When you say a charge, should I think bomb? It's basically a little.
Bomb, right, yeah, larger than a grenade.
Yeah. And then it would so go out to the front and they'd put the bomb on it and they'd go crash it into a Russian tank or something like that.
They sure would. And this goes back to the incredibly novel hybrid warfare that we're seeing in Ukraine, where yes, there are you know, in many respects, the war in Ukraine is a. There is a lot of continuities with the First and the Second World War. There are trench lines, there is artillery, there are tanks and mechanize differenttry that
are moving around. But at the same time in the trenches are drone pilots and they are launching armed drones and sense drones that are watching enemy movements behind enemy lines. And so the battle you know in World War One or World War Two typically happened at the front where forces actually came into contact with one another. Today, Joan Wolfire is pushing the battle behind the front.
It makes the front sort of more porous exactly.
So if you were, example, you know, a Russian armor unit, and you're trying to advance into Ukrainian territory, you have to bring your tanks and your armored personal carriers with your infantry towards the front. Well, the Ukrainian has gone in the habit of using their inexpensive kamikaze drones to loiter and track armored personal carriers, and at the moment at which troops would disembark the armored personal carrier to move towards the front, they would strike. So this you know,
lawn range or sort of short range precisions. Strike has completely changed how the war is being fought. It's being fought unlike any other war in modern times.
So the war has been going on for a few years. Both Russia and Ukraine have clearly been learning and you know, developing new technology at this incredibly rapid pace. How does the war in Ukraine look different now than it did three years ago or two years ago shortly after it started.
Yeah, well, we're you know, back to a place that's a well known place to historians, to military historians, where you know, there was an initial explosion of conflict in the in the confused invasion, where massive Russian armored columns rushed in at multiple points in the country. They were ultimately beaten back to a certain point. Those lines, you know, over time came to freeze. Because each side has very potent precision strike weapons, it's very hard to change those lines.
In other words, the defense in a way has an advantage. It's very very hard to use offensive mass to gain territory, which is why we've seen the battle lines not shift all that much after they settled in the initial months of the year. But what's going on behind them has been this incredible combat through cycles of innovation, who can come up with better electronic countermeasures, who can come up with the next generation of drones that will defeat them?
And so both the Russians and the Ukrainians have been innovating behind the front lines and then bringing that technology forward.
Does the world look different in some way now than it did a couple of years ago? Or is it different on such a subtle level, like the gpsgmming is better and the countermeasures are better, and so the net effect is the same.
I say, it's different in pretty dramatic ways. So today there are such effect of small scale Kamakazi drones that they can effectively hunt individuals. So we are truly today in a black Mere episode where if you're a civilian at a border city near the front, or a soldier at the front, you are in danger of every time you leave the bunker or leave your house literally being pursued by a drone that's trying to kill you.
I mean, how do you survive a kamakazi drone attack?
You essentially have to take cover and so oftentimes you know the sound of a drone buzzing, you know, And this is a challenge, right because on every square mile of the front, they're probably a dozen or more drones in the air at any given time.
So if you're fighting this where there's basically always a drone flying over your head and maybe it's coming to kill you personally, and maybe it's not.
The airspace on both sides of the lines incredibly overrun with drones. I mean they are literally smacking it to each other in the air. It's like flocks of birds at this point. So so no, the changes we're seeing, frankly, are not subtle at all. They're quite dramatic.
So what does the US learn from the war in Ukraine.
It's woken the United States and our allies up in two very profound ways. The first is that, you know, it turns out we had not been stockpiling adequate material for war, whether it was howards to shells or anti tank missiles or sophisticated rockets we had basically you know, ran to the storage cupboard and found it to be almost bare. And that's very alarming. You know, it happened because of course munitions, when the budget gets tired, it's easy to say, oh, we'll buy more next year. So
that's how we ended up in that situation. So we discovered a crisis actually of military readiness on our side, and we discovered a manufacturing base more than that that was not easily able to be mobilized because many factories that produce these weapons had been idle for years or even shuttered. And in a world in which supply chains are threatened globally, if you can't assume the seas are going to be saved for transit, you have to be
able to manufacture this weaponry within your borders. But there was an even more profound wake up call, and let's just go back to the use of Kamakazi drones. We gave the Ukrainians very early in the war and one A one Abraham's battle tanks. These are the most sophisticated battle tank in the world. We've all seen pictures of them. You can be driving down the highway in an Abram's battle tank at seventy miles an hour and you can decide which window of a skyscraper to put around through them.
It's an incredible piece of hardware. And almost all thirty one of those tanks have been destroyed or disabled by Russian Kamakazi drones that each cost no more than a few thousand dollars a piece.
And what's the tank cost the tankst.
Oh gosh, I don't know, twelve or fifteen million dollars or maybe more. So that tells me another military strategist, that if you're an army who bought a lot of tanks, and you presume that your military strength comes from those tanks, you could easily be surprised by an adversary using really an expensive Kama Coze drones. In other words, the era
of manned mechanized warfare is over. In the same way that during the First World War, tanks showed us that the cavalry age was coming to a close, we are now at the end of the era of the battle tank.
So you're sayings, tanks now are about as useful as horses were in nineteen fifteen. Yes, what do we do about that? What does that mean for the United States?
Well, the reason why my cousin I wrote a book all about just how far behind the US military finds itself is that this phenomena of less expensive technology overcoming legacy weapons systems, you know, the sort of building blocks of the Cold War American military isn't true just for tanks, And in fact we're already seeing this in Ukraine, where the Russian black sea fleet is today being kept in port by Ukrainians ingenious use of autonomous boats that are
carrying payloads of high explosives that have rammed in and sunk the largest Russian ship, the Moskova on the black seat. So the naval strategists called the Russian fleet now being a fleet in being, it's a fleet kept in port because of cheap autonomous technology. So, in other words, many of the Americans most powerful militaries systems, aircraft carriers, battle tanks, battleships may now no longer be as effective as they were just a few years ago.
I mean so in terms of changing the equilibrium. Is the effect of essentially autonomous technology drones and then ship drones to give more power to countries with less money. Is that one effect?
Well, there's two effects. So the first and most obvious effect is that a country with fewer financial resources can amass larger degrees of military power. But there's a second and more insidious effect, which is that militaries are inherently conservative institutions for a couple of reasons. One, war is a very unforgiving teacher. And it's almost impossible through just simulations alone to actually replicate combat conditions. And so you can test, you can test new systems all you want,
but you won't actually have the com evident. So they're going to work in the way you suspect them to work or want them to work until you go to war. So this is what leads militaries to almost always perpetually be overfocused on winning the last war rather than winning the war. That's the cop.
Still to come on the show. Why the US military is not ready to fight a war right now according to Chris, and what it might take to change that. So, okay, there is this problem right the US military is not really keeping up with technological change. And this is a problem that Christopher has been working on for a while now.
It's a key part of the reason he was tapped to help launch the Defense Innovation Unit back in twenty sixteen, and to kind of set the scene here right to give a sense of what he and his colleagues were up against, what they were trying to change when they launched the Defense Inovation Unit. I asked Christopher about this story that he and his co author tell in their book, it's the story of something they saw at the US military's Combined Air Operations Center in Katar.
So the Combined Air Operations Center is this nerve center that is in charge of running all air operations in the Middle East in an area that's actually larger than the lower forty eight States. And so any plane, any missile, any air strike that goes on in this extraordinary piece of geography is the responsibility of a set of airmen and women that operate in a base in Qatar. And in twenty sixteen, you'll remember, there was a crisis in
the Middle East. Isis was storming through countries, was encircling the Iraqi population in Mosul, was trying to prosecut the largest genocide of the twenty first century against the Ziti people who were trapped on Mountain Jar. And these civilians were kept alive only by an extraordinary campaign of US air strikes. We were striking ISIS forces every few minutes twenty four to seven for weeks on end, in a
bombing campaign that's the highest intensity in history. And to sustain a bombing campaign of this you have to have, of course fighter aircraft and surveillance, but you also have to have refueling tankers. These fighter jets are gas guzzlers. They need to refill their gas tanks every forty minutes or so, and to do that they have to mate
up with air refueling tankers. And this is actually a very complex thing to figure out because fighters come in different types, they refuel at different altitudes, they have different nozzles on them, and if you have hundreds of aircraft a day aloft operating in combat support roles, you need dozens of tankers to refuel them. So how did the Air Force men and women in uniform and guitar orchestrate the complex planning of tanker routes. Well, they used a whiteboard.
We showed up in twenty sixteen and found people essentially operating like we did in World War Two, except instead of having a chalkboard and magnets, we had a whiteboard. And they were having to do all these calculations by hand, which took an enormous number of hours. Why by hand because the computer program that was supposed to do this
had broken and they were waiting for an upgrade. But they had been waiting seven years for an upgrade that one of the large defense companies was in charge of developing that was way of rebudget and way behind schedule.
They'd been waiting seven years for an upgrade, and they were in the middle of this like multi multi million dollar command center the US had built, coordinating hundreds of flights all the time.
It gets worse. So the minute that there's a deviation from the plan, whether that's bad weather or you know, unexpected combat operations in a different area than you'd plan for now you have to redo the calculations on the fly. So this is actually introducing a lot of unneeded risk in our combat operations that that only puts it risk the men and women of the US military, but also puts it risk the forces on the ground that we're trying to protect, where the civilians were trying to defend.
You describe like somewhere in the process, there'd be like somebody with one spreadsheet and then there was another spreadsheet, but like the two spreadsheets couldn't like talk to each other, right, so somebody had to read out the numbers to get it from one spreadsheet to the other. You describe extra refueling flights all the time because they couldn't, you know, process the data fast enough essentially by hand.
There, they had multiple different databases they had to manually fuel fuse together. Somebody would literally stand behind the computer operator to make sure they were typing in the numbers right, and it would take fifty person hours to put together the refueling plan for the next day. So we realized this was an opportunity to show what Silicon Valley does best, which is rapid iteration using software to create an optimization algorithm that could solve this in seconds.
Why hadn't they just fixed it? Let me ask the naive question, like, it's not as a technical matter, it's not a hard problem. What was the actual problem.
So the problem is that the way the military bias technology and the way Silicon Valley makes it, it's like Mars and Venus, they're on two different planets. So the military way, it's called a requirements based system of developing technology. So a bunch of people sit in a windowless room and they imagine, if I were to need a piece of technology to help with air refueling, what should it look like. And they write a bunch of specifications down
that they think would be great. So it's drawn up in the abstract and then because contracting takes a lot in the military. Oftentimes multiple generations of technology come along before that system ever gets built, but of course it has to be built through the original specifications.
So the thing you get is basically a thing somebody designed ten years ago or more, which is so obviously a bad idea, right, And I mean certainly you make a compelling case in the book that it's obviously a bad idea. Is there any counter argument? Is there any case for the way they do it?
There is, And we have to be very sympathetic to the people that are a part of the system because the system actually was built to do something a little bit different, and that is, you know, there aren't too many companies that can build a nuclear powered aircraft carrier or a submarine or a fifth generation stuff fighter, and so you can't exactly go on Amazon and comparison price shop if the're you're the US govern.
There is no market. There is no market for a nuclear sub in a conventional sense. Yeah.
So the Defense Procurement system and the associated auditing system that governs it is designed to save taxpayer dollar by coming up with exceptionally precise specifications through which things must be built and then a cost accounting system that is meant to make sure that the taxpayer is never overcharged from the one or two companies that can supply this technology. So it was built for a different era, and in terms of building aircraft carriers it actually does a reasonable job.
But commercial technology, there is a market. You don't have to do it that way, and in fact, doing it that way only slows you down.
So okay, so you go, you see this terrible, inefficient, dangerous, expensive technology failure that is happening at this base and cutter, what do you do?
Phone home? So my Cothur Raj got on the phone that night and we immediately fedexed Imax and deployed a few Air Force courders and a few coders from the Silicon Valley out with the Pivotal to Qatar to sit side by side with the airmen and women doing the tanker planning and to build an automated app that they could use eventually to push a button and have this done. And it took just over a million dollars and just over one hundred days to do it.
And you said it saved a million dollars in excess flights in like a couple of days.
Right, So before the Air Force had to actually keep a couple refueling tankers on standby to scramble because because they weren't able to pivot the plan fast enough. And of course, each time you scramble an enormous seven sixty seven sized aircraft full of kerosene, it's energetically expensive. So the tool that we use to optimize air tanker refueling actually has saved the Air Force today something like I
think twenty four million gallons of kerosene a year. So we're talking, you know, a lot of money of savings, right, So this is this is a no brainer. Spend the money, automate it. You get better mission outcomes for lower cost. It's the Silicon Valley way of co developing technology with the users, of iterating on a minimum, minimum viral product and then adding features to that product as you go along.
So this is like an early win in a fairly straightforward way. I mean also worth noting that having the coders go sit with the people who are using the thing is like classic Silicon Valley, right, get close to the user, see what they need, as opposed to like have a guy in Washington ten years ago drop a list of things he thinks people are going to need
ten years from now. Right, there are a couple of new companies selling weapons and technology to the government that seem worth talking about, that seem quite different than the sort of twentieth century giant military contractors, Right, Pallanteer and Anderil. Tell me about those two companies and how they fit with this broader story.
Well, it's important to recognize that the defense contracts we have today, particularly the large ones that we know about, the household names so Lockheed, Martin, Boeing, and so on and so forth, they are really best thought of kind of like utility companies, as highly regulated enterprises that are not like traditional product companies because they have to adhere to this massive system of cost accounting and auditing, this two thousand page re regulation that governs everything they do.
And so in fact, the Pentagon has created these companies and the way they operate, which is extraordinarily inefficient in terms of how modern technology is produced by the commercial sector. So what's so interesting about companies like Anderole, like Palanteer, like shield Ai. They are traditional product companies in the
sense that Google and Microsoft are. They are unburdened by having to erect this massive, separate accounting system to deal with the Department of Defense bureaucracy because they are primarily selling their technology through a completely different kind of contract that we pioneered at DIU that our twenty nine year old employee, Lauren Daily, came up with by seeing up late at night and reading congressional legalation.
Basically a way that the Department of Defense can buy weapons without going through the traditional two thousand page process.
Exactly.
It's so amazing that that's what's driving the technological change right in a way, not surprising but quite interesting, right that this, this the process ends up driving the outcome in a backwards way.
Well, political scientists and sociologists all, you know, are always keen to look at what you know, structures are operating, and there is of course a whole discipline looks at public institutions, which are constrained quite a bit more than public companies. So once it isn't surprised. But what is a surprise is the difference that one individual made in this. So Lauren Daly, you know, the daughter of a tank driver,
somebody who entered the army as an acquisition specialist. As her way of serving the country, figures out that there's a loophole a Congress that a renegae Senate staffer is just inserted in the National Defense Authorization Act that that DAU can drive a truck through, and in a matter of weeks, through some hustling at the Pentagon and the support of thatsh, Carter codified this new method of procuring technology called the Commercial Solutions Offering that essentially all want
us to negotiate with companies for a few weeks very openly, in a very conversational way on commercial terms, so the way they were used to negotiating and then to sign a contract. So this was a game changer. And I am so proud to report that since twenty sixteen to date, about eighty billion dollars of technology acquisition by the departed
Defense has gone through this new mechanism. And in fact, last week Trump Secretary of Defense Pete Hegseth signed out a memo mandating that all software bought by the Department of Defense be acquired using this means, not the old two thousand page, two thousand pageway.
Huh So it's sort of the end of the old pretty clearly bad way of buying software for the military.
Yes, And so you know, there's two things to keep in mind here. You know, one, it doesn't make sense to use a two thousand page rule book to buy things where you don't need it, where there's an actual market. But the second thing that's really important. We would never go to Microsoft or Amazon and ask them to build an aircraft carrier. So why were we going to Raytheon or Lockheed Martin and asking them or north remmen to
build software. It's not their core expertise. So the Pentagon, in the last eight years, thanks in part to its Silicon Valley office, has gotten a lot smarter about how to procure advanced technology. And so fast forward to today. The most exciting contract the Air Force is letting to build up to potentially ten thousand supersonic autonomous drones that
will fight alongside our fifth generation man fighters. That contract went to anderall In General Atomics, not de Boeing, not to Lockheed, not to one of the prime So we're seeing this complete sea change in who is being successful at pulling down the biggest contracts with the most advanced technology come out of the Department today.
So do you think the US is ready to fight a war right now, No, go on.
We are in the unfortunate position of having made major decatal investments in weapons systems that have essentially reached the end of their lifespan.
You mean like tank and aircraft carrier. Yes, not a particular tank or aircraft carrier, but just any tank or aircraft carrier is now basically obsolete.
And we have a three part challenge. We're not yet exactly sure because we haven't experimented enough what technology to buy to replace them. We're also not sure exactly how to use that new technology and what the military calls concepts of operation. So when we invented care aviation, that's a whole new way of using aircraft to fight naval.
Battles, you have to figure out what to do with ten thousand supersonic drones.
And the third problem we have is that once we do figure out what technology to buy and what operating concepts to insert them into, we have to scale that up.
Is there not a fourth problem, which was we're still spending billions and billions of dollars on weapons that you're arguing aren't going to work, are obsolete.
Yes, that is a fourth problem. The Biden administration shifted less than one percent of the Department of Defense's procurement budget towards autonomous.
Weapons less than one percent, meaning basically.
Zero, basically zero.
Yes, I feel like that's a bigger problem than the other three. And I mean that comes through in the book in a way, right, like your antagonist in the book, in a narrative sense, is not Russia or China or isis right. It's bureaucrats. It's it's you know, you go to Capitol Hill and somebody's like, we're going to take your money away because you're not spending any money in Indiana, and that's where my congressman is from. And it seems like that is still a problem.
It is, And you know, this is one of these moments where the strategic environment is changing so fast and anybody watching or reading the news knows it from what's happening in Ukraine, what's happening in the Middleize, what's happening elsewhere. And so we're in this situation where the ordinary way we buy technology and manage the military won't work. If we keep doing it, we're going to walk the plank.
So we're now in a fierce battle to outmaneuver our own system to transform it before there's another common.
Frankly, it's hard for me to imagine the kind of change you're suggesting is needed. Like when I think of how whatever Congress works, and how the country works, and how the military works and how everything works, it's hard for him to imagine. I mean, unless there was a war, which I don't want to happen obviously, right, Like, I don't know, what do you think is going to happen?
Well, you know, to make a more optimistic case, there are a lot of things that are now going in our favor. Most fundamentally, there is a new consensus in the Pentagon and in the Congress that we're in a crisis and the way to solve it is to pivot the way the Department of Defense buys technology and use this technology. And at the same time, there's a new
commitment in industry. We have a very dynamic economy the United States of America, with venture backed investment creating small scale infusions of technology that scales up companies in incredible ways. So since Defense Innovation Unit figure it out this more seamless way to quickly buy technology and actually make the
government a good customer. In Silicon Valley, there's been about two hundred and fifty billion dollars of venture funds now in play to back defense tech startups like and Andro Al's am on the most successful, but there are hundreds more. So the market is already responding. And today there is more dynamic invention going on in the defense tech sector
than I think anywhere else. And so if the consensus in the Pentagon in Congress can actually be changed into shifts in the budget, we will quickly beheaded in the right direction.
I was with you until you said, if consensus in the Pentagon in Congress can be changed, right, that's like there's so much money there and so much inertia, and like the government seems quite bad in many domains at changing when change is needed. That it seems I don't know, you think that might happen. It seems unlikely to me.
Well, you know, again to paint the optimist picture.
Well, just for a sec what do you think? What do you think is going to happen?
I am worried that people now understand the problem, but the pendulum of change is going to swing too slowly. Huh.
So I read that you recently did a residence at Anthropic, which is one of the big frontier AI companies that may claude the large language model, And so I'm curious. This is very interesting to me, Like what did you learn there? Like, what do you know now that you didn't know before you you were there?
The United States is unquestionably right now ahead at least by a little bit in artificial intelligence. So's it's a technology that we're better at than anybody else. And so I think that the sort of sixty four million dollar question in terms of military strategy for the United States is whether or not we can harness AI in designing a new military because right now that is the one thing that will give us an unquestioned advantage in the battlefield.
I mean, more generally, how do you think about AI in the future of war? I mean, what does it mean for the equilibrium? And what does it mean for the sort of relative abilities of big countries and small countries and kind of you know, rogue actors, Like how does it change the power dynamics?
Well, I'm really glad you used the word equilibrium because it's a it's a core concept in strategy. There's a whole literature on something called strategic stability, and it's a literature that comes out in part of the Cold War.
It's a literature that has to reckon with the counterintuitive finding that the proliferation of nuclear weapons otherds, the more countries that became nuclear powers, the less incidents of conflict between them there were, so nuclear weapons, an incredibly dangerous technology, in a strange way, actually made the world safer.
So far, so far.
Now, with this whole true, if everybody had a nuclear weapon, probably not.
Right, And there's also a tail risk where it's like, maybe this was just the lucky whatever eighty years, but go on, yeah, so far, so far.
So I think that the question is what will AI do to strategic stability? Will it actually enhance it in the following sense of AI is clearly going to make warfare much more destructive than it is today, and more destructive systems will raise the cost of going to war. And so if multiple countries have really advanced AI, even if one country is a little bit ahead of the other, that won't change the fact that either one going to
war starting a war will be massively costly. So it could be that AI, as a very diffuse technology, actually increases to deterrence, in effect creating a higher bar to
starting a war. The optimistic scenario, the truly optimistic scenario, is that not only does AI the diffusion of AI technology make great power war less likely, but because there are going to be some bad actors that get a hold of this stuff and go cause a RUCKUS, that's actually going to increase the interest of nations like the US and China working together to ensure they retain their own sovereignty and moopolying the means of violence to keep
their citizens safe. So it could actually not only decrease the likelihood of great power of war, it could increase international cop ration because we desperately need it if we're all going to stay safe. Now. There is, of course, a darker alternative. One flavor of that darker alternative is that there is a country that believes they have a breakout uh uh you know AI, and that that changes for them the calculus of wanting to go to war because they think that they'll just be able to eke
out enough advantage that that they'll launch an attack. Also, I mean there's another danger too. We have to remember that, you know, Vietnam started because in part the US Navy thought its destroyers were attacked in the Gulf of Talking, leading the you know US Congress to pass the Golf of Tonkin Resolutions, which later led to the deployment of troops in Vietnam. Well, turns out the attack never happened. It was just mass confusion. Uh, And so AI introduces
a lot more uncertainty into the battlefield. That that is a risk, that is that is worth knowing. And then truly dark scenario is that AI allows small grop groups to project lethality in a way that's very hard to control.
What do you what do you mean when you say that?
What means that you could go by or make a weapon for low cost that there is no that there is no effective defense for, so that you could go toe to toe, you know, with with the most advanced police force or homeland security force in military and still cause a bunch of damage. So that is indeed a very legitimate worry. It always has been. AI would just be the latest technological innovation to land in the hands of bad actors that are looking to do bad things.
We'll be back in a minute with the lightning round. Okay, let's finish with the lightning round, which is going to be a little lighter. It's going to be like then the rest of this conversation. I read in a bio of yours online that you have backpacked in over thirty countries, including a trek from Moscow to Singapore overland, and I'm curious about that trip in particular. What was one really bad moment on that trip.
I'll go with food poisoning in eastern Siberia that caused me to hallucinate at a time. I was staying in a Russian farmhouse up a ladder in a loft, so I had to while I was hallucinating, climb down to use the bathroom every few minutes late late one night. But I will say it was an incredible trip.
That wait, I want to get to the incredible trip part. But what was the nature of the hallucination.
It was like profoundly lifelike dreams, and I recognized at the time for you know, what they were, and it was just bizarre to be, you know, on this rural farm in eastern Siberia hallucinating and trying to climb down a ladder. So don't you know crap where I'm sleeping?
Yes, yes, okay, you were going to say how great the trip was, tell me that part.
It was just absolutely astonishing to go from Moscow ultimately all the way to Singapore over a land because we got to watch out the window of a train moving into about twenty miles an hour Europe turned into Asia, and we brought history books along and read them and it was an incredible education and I'm so glad we took that trip. It certainly shaped how I see the world and I'll never forget it.
And is it right that you have like some super top secret clearance.
Well, they haven't told me where the aliens are, so I don't know how high have a clearance. I actually have it.
Across that question off the list. I'm like, I'm curious, what, like, what do you have to do to get a you know, a really high level clearance.
Well you just had to fill out a bunch of paperwork and and take take a drug test.
Uh.
But what I will say is that, you know, holding a security clearance, you know, is not all fun, right, because you you learn dings about the world, and you you see things that you you often can't can't forget. So although I was really thrilled when I first was granted a top secret clearance. There are days now that I wish I never had one.
Huh, because the world is scarier or that than you than you would have thought.
Uh, the world is is scarier and and and more dangerous and and uh people people sometimes do really really awful things.
Yes, anything else you want to talking about, I would.
I would just say to everybody that's listening that that now is an incredible time to be watching the world, because we really are seeing historic things happen almost every day. And anybody that pick us up a news story about Ukraine or about China or about the Middle East, it's sort of like reading the papers. I suppose you know, in the run up to the First World War the Second World War, you just see things that are so striking and that tell you so much about where the
world is going. Uh, that that that it's worth it's worth paying attention.
I got nervous when you said run up to the First World War or Second World War.
Well, I say that quite quite specifically in the sense that you know, warfare is changing remarkably fast, and we've seen a succession of advanced militaries fail, the Russian military, the Israeli military demonstrably failed during the attacks that Moss waged on October seventh, and I think it would be humorous to imagine that our military, the American military, couldn't
fail in a similar way. So that's why I think it's really important to pay attention to what's happening in conflict around the world and to realize that we as Americans now need to pay a lot more attention because we no longer have the margin of error that we're accustomed to in being able to project military power.
Christopher Kierkoff is an advisor at the Defense Innovation Unit. Today's show was produced by Gabriel Hunter Cheng. It was edited by Lyddy Jane Kott and engineered by Sarah Brugier. You can email us at problem at Pushkin dot FM. I'm Jacob Goldstein and we'll be back next week with another episode of What's Your Problem? Doctor A sign instituted the fat FA