M.6: Burning Man

Post 7 of 10 in a series about how cheap drones changed warfare.

Ukraine tests its thermite ‘Dragon Drone’. Picture from CNN.

Click here to for the rest of the series.

Both sides of the drone debate are guilty of bad faith arguments. These ‘fallacy fouls’ aren’t just wasting ink (digits?) talking past each other. They are wasting time, the one resource we definitely aren’t getting any more of.

So, before we continue on and examine the impacts cheap drones are going to have on every domain of conflict, I thought this would be a good place to conduct a controlled burn. I’ve got four strawmen in particular I’d like to torch:

1. [Insert platform here] is no longer relevant because drones

2. The ‘Transparent Battlefield’

3. Robots mean we don’t need people to fight anymore

4. EW (and lasers) will save us

The Tank is Dead. Long Live the Tank.

Back in the Second Nagorno-Karabakh War, there were people quick to proclaim the death of the tank. Of course, there were also people who claimed the tank was dead after 1973’s Yom Kippur War. This strawman is actually pretty easy to put to bed.

We don’t stop using things in the military because they can be killed. We drop kit when it’s been replaced by something that does the job better.

Bullets kill infantry all the time. Have been for over five hundred years. The cost-per-bullet is vanishingly small in comparison to the soldier killed, and yet we still have tons of infantry running around. Hundreds of millions of dead infantry across every war since the 16th century, and plenty more lying on the front in Eastern Ukraine today. Why? Because no one has found a better way to hold and clear ground, nor a more efficient bullet sink, than infantry.

Tanks aren’t much different. We didn’t ditch horses as cavalry because horses could get killed by machine guns. We actually still have horses in the army today. But tanks took over their heavy cavalry jobs because they were better heavy shock troops. Drones do a lot of things, but they aren’t doing heavy cavalry’s job yet. Drones are complimenting the light cavalry job, but tanks weren’t ever really great at that job anyway. Light scouts are using drones, but there are still light scouts — as in the people — out there, and I suspect there will be for some time.

FPVs aren't doing artillery jobs or infantry jobs for that matter. So, they aren’t going to replace them. Disable them? Sure. Kill them? Oh, yeah. But pigeon-holing drones as just replacement tools is missing their value. Cheap drones work with artillery, infantry, and yes, even armor. Combining different combat units into a multiplier effect was well established military strategy before even Alexander the Great’s time.(77) Drones don’t undermine this combined arms strategy, they expand it.

Large caliber artillery and high-end strike platforms still have their unique roles. A drone is not going to move at the speeds of a hypersonic missile, nor reach out to its ranges.¹ Nor can drones deliver the destructive force tube- and rocket-launched artillery can. A 155mm round can, in an instant, strike with the net force of 100 Volkswagen Beetles crashing at highway speeds into a very narrow spot on the earth. Few drones could even lift a single 155mm round, though the Russian’s have adapted their dumb bombs into a ‘wonder weapon’ with a $20,000 GPS kit and a steerable tail.²

Instead, drones and artillery pair very nicely.(78)

Drones have incredible scouting abilities, but also a myriad other uses. The boom in diverse designs on today’s battlefields has led to comparisons with the Cambrian Explosion. While large drones have been able to reach hundreds of kilometers for decades, small cheap ones today consistently reach between 15 and 30km.(79) This increased range of awareness coupled with real-time video from the edge has resulted in dramatically reduced time-to-fire for artillery. Howitzer batteries have cut the time to execute unplanned fire missions in half and counter-battery responses to just 30 seconds.(80)

Replacing tanks is not why drones have changed war’s character, any more than failure to replace a tank is an argument for why war’s character hasn’t changed.

Gladwell’s Razor

The US Army has unfortunately adopted the ‘Transparent Battlefield’ concept. ‘One where almost everything soldiers do or say on the battlefield will be visible, identifiable and trackable by the enemy or the public.’

The problem with the ‘transparent battlefield’ is it’s only mostly true. These days, cheap sensors are everywhere, even in space. High end spy satellites used to be the only way to get high-definition imagery, which was restricted by classification from the public. Today you can just order updated satellite imagery with a credit card.³ As I discussed previously the variety and capability of sensors on the battlefield has exploded, and there’s no indication that trend will reverse as Jevon’s Paradox just keeps lapping us.

But regardless of how many sensors there are, whether the cheap ones we’ve been discussing or the high-end ones like GMTI (Ground Movement Target Indicator), we’ll never be able to see and sense everything. The army knows as much, which is why they walk back ‘transparent battlefield’ with the caveat that this doesn’t mean ‘knowing’ everything.

But Gladwell’s Razor doesn’t care. As in the author Malcolm Gladwell. Gladwell is somewhat notorious for writing very widely read books where he shares another person’s theory. However, he often shaves off important subtleties and critical details, undercutting the original theory. His ‘10,000-hour rule’ was wrong, and probably the worst thing to happen to youth sports in my lifetime.⁴ Gladwell’s Razor is what happens when a good idea gets picked up by a bunch of people who didn’t actually do the reading.

In this case, when the battlefield isn’t fully ‘transparent’ but instead harder to hide on, the drone skeptics can sit back and claim nothing’s changed.

Another critique in this vein with a kernel of truth is ‘there’s too much data’. The proliferation of sensors has dramatically increased the volume of data to sift and sort through. In the wars of Iraq and Afghanistan we often saturated targets for a week with video cameras to determine if an insurgent was on the objective. We also tried to minimize the civilian casualties with these strikes and raids. Parsing over 168 hours of video per target was crushing the available analysts, which is a large part of why we sought AI augmentation with tools like Project Maven.

But while it is indisputable there is more data today than yesterday, and highly likely tomorrow there will be even more, not all data is the same. Some data is hard to parse out from background noise, what we can call ‘quiet’ data. This is the sort of data those analysts were struggling with in Iraq and Afghanistan. At night people just look like monochrome figures on most drone cameras. It can be hard to determine and track who is who. ‘Quiet’ data blends in. Think a single cell phone’s electromagnetic signature in downtown San Francisco.

Not all data is quiet though. Tanks aren’t just loud in person, they are ‘loud’ in the data. When you start layering different sensors over top of one another, a lot of the major military platforms start to stand out just as plainly as they would sitting next to you in highway traffic. The Ghost Army was able to fool the Germans in the Second World War with something as simple as an inflatable tank. Today that balloon also needs to have a way to disturb the ground around it and a way to give off the right thermal signature as well as the right EM footprint. If it doesn’t, it is less and less likely to fool the disparate sensors that are going to find it.

But there’s still value in making them look. AI can struggle here, because AI has a hard time ‘looking away’. Sometimes it can be beaten with simple edge case tricks. Marines in a recent exercise were able to sneak by an AI simply by pulling a Bugs Bunny and wearing a cardboard box on their head.(81) Sensors can still be fooled, both into seeing what’s not there and not seeing what is.

And time is, and will always be, a factor in the fog of war. The less time you have to make a decision, the less time there is to take in that mountain of data. This is, in many ways, the entire reason that McCraven’s theory of relative superiority works.

We can see more of the battlefield than ever before, but we still can’t know all of it. But not being able to see all of it doesn’t mean cheap sensors don’t matter. Cheap sensors are absolutely a challenge on the battlefield, one that’s not going away. But the fog of war is and still will be a thing. Surprise is possible. Deception is essential. In offense or defense, combating enemy sensors needs to be part of a commander’s plan.

The Most Dangerous Demographic

The robot revolution has been prophesized for longer than I’ve been alive, perhaps for all of human history. Almost every ancient culture dreamt up their own version of golems before we even had batteries. Satirists, then science fiction writers, and later even military theorists have all envisioned a world where robots do the dirty work, freeing up humans from having to toil, to bleed, and to die.

And I for one welcome it. ‘Soldier’ might be the only profession in the world that looks forward to the day when robots have taken our jobs. But today is not that day. And neither is tomorrow.

This is in part because, just like with tanks and artillery, it’s not ‘either / or’. It’s ‘yes, and’. Humans and cheap drones pair very well together too. In the USMC anecdote above, those same crayon eaters would quickly correct an AI camera that missed something so obvious. While we’ve been able to get computers to best humans in FPV races, these are typically in controlled environments where the computers get to study and practice the course. Humans are good at improvising on the fly. Edge cases like these are why the US military is investing in ‘centaur’ human machine teams.

And more drones currently means you need more people. Jack Watling found ‘automation does not lead to leaner land forces’.(82) Instead drones need handlers and maintainers. Heavy ground drones can’t be moved by light forces when they break down, so we still need recovery teams.⁵ You need people to update the code, tweak the settings, and equip the munitions. While I’m not arguing you will always need a human to do all these things, we currently live in a world where it is easier to replace a lawyer with AI than a construction worker.

There is also a cold, and possibly unsettling, calculus to keeping humans in the fight. When I was studying terrorism at King’s College, one of our professors offhandedly remarked that the most dangerous demographic in any country was ‘underemployed graduate students’.(83) She wasn’t entirely wrong either. Al Qaeda had more PhDs than any battalion I’ve ever served in.

Underemployed people can be very disruptive for a state, and there’s only so many ‘bread and circuses’ you can throw to distract the masses. So, regardless of how cheap and how lethal drones can get, it’s not likely they will offset the potential return on ‘non-investment’ that comes from some nation states simply throwing a chunk of their population into war’s meat grinder. If one state is willing to use human waves to suck up drones — <cough> Russia — it’s hard to see how others will be able to avoid putting more humans on the line.

Napoleon wasn’t the first person to realize there’s military utility in dressing a chunk of your human capital in uniforms, nor was he the last. Cheap drones haven’t put me into early retirement yet, regardless of how much I’d enjoy it. But that doesn’t mean they haven’t fundamentally changed the way we fight.

Sharks With Frikin’ Lasers

As I mentioned in the last post, EW jamming isn’t going to magically save us. While EW is absolutely a critical component, and something we desperately need to get smart on immediately, it is not a cure-all. EW jamming means pushing out signal. Signal that can be used to call in fire on your own position. Your own forces are also using some portion of the spectrum to communicate. The Red Queen always gets her races.

Lasers may be hugely impactful in combatting cheap drones. Thus far the technology has been ‘a couple years away’ for over a decade.

But directed energy, which includes microwaves, will likely be restricted to defending fixed sites for some time. Mounting these tools onto a small mobile platform that can significantly impact the mass effect across the battlefield is going to require some novel power generation.

And on that day, when we do have cheap portable lasers capable of shooting drones and artillery out of the sky, well somebody’s certainly going to wonder what happens when we point them at other things. Now we’ve got a whole new Red Queen’s race on our hands, don’t we?

Ironman

Instead of wasting time strawmaning, a better alternative is for ‘steelman’ arguments.⁶ Above, I’ve tried to tackle strawmen on both sides in order clear the field for the best case I can for cheap drones.

The best counter ‘steelmen’ arguments I have managed to come up with thus far for why drones haven’t changed the character of war are:

• because they are only transitory,

or

• because they don’t change the way we fight.

We’ve addressed the first argument throughout this series thus far. Cheap compute + cheap manufacturing has made death cheaper too. The trend lines continue relentlessly toward zero and show little to no sign of slowing down.

There are pessimistic outlooks though. Authors such as Pete Zeihan argue we may soon see the collapse of the global economy and the free trade it’s built upon. However, if that does come to pass, and we see another microchip disruption like we did during COVID, this will have an outsized impact on the luxury brands of defense industry. This will actually increase the drive toward cheaper and easier to manufacture drones, as everyone looks for cost savings.

This leaves the second, ‘war never changes’ argument. And this just doesn’t seem to be the case. Even the slow to adapt US military is looking at building out a drone corps. I don’t personally think this is going to be enough. This is because cheap drones impact every formation. Logistics convoys need them. Infantry squads need them. Artillery already loves them. While a drone corps might help centralize the larger, more bespoke drones — the mid-tier, Nike and Toyota brands of drones — the cheap ones are still going to find their way onto every corner and into every pocket of the battlefield. I think it’s hard to argue this proliferation won’t change the way we fight.

Drones aren’t going to replace most major military platforms. But they are going to destroy a lot of them. They aren’t going to eliminate the ‘fog of war’, but they are going to challenge it. Drones will not take human’s place on the fields of battle, but they are going to kill a lot of us. And finding ways to combat them is essential, even if those won’t magically solve the problem.

Military pundits predicting a ‘Revolution in Military Affairs’ have almost as bad a track record as economists predicting recessions. ‘RMA’ has been so overused and abused over just my twenty-year career, I don’t even know what it actually means anymore. This is why I don’t bother using it.

But I do believe drones have changed war’s character. They are ubiquitous, they are lethal, and they are not going away. Jevon’s Paradox predicts we will see them more on tomorrow’s battlefield, not less. There are already more cheap drones on the battlefield than soldiers. We should start figuring out how to adapt now.

1. A hypersonic missile is not not a drone, but for the purposes of this series we are regarding them as different tools.

2. Also not not a drone, but this time pretty cheap. Not sure where this lands on the ‘drone’ v. ‘not a drone’ spectrum, but I’m inclined to include it.

3. Civilian satellite imagery has been around for a long time actually. Two decades ago, Second Lieutenant Davis used Google Earth to plan missions because their color imagery was better than the black and white ‘classified’ stuff the Army gave me.

4. Highly recommend David Epstein’s The Sports Gene for a great take on what really matters.

5. May god forever bless the 88Ms and their eponymous M88.

6. I first came across ‘steelman’ last year in Nate Silver’s newest book On the Edge. The idea is instead of wasting your time tearing down bad faith arguments, you should build the strongest possible version of an opponent’s argument, in particular one you disagree with, and then attack that.