Drone warfare has always been equal parts distressing (it’s impersonal, it’s too easy to leverage, and therefore too easy to abuse) and incredible (ultra-thin fiber as a kite line? Brilliant!) to me. These advances are no different.
That said, I do fret we’re staring down a new age of guerrilla warfare. Drones are cheap, widely available, and increasingly autonomous. Their countermeasures are either impractical for communities (AA Cannons or automated firearms) or costly (jammers, interceptors). The programming can be set-and-forget, meaning operations can be staged months ahead of deployment and make it difficult to find or prevent. The autonomy of target termination specifically raises concerns for the immediate future of violent uprisings, coups, and civil wars.
As an engineer, I am fascinated by it all. As a human, I am horrified that we democratized violence on this scale.
nostrademons · 10h ago
I'm also fascinated by the political implications on state formation, state size, and form of government.
State formation tends to track the relative military effectiveness of large highly-trained standing armies vs. small distributed arms making. The Roman Empire collapsed when they ran out of money to pay their legions. The smaller tribes and kingdoms of the Early Middle Ages unified into the larger kingdoms of the High Middle Ages as the longbow and mounted knight gave the advantage again to large, highly trained standing armies. These collapsed into the city-states of the Rennaissance because the gunpowder musket rendered all the armor of the knights useless. Then the nation-state took over as mechanized arms and airplanes became military weapons, and needed the resources of a large territory to produce them.
It's likely that the drone, being both cheap to produce, easy to use, and extremely lethal to existing weapon systems, will produce a similar political revolution. And it seems tailor-made for smaller political units: drones can lay waste to an invading army, but they suck at power projection because their range is only ~10-20 miles. Might we see a return to city-states as the primary form of political organization? Maybe all the arguments about whether Russia vs. the U.S. vs. China will come out on top are moot, because the very concept of a nation-state will disintegrate, and instead we'll have Beijing vs. Shanghai vs. Shenzhen vs. Moscow vs. Kiev vs. the Bay Area vs. NYC vs. Washington DC? Drones are also ideal for defending shipping lanes, so perhaps we'll see a loose confederation of economically-bound city-states, but each having their own culture and social laws.
heavyset_go · 8h ago
IMO the supply chain for drones requires the stability and resource extraction on the scale of a state. SoCs, radios, optics, batteries etc all require high tech manufacturing.
Guerilla use of drones need off the shelf microcontrollers from somewhere, they aren't fabbing them in their backyard.
klooney · 4h ago
> supply chain for drones requires the stability and resource extraction on the scale of a state
It's all about China. They have the ability to cut off drone production for the rest of the world.
lukan · 10h ago
"It's likely that the drone, being both cheap to produce"
If you can buy the parts, then yes, but producing a complete drone all on your own is not that easy I think.
Still easier than a stealth air plane or a cruise missile, though. So your predictions might come true, because I also see most state armies being really slow to adopt to this new reality.
vasco · 14h ago
> As an engineer, I am fascinated by it all. As a human, I am horrified that we democratized violence on this scale.
How do you even defend against this in a terrorist use case? When a small drone with a grenade or homemade explosive is so accessible? Any Christmas market in central Europe these days is surrounded by car barriers to prevent mass run-overs, but what do you do when soon someone has the idea of dropping some molotov cocktails from drones in public places? Answering my own question I guess you can already throw one manually without a drone. Securing public places is weird, I'm glad it's not my job.
GoatInGrey · 12h ago
It's going to be an analogue of the situation with firearms today where the assailant has an asymmetrical advantage over civilians but not the countermeasure (police).
Also, the US military has been stockpiling kinetic drone countermeasures for about four years now. The idea is you get a hardened, ~11 pound autonomous drone that slams into the target at roughly 90 mph and physically destroys it before returning home. Add on 1-2 year-old US EW technology that now disables autonomous drones (yep, even autonomous drones), and you can establish a very comprehensive defense. The point I'm making here is that the tech is not only possible, but it exists.
Is it perfect? No. Though defense against firearms and explosives today isn't perfect either. Namely because of response times of the countermeasure. So in that sense, we aren't entering a uniquely dangerous situation.
Edit:
I think what will happen is that the first time a UAS is used on civilians, flying drones around population centers will be banned without permit. That way, if a drone is seen flying without a permit, it gets taken down on sight.
ethbr1 · 12h ago
The "unique" part is two-fold.
1. It's cheap.
2. It's anonymous.
The fact that we haven't had more drone terror attacks says more about the technological slowness of terrorists than its infeasibility.
And eventually terrorists catch up. (Probably after the Russia-Ukraine war ends and some skilled people from both sides are unemployed)
It's infeasible to blanket every inch of civilian space with kinetic or EW anti-drone systems.
They may become commonplace at mass events (concerts, parades, gatherings, etc.) but will never cover all soft targets.
And unlike the nearest analog in chemical weapons, drones are dual-use, stable, and easily assembled.
The only reason the 1995 Tokyo Subway attacks [0] weren't worse was because of ineptitude.
Someone could be a quarter as intelligent and successfully fly an FPV drone into a target.
> The fact that we haven't had more drone terror attacks says more about the technological slowness of terrorists than its infeasibility.
Also motivation and incentives. The reason we haven't seen many drone attacks on civilians is that it is far more lucrative to get civilians to buy your product than to kill them, and the companies that actually have the resources to mount a credible drone attack are making a lot more money doing the former.
Terrorism in general has always been far more overhyped than actually a problem - the median number of terrorist deaths per year in the U.S. from 1970-2020 is 4, making your odds of being killed in a terrorist attack significantly lower than being struck by lightning. And the reason is simply that it's deeply irrational. What do you have to gain from killing a random stranger?
ethbr1 · 9h ago
The calculus to make a terrorist is simple
non-terror life opportunity
vs
terror life opportunity
Why you tend to have domestic terror in places with bad economies (and high economic equality). And why beliefs often drive it (religious, political, etc.), by adding righteousness to the terror side of the equation.
But requisite minimum skillset is also a consideration, and that's where drones are dangerous.
We're talking (play videogames and some soldering) instead of (chemistry or biology).
seadan83 · 12h ago
> How do you even defend against this in a terrorist use case?
You can't, and you don't. That is why it is called terrorism. Safety and freedom are sometimes antagonistic goals. This is an example. Terrorism is defended against by not changing society despite the terrorism. It is violence with a political goal, if the politics do not change, the terrorism fails. Not every soft target can (nor should be) hardened, there will always be soft targets.
Thank you for the video! This is so much coming. Plus with thermal cameras too which is even more scary.
On the same topic it reminds me Splinter Cell: Chaos Theory (from 2005!) where an AI called "Masse Kernels" was automatically creating missions sent to different PMCs, managing war assets, supervising the war effort, coms, and the like... Feels also that's coming, at least in some forms for now.
In Ukraine one way that they partially secure bunkers or buildings is by putting up nets around them.
nkmnz · 13h ago
fwiw, I think the car barriers never had the purpose to stop cars, but to show people that at least on some level someone cares about their perceived safety.
forgotTheLast · 11h ago
Those barriers do serve their purpose but it's rarely a terrorist threat. Just do a web search for "drunk driver bollard".
netsharc · 9h ago
Feels chicken-and-eggy.
At the weekend, I was at a small town which had a half a mile long main road blocked off for a market day. They put up bollards to do so.
Chances are, there were 0 persons planning a car attack on it. So there was an element of "We don't think anything's going to happen, but if it were to happen, we're prepared.". A bit like having a fire extinguisher when there's never been a fire.
But would seeing the bollards also have the effect of discouraging the insane people of the idea of driving the car through the crowds the next time a market day is held?
Oppositely, if they didn't put up any barriers, a psychopath seeing this and the realization that cars can be weapons might give them the idea of "I know what I can do for my act of terrorism..."
stego-tech · 13h ago
I mean…you can’t, at least not right now, not for civvies. Let’s consider some of the current countermeasures:
* Flak/Shrapnel/Birdshot: An excellent last-minute defense if you’re calm enough to line up an accurate shot, but data shows that equipping civilians with these sorts of weapons en masse is a bad idea for safety and well-being. That’s a no-go.
* Nets: Popular for defense, but it’s a matter of time before drones adapt by flying under the nets or changing payload to something to dissolve it. A kamikaze drone could also be enough to destroy an opening for more to swarm. In a civilian context, they’re an excellent deterrent for high-population areas, for now, albeit unsightly.
* Buildings: Safest for now, provided the structure is relatively hardened and the windows are secured. But most civilian structures aren’t guarded against explosions or external attacks, and even those that are require a human to vacate it eventually. Once inside however, there’s more options for stopping an attack - for now - like interior netting, small arms with pellets or buckshot, or even lasers to blind the optical sensors. Impractical for civilian deployment at scale, presently, and highly variable.
* Jammers: Good against piloted drones, but as the article points out, the current crop of dev work is geared towards autonomous slaughterbots instead of human decision-making. Jammers are restricted by most countries and, if left functioning after an attack, could hinder first responders. If left on constantly, would disrupt civilian work. So that’s a no-go.
* LASERS! Probably the best deterrent in the short term for civilians, I would wager. A randomized strobe of a high-powered IR laser could devastate a swarm of drones’ optics, making navigation or target acquisition difficult or impossible. Sticking a piece of protective glass on the sensor would likely nullify it long enough to finish its mission, though.
And that’s what distresses me, ultimately. The future depicted in Slaughterbots or Horizon is rapidly approaching, where autonomous drones can murder with impunity and are affordable enough that any threat actor could get their hands on it. Combined with modern databases of humans - faces, biometrics, profiles, locations, habits, schedules - we’re nearing an era where assassination or murder is a drone away.
That is what horrifies me. And if there’s one thing my time in the defense industry taught me, it’s that nobody is trustworthy with that kind of power. Companies making these absolutely will use them (or condone their use) against dissidents, opposition, regulators, and governments. Pandora’s Box is already open, and I don’t think enough folks appreciate the horrors it will bring.
southernplaces7 · 13h ago
>Nets: Popular for defense, but it’s a matter of time before drones adapt by flying under the nets or changing payload to something to dissolve it.
Worth noting here that the Ukrainian armed forces have already repeatedly deployed drones with the ability to spray pretty impressive amounts of napalm all over their targets from fairly high altitudes. Fittingly, they've been called "Dragon drones", and I wouldn't want to be under any anti-drone net if one of those arrives.
The "Dragon Drones" also drop thermite [0], a material with massively exothermic high temperature 2000°C+ redox reaction, which can melt through most metals. Thermite is often used to weld rail tracks [1].
For civvies the protection is the same as stopping someone shooting you with a rifle - have the cops jail the perpetrator.
varjag · 13h ago
The perp can be a thousand miles away tho.
im3w1l · 13h ago
Shooting people with a rifle is not scalable and there is a high risk of getting caught. Imagine if school shooters could kill thousands instead of dozens, and at zero risk.
01HNNWZ0MV43FF · 13h ago
Oof. Bad news for people like me who cops might not like
dyauspitr · 10h ago
Well you can’t just throw a Molotov because in that case you’re going to get caught. You can place a drone, go home and have it drop the Molotov the next day and no one would know who did it.
NoMoreNicksLeft · 14h ago
>How do you even defend against this in a terrorist use case? When a small drone with a grenade or homemade explosive is so accessible?
I can see this from the other perspective. I've read stories for the past 30 years now about police forces and swat teams abusing people, murdering people because someone filled out the address wrong on a warrant, etc. And I wonder, in the coming years, if those sorts of scenarios will be quite as one-sided as they are today. What will that world look like?
>Answering my own question I guess you can already throw one manually without a drone.
Sure, but can you throw one from 2 miles away, can you throw 30 simultaneously? Can you then instantly escape without much of any evidence of who you are being left behind, not even a blurry traffic cam picture? The scale of the mayhem is a quality all of its own.
AnimalMuppet · 13h ago
Speaking just for myself, I am not currently stockpiling explosive-carrying drones in preparation for an encounter with over-zealous police. Yes, such incidents happen. They could even happen to me. Still, I don't worry about them enough to go to that length. (Nor do I see things ending well for me if I did.)
fancyfredbot · 11h ago
Give it a go and you might be surprised! Stockpiling explosive-carrying drones in preparation for an encounter with over-zealous police can be an educational and absorbing hobby. Many people are sceptical of how things will end for them, but once they start they are drawn in by the friendly and supportive community and doubts like this are quickly forgotten.
BobaFloutist · 11h ago
Also, the fear of drones is their offensive use. I'm not sure a cop intending to officer-involved-shooting you for looking vaguely like a shoplifter is going to clearly announce his intentions, then wait for you to dig your weaponized drone out of the trunk of your car, boot it up, and get him with it.
NoMoreNicksLeft · 10h ago
>Also, the fear of drones is their offensive use.
Whatever fear Russians have of drones isn't in their offensive use. They're only being used defensively by Ukraine. Similarly, if you're a jackbooted goon, you might well fear their defensive use as I outlined above. But yeh, you have nothing to fear from the defensive use of drones, because you're not attacking anyone... in the coming years, however, the narrative will likely be twisted so that you do come to fear such people as might use them defensively, because the establishment needs you to despise them.
NoMoreNicksLeft · 10h ago
>Still, I don't worry about them enough
MoonGhost · 6h ago
> jammers
Jammers don't work against optical cable or AI vision controlled drones. That's a big problem today in Ukraine for both sides.
As for defense, first of all it's detection and tracking. Copters and long range gas powered drones are very loud and easily detectable. Ukraine uses a net of cell phones. Several devices with microphones can accurately pinpoint all drone like sources in real time. That's cheap to install miles around important targets. Then we need just fast AI interceptors 'on hold', in the air if can afford. The last part is missing today, but we'll get there soon.
As for danger, etc. Small remote controlled firearms were easily available for decades. Drones _are_ trackable. When one takes off in big city Russians know immediately where. By using radio scanners. All DJI drones, and most others, communicate and simply broadcast their coordinates. This is used in Ukraine to find their operators.
myflash13 · 14h ago
This article did not even cover some of the weird solutions being deployed against drones. For example, Russia has surrounded many of their critical infrastructure sites with huge nets (similar to golf course barrier netting). They have also developed anti-drone drones that drop nets from above, catching and tangling target drones in a bunch of netting that simply snags the blades.
sreekanth850 · 4h ago
In my opinion, AA guns will become again popular to counter drone swarms. None of the modern weapons are that effective like old AA Guns.India proved this with Pakisthan massive drone attacks recently.
rollcat · 13h ago
> They have also developed anti-drone drones that drop nets from above, catching and tangling target drones in a bunch of netting that simply snags the blades.
Yeah if we could just all agree that from now on, all warfare is limited to drone-on-drone engagements.
vasco · 12h ago
If we could agree on that we'd agree on simply simulating war, no need to actually send drones. But it's unlikely to happen for the same reason hero fights instead of army fights didn't happen like in the movies. Ultimately war is about a difference of opinion strong enough to not care about loss of life, so after a country would lose the simulation they'd just go onto the real war part anyway.
myflash13 · 31m ago
Actually simulated war does happen. Most of the Western world agreed to fight economic and culture wars since WW2 instead of actual killing.
reaperducer · 11h ago
If we could agree on that we'd agree on simply simulating war, no need to actually send drones.
That was an episode of Star Trek.
Two warring cultures simulated war. The computers told each side how many casualties there were, and people reported to the extermination chambers.
Y_Y · 14h ago
> The programming can be set-and-forget, meaning operations can be staged months ahead of deployment and make it difficult to find or prevent.
This is something I haven't considered before. What's the worst case here? Is it feasible for me to go live on a farm in <country I want to harm>, buy a fleet of DJI drones at flea markets etc, stick something harmful to them, then hide them in the woods.
I can move away, wait a year or two, and then have them fly to the nearest metro area and wreak havoc. This seems to be cheap and relatively straightforward, and hard to detect. What am I missing?
wongarsu · 13h ago
In a pre-drone world you can get explosives, divide them into X equal sized packages, add a timer set to the same point in time to each, then travel around the country hiding them in high-traffic areas.
Yes, that approach is inferior to the drone version. You have to hide them inconspicuously, and a bomb sniffing dog could find them. But you can visit a lot of places in a single European country or US state within one day, and unless the country is already on high alert you can hide something for that time span in public. Yet this doesn't happen. Even regular bombings are rare.
The reasons are manifold: In most places getting explosives isn't actually all that easy (unless you go the homemade route) and is a good method to get attention from authorities. But another factor is that there just doesn't seem to be a large interest in doing that kind of complex attack unless there is already an ongoing civil war. Actual terrorism is fairly rare, and the terrorists tend to be not all that sophisticated.
Are these kinds of drone attacks a scary new possibility? Yes, absolutely! Are they likely to happen? Not really. We might see it as a method to assassinate officials (imagine staging drones at a place where you know the US President will hold a speech in a couple months), but I doubt it will play a major role against the general population
impossiblefork · 10h ago
Attentive people would find the bombs. Especially if they're on alert, which everyone who knows about terrorism will be at least to some degree.
Back in the day, if you forgot a bag on a British bus the driver would get it and run after you, so that it wouldn't be a bomb issue taking the whole day.
adammarples · 12h ago
The drone scenario is very different, all you have to do is get close-ish to your target. This is very different than having a stick of c4 sit on your target for months. You have to get them to drone-range, say in the back of a lorry, on a roof somewhere, in a box. Then the programming and or AI can kick in and do the last mile for you, whenever you need. Case in point is the Ukrainian attack on the air fields, they parked a lorry nearby.
rurp · 14h ago
Not much. Autonomous targeting and control are quite new and currently take a fair bit of knowledge and skill to get right, but I expect those barriers to lower dramatically in the coming years. There might be power issues with such a long delay, but I'm not sure. I think the main drawback once this tech gets slightly more widespread is that most people who want to terrorize cities don't want to wait a year or two to do it.
For long running conflicts (Israel vs Iran for example) I expect we'll see some fascinating and horrifying attacks in the near to medium term. Of course anti-drone tech is also evolving quickly and I expect that to continue so the shelf life of any specific attack will probably continue to be rather short.
stavros · 1h ago
It doesn't really take much knowledge to set up an autonomous drone mission. It's not DJI-levels of consumer friendliness, but I know multiple people that made and fly their own drones, and it's not something you can't do with a few YouTube tutorials.
lapetitejort · 11h ago
What you're missing is the will to go through with it. Even state actors would get spooked spending months or years setting this up. One slip-up and you're in prison for life. When your country's existence is at stake, the process is easier.
heavyset_go · 8h ago
> What am I missing?
Battery degradation, a year or two's worth of leaves and debris accumulating on and around the drones, literally all of the elements affecting them, animals, etc.
stavros · 1h ago
If I've gone to the hassle of setting up a high tech bombing attack, wouldn't I have also gone to the trouble of putting the drone in a self-opening container?
maxglute · 13h ago
You can probably get unsuspecting couriers to deliver last mile. Build a drone dock with cutter to open top of boxes. Mail a weaponized drone to an address within range of target.
NoMoreNicksLeft · 13h ago
They're battery operated, so I think there's a time limit of a few months. Then, you want to be very careful with the infrastructure you leave behind (little pop-open doors/roofs for them to fly out of) to avoid future investigation. You're going to need some practice, your first try will just go to shit perhaps. Opsec while you're setting all this up is still a big deal that amateurs will have trouble getting right. But none of these are particularly insurmountable. With the correct software and careful planning, this will succeed at its goals.
Things you can't help: they will discover the remains of the drones, and also their origin. This evidence will eventually lead back to you (unless you have the aid of a enemy nation-state). Not a big deal if you're dying in a suicide attack, but maybe you don't want the extended vacation in the CIA's worst black ops rendition site.
krunck · 14h ago
What is also distressing is that drones make false flag attacks even easier. Add to that the fact of AI generated media/propaganda means no war will be factually comprehensible to anybody.
tim333 · 13h ago
>horrified that we democratized violence on this scale
Violence has always been pretty democratic - you've always been able to punch someone or hit them with a rock and the US seems to have more guns than people.
ulnarkressty · 13h ago
Pandora's box is now open and multiple groups have access to drones[0].
This is something that I think escapes engineers in this line of work - that something they invented will eventually end up (legally or not) in the hands of people with no scruples.
I think watch out for Mexico. The cartels are swimming in cash and are already using drones to attack security forces and assassinate people.
agumonkey · 13h ago
When you blend mosquitos and terminators..
drdrey · 10h ago
are jammers really that expensive? They shouldn't be particularly sophisticated
bufferoverflow · 8h ago
Doesn't work on fiberoptic drones. Doesn't work on drones that use laser/optical comms.
nadermx · 14h ago
I mean if war has existed since the dawn of man. Maybe we are trying to fix the problem incorrectly.
nradov · 14h ago
War has (probably) existed since before the dawn of man. Even chimpanzees engage in something like tribal warfare so that behavior probably goes back to a common ancestor species at least 6M years ago.
Yeul · 10h ago
I think drones are fantastic. Why throw good money after bad?
Dutch soldier lives have been ruined because they had to be sent to places like Lebanon and Bosnia. Nobody decent deserves that.
Teever · 13h ago
I'm very interested in the consequences of vigilante applications of this kind of technologies. Imagine a scenario where people start taking out corrupt police officers who have until now been able to terrorize small communities with impunity.
It may motivate actual reform in policing because law enforcement will realize that police officers who kill innocent people with no regard for the law are safer in prison than out on the streets with a paid vacation / desk job punishment.
HeyLaughingBoy · 11h ago
You don't need drones to do that.
Teever · 11h ago
You need drones to do it in a way that significantly reduces the risk of getting caught.
As it is vigilante action against law enforcement in the west is a sure death sentence and probably life long reprisal against your family once you're dead which is what keeps people in line.
If the development of drone technology significantly reduces the risk of that then you're likely to see many more people respond to violent abuses of authority by law enforcement with vigilante action.
UncleOxidant · 14h ago
> or costly (jammers,
Are jammers really that costly?
Also spoofers that could take over a drone - not sure how much encryption is used in most of these off-the-shelf drones, but it would seem like it wouldn't be too difficult to create a Flipper Zero-type device that could spoof the codes used between controller and drone.
Ciantic · 19h ago
The latest attack deep inside Russian borders were apparently using ArduPilot [1], it is mentioned in the Atlantic article [2]. ArduPilot also has C++ source code in Github [3], also adding an article specifically about ArduPilot and Ukraine [4]
Apparently, there were hundred+ drone pilots available who were spread across several timezones should the AI not be able to find their way visually. If you see some of the streaming video snippets, it says GPS fix unavailable likely due to jamming of GNSS systems and disabling of civilian GLONASS near bases. As such, Russian domestic cell carriers were reportedly used to stream video and for manual terminal guidance following visual cues, hence the daylight timing of the raid.
throwoutway · 5h ago
Source on the 100+ drone pilots an cell carriers? I've been hoping to find an article on how they staged the pilots for manual guidance. I imagined they would have had to tunnelled all that traffic to each operator from the truck, but the cell towers is clever
ZhiqiangWang · 16h ago
Andrew Tridgell, from rsync/samba to drone ....
aiiizzz · 39m ago
What goes on in a person's mind when they pivot to making killing machines?
ajross · 17h ago
To be clear, that's because Ardupilot is a really pretty routine UAV navigation project with obvious civilian/industrial/enthusiast application. People hack on Ardupilot because it's fun to see your drone fly around on its own. That you can also put a grenade on it is sort of an obvious extension, but the payload is very much a separate feature.
masom · 17h ago
I remember NodeCopter and running OpenCV to control them years ago.
The big difference is they can now run this on the copter instead of being remotely controlled; a 100$ raspberry pi has enough processing power for this, and so does several other off-the-shelf mini computers powered by lithium batteries.
Crazy times.
ivape · 17h ago
So what did they do, just stick GPS coordinates and the drones were that successful autonomously (talking of the Russian fighter base attack from the last 48 hours)? I'd be shocked if they didn't manually pilot these things.
faitswulff · 17h ago
You are correct:
> ...each of the 117 drones launched had its own pilot.
They need to make a movie behind this one, that just looks like such a cyberpunk operation.
consumer451 · 14h ago
There are so many possible scripts that could be based on what has happened since 2022 in Ukraine. There would be no need for exaggeration of the heroism, bravery, and loss. The only issue might be that people would not believe it actually happened.
From Zelenskyy, a previous comedic actor refusing to flee, "I need ammunition, not a ride," to the defense of Snake Island "Russian warship, go fuck yourself," to all the brave women who volunteered, the farmers towing abandoned Russian tanks, the constant drone attacks on residential and commercial areas, the 40,000 stolen Ukrainian children, this most recent attack on Russian air bases...
If this was a movie, I would probably think it was a bit much myself, but this all happened. We witnessed it.
"Glory to Ukraine! Glory to the heroes!"
irthomasthomas · 14h ago
An actor who's last role was the lead in a sitcom about an ordinary citizen who becomes the unlikely president of Ukraine.
HeyLaughingBoy · 11h ago
... is not particularly more surprising than the US president who starred in "Bedtime for Bonzo"
OpenCV and other onboard computer softwares can be trained to recognize shapes, 10+ years ago there was a demo of a NodeCopter controlled small drone following red flags.
Stick the GPS coord, fly there, and once in a geofence look for a shape to crash into doesn't seem impossible given what was possible 10 years ago.
spacedcowboy · 14h ago
Hell, 30 years ago I was working for the MOD (they sponsored my PhD and turned it into an RA) in the UK creating context-aware neural network inference engines for FLIR (Forward-looking infra-red) data. We had all sorts of "fun" stuff running on a Meiko computing surface, with parallelised network training and implementations, temporal and spatial averaging, and relaxation labelling all thrown into the mix to aid the recognition engine, done with a voting system of various architectures sharing to a "blackboard" where information could be posted to and read from. Visualisation was all on high-end (for the time) Silicon Graphics workstations.
The context (together with the features extracted) was the killer (forgive the pun) feature though - everything else reduced noise, but context increased signal.
My gast remains flabbered that the sort of thing I was working on back then hasn't become commonplace in the interim. The computing power available today, compared to then, and the accuracy we had (I know for a fact at least one of the designs was made into real hardware, it was called RH7, and "RH" stood for "Red Herring" - oh how we laughed) ... It beggars belief that it was just left to digitally rot.
m4rtink · 16h ago
There is often quite heavy GPS jamming or spoofing. Also in some of the published videos I think you can see a "no GPS lock" status message - but maybe they just did not bother with GPS if all the drones were manually piloted anyway.
mrandish · 14h ago
Yes, I assumed they didn't need GPS because they knew exactly where the trucks that were the launch sites were to be placed and they knew approximately their targets would be sitting on a certain section of airport tarmac. The pilots would have had a detailed satellite photo map of their entire route until visual target ID was possible. While GPS was probably partially jammed, that deep in Russia I doubt it would have been as severe as near the front lines. Plus there wouldn't have been heavy jamming of the local drone control frequencies because they weren't expecting a drone attack there.
To me the more interesting question is how they managed sending the real-time video feeds and control data. Since the trucks were mobile, I assume it had to be via a bunch of mobile phones signed up to Russian service providers since Starlink doesn't work inside Russia. To reduce latency, I wonder if the phones were connecting to a covert site in Russia which had a high-bandwidth wired link, maybe a front company established for the operation with servers and broadband internet connections.
01HNNWZ0MV43FF · 12h ago
I assume drone guidance builds on missile guidance. Old cruise missiles were loaded with mission-specific terrain maps to follow to their targets
luma · 16h ago
GPS is heavily jammed throughout Russia, and the ArduPilot overlays shown on the videos released directly show there was no GPS lock (might not have been equipped as I'm sure they'd be expecting this).
0cf8612b2e1e · 16h ago
Given these are static targets, it might be possible to relay precise GPS information from that morning’s satellite data. No real time intelligence required. Just dead reckoning to the target coordinates.
Sohcahtoa82 · 16h ago
Dead reckoning in a drone would be a nearly impossible feat considering how variable wind can be.
jdietrich · 13h ago
Operation in GNSS-denied areas is already a stock feature on many relatively inexpensive commercial drones. The manufacturers talk about it euphemistically for obvious reasons, but they're designing drones specifically for the Ukraine war. There's a huge amount of engineering effort going into building drones that can remain operational in an extremely hostile RF environment.
Compensating for wind drift is a fairly straightforward software problem when you've got a fast processor, a bunch of high-resolution cameras and a laser rangefinder.
If you have a downward facing camera you can track your movement like an optical mouse by just watching the terrain. Error will creep in, but you only need to fly a few kilometers till you find something that looks like a strategic bomber.
Sohcahtoa82 · 13h ago
I'm dumb. I don't know why I didn't think about the cameras being used to maintain location.
eloisius · 15h ago
Dead reckoning with error correction using known landmarks like highways, maybe
0cf8612b2e1e · 15h ago
The wrong term, but I know there has been extensive research in maintaining accuracy without GPS.
bamboozled · 19h ago
"He said that each of the 117 drones launched had its own pilot."
He could be telling the truth, he could be lying... A drone programmed to automatically boot up , check its location, and if it's at the right coordinates, take off and crash at some other coordinates (the airfield) is more satisfying to "fans" of automated warfare.
For extra fun, add some other code to "look for plane-like objects to crash into", but now you're approaching dangerous territory of "What if a civilian 737 happens to be boarding at this airfield"...
The reports also mention the truck roof opening remotely, one could also use GPS coordinates to trigger this. But doing it manually from a distance, after checking the surveillance cameras that the coast is clear, is more reliable.
I guess they used smartphones and SIM cards with mobile data for the remote communication...
wiseowise · 18h ago
> For extra fun, add some other code to "look for plane-like objects to crash into", but now you're approaching dangerous territory of "What if a civilian 737 happens to be boarding at this airfield"...
Civilian 737 boarding airfield where Russia keeps strategic nuclear bombings? Russians would shoot them down faster than any drone could get them.
cmcaleer · 15h ago
>Civilian 737 boarding airfield where Russia keeps strategic nuclear bombings?
I'm not sure if the Tu-95 is hosted at any joint-use airports, but joint-use themselves airports are not uncommon. Pskov is joint-use, Ukraine launched a smaller-scale attack on some Il-76s there a couple years back. The scenario that an attack on legitimate target aircraft could be happening metres away from civilian aircraft is realistic.
oneshtein · 14h ago
It's not a problem when Russia, Israel, USA, or other nations attacks civilians, because of a military reason. Why it will be a problem in this case?
BobaFloutist · 11h ago
Because whether or not they're justified in doing so, Ukraine has made it clear that they have no interest in targeting civilians. They've been incredibly surgical and precise in their attacks. Unbelievably so, honestly.
dzhiurgis · 6h ago
Whichever way this war ends you can be sure there will be lone Ukrainian terror attacks on Russia for next 50 years. People will not forgive.
bamboozled · 12h ago
Because the makeup men don’t like Ukraine
Muromec · 18h ago
It's a military airfield, so no civilian 737 there. There seems to be a video from the drone, meaning some kind of connectivity was present with or without autopilot.
m4rtink · 16h ago
Reportedly it was just running over local mobile internet connectivity. The attack was over so quickly they would likely not even have time to shut it down.
pjc50 · 16h ago
> What if a civilian 737 happens to be boarding at this airfield
Good old non-AI radar-guided missile launched by human crew of Russians.
2OEH8eoCRo0 · 18h ago
They did more than crash at coordinates, they targeted specific parts of the different aircraft.
keepamovin · 18h ago
It's strange - blocking GPS is typical around military sites. So, assume the drones were hard-coded to zero a location - they couldn't do it, as GPS would be blocked. They had to be piloted. Interesting.
XorNot · 18h ago
Inertial navigation is a thing though. Your smart phone has a very capable accelerometer for a short flight.
gosub100 · 15h ago
In addition to the other reply, it might have been done with CV. Identify landmarks to get bearings, drop bomb on thing that looks like airplane.
cma · 14h ago
Saw something saying they trained them on old Soviet planes at a military museum
Only need one camera drone capable of identifying targets. And just tells another drone to bomb it.
shrubble · 16h ago
Or can “paint it” with an infrared laser point and then the drone can use simple sensors to guide itself to the target.
XorNot · 18h ago
Cameras are incredibly cheap though. There's basically no reason not to put them on everything.
4gotunameagain · 18h ago
..in theory. In practice, building such a complex system in a fail safe way is not that easy.
gosub100 · 15h ago
"in a fail safe way" goes out the window in wartime.
4gotunameagain · 3h ago
I meant safe for the mission, not for any innocent souls around..
RenThraysk · 18h ago
Still only need one camera drone if a human is spotting targets.
4gotunameagain · 18h ago
In theory. In practice you would not allow a single camera drone to be the single point of failure of a mission with such lengthy and risky planning, and dire consequences.
RenThraysk · 18h ago
Still only need one flying drone to identify all targets. There maybe more camera drones available to pilot, but still only need one flying to spot.
A static target only needs to be seen once.
dmkolobov · 16h ago
I understand that you’re probably just gonna reply with “still only need one camera”
…but if GPS is jammed, and there’s only one camera per fleet, how exactly are the other drones supposed to navigate towards the spotted targets unless they’re all equipped with cameras?
RenThraysk · 16h ago
One camera drone can see if another drone is on target.
dmkolobov · 16h ago
So the old “use a single unreliable 2D instrument to coordinate multiple fast-moving projectiles in three dimensional space” approach.
RenThraysk · 15h ago
You are just continuing to spout nonsense.
Camera drone hovers above target and kamakazi drone intersects the line between camera and target, and drops.
nradov · 14h ago
You are just continuing to spout nonsense. All of the drones have cameras. Using a single designated camera drone is a stupid idea, overly complex and completely unnecessary.
dmkolobov · 9h ago
I think the problem is an assumption that people are too stupid to grasp their brilliant idea.
That being said, having all drones equipped with cameras could enable a more robust version of what they’re talking about:
If uplink with human operators is lost, but short-range comms between drones exist, they could use their video feeds to autonomously coordinate amongst themselves.
dmkolobov · 15h ago
So now the camera is pointed at the target? How is it checking that the other drones are headed in the right direction? And the personnel on the ground? They're just chillin' waiting for those other drones to come intersect with the stationary spotter drone's line of sight?
RenThraysk · 15h ago
You are raging, and your thinking has ceased.
We've had two years of footage of drones being flown over tanks, and bombs dropped directly down into them.
dmkolobov · 15h ago
No one is arguing the merits of drone warfare.
We have two years of footage from Ukraine, where camera-equipped drones are launched from a several miles away at most, and where there are networks of pilots and support specialists to assemble and launch more drones in case of (frequent) failure.
I don’t think it’s wise to wager the success of a 6-month mission deep in enemy territory on a plan with a single point of failure, especially when the alternative is equipping each drone with < $100 cameras.
But sure, you’re clearly the better thinker.
0x457 · 15h ago
lmao what? You want to loiter with a camera drone to guide other drones to target? How would that work if neither drone knows where it is (drones had no GPS lock, it's a fact, not a speculation)?
RenThraysk · 15h ago
They knew where they started from. Know where the target is relative to the start point.
4gotunameagain · 3h ago
You need to read a bit more on autonomous systems and navigation, it will surely tame your hubris. Everything is simple if you don't understand it.
Yeul · 18h ago
They had a limited amount of drones in those containers they needed to make them count. My money is on operators.
varjag · 19h ago
The drones reportedly flew from their containers to the staging area autonomously, where they were taken over by the pilots for the attack approach.
ladyanita22 · 17h ago
So what about the AI part that has been mentioned by several outlets?
boxed · 17h ago
I would assume someone speculated and then the hype brain took over and everyone reported it as AI.
I've had that happen to the company I work at and we literally have zero AI stuff.
dboreham · 17h ago
Hype Brain :) And the language changed just a little bit..
shagmin · 17h ago
From my understanding that was to be used as a last resort.
DyslexicAtheist · 17h ago
there was this German talking head "Nico Lange" who made this claim first without providing evidence for it. He is an ex politician and a regular in the Munich Security Conference and I assume this is who was (mainly) responsible for spreading it.
AI gets so much boost from this nonsense. Because now it's about saving our lives.
"ArduPilot can handle tasks like stabilizing a drone in the air while the pilot focuses on moving to their next objective. Pilots can switch them into loitering mode, for example, if they need to step away or perform another task, and it has failsafe modes that keep a drone aloft if signal is lost."
So it is not fully autonomous.
originalvichy · 17h ago
Not for this task, but could be used autonomously. If they trusted that these planes were still in the same spot, and their GPS coordinates were accurate to 10 cm, then what they could do is just program the drones to fly a preset route at preset heights, stop over the plane's wing and then descend all the way to 0 meters.
mandevil · 16h ago
Even with that level of target knowledge (I suspect the US has the investment in the sensor-targeting links to be able to use satellites to know within cm where planes were within a five minute window, but am not sure about other nations) you'd want to have some that were available for later re-targeting to handle misses. Nuclear weapons war plans solve this by relentlessly re-targeting again and again (declassified 1960's USAF war plans called for over 70 different missiles to hit Moscow alone) but with the smaller damage radius of conventional weapons you either end up with a second strike to make sure you get all the survivors of the first strike- or have a trained human who knows the targeting priority in the loop available to update targeting on the fly.
varjag · 16h ago
GPS can't be relied on in that heavily denied area. And indeed the screenshots show drones in failsafe mode.
You can also see the careful departure of drones from containers in the videos, without extra panning or yaw. Not quite how a human operator would fly them.
When you watch some of the footage I feel like it’s clear that there were at least some human controlled drones there.
randomtoast · 18h ago
How could they be navigated from such a long distance? Satellite communication? Wouldn't the lag be too high?
sorenjan · 17h ago
Via the regular mobile network according to one article I read. The Ukrainians said that all operators where safely out of Russia when the news broke, so I doubt they where at the airfields several hundreds of kilometers from Ukraine.
mrandish · 14h ago
To reduce latency, I wonder if the phones were connecting to a covert site in Russia which had a high-bandwidth, lower-latency wired link, maybe a front company established in Russia for the operation with servers and broadband internet connections. Or maybe just a colocated server at a major backbone site in Russia was rented by a Russian front company. Seems like the kind of thing intelligence services do. While I'm sure Russia has more restrictions on renting colocated servers than the U.S., it's still something that needs to happen every day. Russia also has a fairly robust underground economy of less-than-legitimate companies doing illicit things, so there have to be ways for those companies to avoid restrictions (probably involving bribing certain people).
If the attack was coordinated this way, I assume whoever sold the colo to Ukrainian intelligence thought they were simply setting up yet another server for a shady Russian scam company. Foreign intelligence services often avoid scrutiny by using the same methods as domestic criminals in the target country.
randomtoast · 4m ago
Okay, and GPS is restricted near military installations, but mobile internet remains accessible?
yieldcrv · 17h ago
Everyone was right next to the bases in Russia
twothreeone · 18h ago
Assuming they were remotely operated - at least partially during the final few minutes of the attack - I wonder if the pilots remained in Ukraine or were hidden somewhere close by. I'm assuming they remained in Ukraine, thousands of kilometers away. If so, how did they pull off the remote connection over enemy territory? The only option somewhere as remote as Irkutsk seems to be Starlink, unless the trucks carried custom transceivers (which seems like it would be easily discovered during transit).
alwillis · 11h ago
Regarding jamming:
Now in its third generation, the Ghost Dragon has come a long way since 2022. Its original command-and-control-band radio was quickly replaced with a smart frequency-hopping system that constantly scans the available spectrum, looking for bands that aren’t jammed. It allows operators to switch among six radio-frequency bands to maintain control and also send back video even in the face of hostile jamming.
giantrobot · 17h ago
Over cellular modems in the drones with Russian SIM cards. The operation was prepared for months in advance. Getting a bunch of pre-paid SIM cards from the Russian equivalent to 7-11 (which might just be 7-11) was probably the easiest part of the operation.
hofrogs · 16h ago
Stores won't sell a SIM card without the buyer providing valid state ID, and SIM cards are disabled by carriers if they suspect you are using someone else's card and you can't provide an ID for it. This is one of the recent laws. Getting a phone number/data plan that isn't associated with your real identity (and instead registered to someone else, usually a homeless guy somewhere) isn't impossible, but those wouldn't come from a grocery store.
luma · 16h ago
They've been planning this for over a year and this is the SBU we're talking about. I'm pretty sure they could figure out a way to light up a data plan on a cell phone in Russia when needed.
I wonder if identity documents were taken from POW's captured at the front lines.
pjc50 · 16h ago
Or just their civilian phones. Suddenly occurs to me how effective that would be as a source of SIMs. Won't last forever but can be produced in the month preceding the op.
lordnacho · 15h ago
Maybe just tourist SIMs that let you roam into other countries? Buy a load of SIMs in Kazakhstan or somewhere like that, roam into Russia, now you have internet?
twothreeone · 16h ago
Isn't GSM super easy to jam? Also, there's civilian cellular coverage in Siberia? And right next to military bases?? Wow..
giantrobot · 11h ago
Jamming only works when there's equipment in place and activated. Prior to this attack there was no reason for Russia to even have jamming equipment located at a base deep inside their borders let alone active.
sreekanth850 · 17h ago
India used oldschool L70 Guns, zu 23 and ZSU-23-4 Shilka against pakisthan's drone swarm attacks. They are modernized to track, lock, and fire automatically. But they are cheap.
Immediate post-WW2 vintage. The classic design of AA gun.
As far as I understand it from talking about Turkish drones, you mean https://en.wikipedia.org/wiki/Bayraktar_TB2 style, aircraft size drones, rather than the quadcopter size ones? The latter can more easily hide in terrain.
sreekanth850 · 16h ago
Asisguard Songar drones.
pjc50 · 16h ago
Ah, thanks for the detail.
cenamus · 17h ago
What kinda targeting system do they use? Probably no significant IR signature and radar one's pretty small aswell?
sreekanth850 · 17h ago
We have multi-layered low signature and passive drone detection system, including radar, radio frequency (RF)-based systems, electro-optical sensors, IR, and video tracking. We also have Thermal imaging for all weather use. And all are integrated into a single command and control system. These upgraded Old school system proved to be worthy to shot down more than 500 turkish drones sent by Pak.
cenamus · 1h ago
Ah so more like traditional, larger drones with gas engines/turbines, as opposed to the quadcopters?
petra · 16h ago
What was the kill rate of this system ?
sreekanth850 · 16h ago
None of them caused any casualities or any damage to infrastructure. As we have multi layered defense system with
L70 and Zu23 for very short range Defense agains drones and UAV, Spyder for Short Range, Akash for Medium range, s400 for Long range and high value target, exact kill rate of Guns are unknown. In press briefing, officials said that these were so effective against drones.
petra · 15h ago
That's very impressive, especially at low cost.
sreekanth850 · 5h ago
Yes, Cost is the main factor, one cannot use expensive air defense to shot down drones that are relatively cheap.
ahartmetz · 13h ago
The strength of a reflected radar signal generally decreases with the fourth power of distance: r-squared to the reflector, and r-squared back from the reflector (assuming more-or-less uniform backscattering). Which means that a low radar signature is usually still detectable, the object just needs to get moderately closer.
So being stealthy in the radar spectrum is pretty difficult, and I often wonder if stealth planes are mostly a means to transfer money from the state to defense companies.
potato3732842 · 16h ago
There hasn't been much coverage because it's not as sexy as drones but the degree to which AA tech that was formerly the domain of well funded armies has proliferated down the economic spectrum in recent years is really hard to overstate.
dragonwriter · 1h ago
> There hasn't been much coverage because it's not as sexy as drones but the degree to which AA tech that was formerly the domain of well funded armies has proliferated down the economic spectrum in recent years is really hard to overstate.
A lot of the key AA tech that has suddenly become important in the era of drone swarms began proliferating to mid-tier forces around the 1980s (or earlier), and was retired by well-funded armies between then and the 2000s, because compared to SAM systems, it was suitable only for lower, slower, less capable targets.
Turns out, suddenly large numbers of lower, slower, less capable targets are being fielded, and its really expensive to take them on with SAM systems optimized for dealing with modern manned aircraft, cruise missiles, and/or ballistic missiles.
sreekanth850 · 16h ago
Exactly, one cannot use patriot or s400 to dhot down drones, they are way expensive compared to a cheaper UAV.
No comments yet
originalvichy · 17h ago
If your country is an adversary of China, I would be scared. I've seen videos on YouTube of how the drones from Chinese drone light shows take off and return to their launch areas. They have remarkable accuracy.
dji4321234 · 15h ago
These generally just use RTK and a base station; nothing interesting and extremely easily rejected by EW (since they need both accurate global positioning signal _and_ RTK signal).
Inside-out SLAM strategies and on-device ML are much more interesting and are starting to trickle into COTS drones. For example, the latest DJI drones all use SLAM for return-to-home even when GPS denied: https://www.facebook.com/reel/440875398703491 , and the latest Matrice 4 enterprise drones also have end-user ML model runtimes that can fine-tune flight plans using user-provided logic.
Inside-out last-second targeting is also very popular in Ukraine, with off-the-shelf "find the nearest car/person in analog video, lock to it on signal lost, and send Betaflight MSP stick commands to hit it" modules readily accessible on Aliexpress.
mountainriver · 15h ago
We do that in the US as well, it’s just more regulated (for some reason)
empath75 · 14h ago
But the US can't manufacture them at the scale china can.
jxjnskkzxxhx · 14h ago
Boards don't hit back.
bigyabai · 15h ago
If GPS geolocation makes you scared then don't look up what America did when we invented the JDAM...
avereveard · 15h ago
Eh that's easy enough with inertial or ground navigation and some kind of beacon on the charging bay, the hard part is in enemy territory under the fog of war, jamming, and spotting targets on the fly.
overfeed · 13h ago
If the drones are self-guided there's nothing to jam[1], and what can you do after spotting a swarm of drones? Shoot $5m missiles at each one of them?
maxglute · 13h ago
Not to mention DJI has AESA radars on $8000 agricultural drones. COTS drones withsensors + sensor fusion on par with best smart munitions.
slicktux · 19h ago
Awesome technology!
Nice to see Dead Reckoning being used with computer vision and offline maps!
Something college students have been doing in robotics competitions here in the USA ;)
major505 · 18h ago
TO know more, here is a brief presentation about drones in Ukrain War.
Shaheds and Ukrainian long-range drones are based on inertial navigation, the drone knows its coordinates and the coordinates of the target it is to hit, and the entire route between them is covered based on data from accelerometers, gyroscopes and magnetometers.
However, the decision-making based on image recognition mentioned in the article is undoubtedly more effective in more changing fields, when the target is moving
Oarch · 20h ago
Beyond jamming, I imagine some kind of autonomous laser system could also be pretty effective at downing large numbers of drones within a given radius.
Cthulhu_ · 19h ago
There's a few, but they're large, expensive, require a lot of electricity and have limited range; there's the Silent Hunter [0] which is 30-100 kilowatts max power but which has a range of up to 4 kilometers. Raytheon has a 10 kilowatt palletized version that can go on a truck bed [1]; I can't find any numbers but it's listed as short-range, so I presume it's only effective at distances of less than 1 km, probably only tens or hundreds of meters. Plus they need to detect the drones first, but there's multiple ways to do that. It likely needs a network of detectors though.
What they are good at is target tracking, having started out in satellite communication.
Their tracking system paired with a 30mm Bushmaster cannon and proximity ammo is another solution, and there are apparently 160 of them heading for Ukraine to be mounted on M113 and Kozak vehicles.
Israel published last week that it made a trial run of the iron beam (10km range or so) during conflict with hezbollai. It had 40 intercepts and full operational deployment is scheduled for this year
glitchc · 17h ago
Lasers aren't effective. Most of the drone is just an empty frame. The control board is pretty tiny, as is the ordnance. Targeting those or the propulsion systems is quite difficult. Sure, you can punch holes in the chassis but it takes a lot of guesswork to hit something vital. It's the wrong weapon. Something with an area of effect, like a shotgun or a net is much better suited to stopping drones.
00N8 · 15h ago
Are you sure about that? AFAIK effective laser drone defenses are not yet widely deployed proven technology, but I don't think small beam size is a limiting factor. Getting enough power onto the target to disable it is a big challenge, but part of that is fighting the natural tendency of the beam to spread out & be attenuated by the atmosphere - not that the beam affects too small of a spot on the drone.
Having a laser that spreads out to e.g. 30cm radius at 500m is not hard to do if you need an area of effect weapon & can push enough power (ie. your laser is powerful enough, but not so intense that it ionizes the air & blocks itself). Reflections seem like a bigger problem: If the most effective defense includes guys with shotguns &/or there are a lot of unprotected personnel in the area, how do you make sure stray reflections don't end up blinding them?
glitchc · 14h ago
The point is that a focused laser will put a hole through the drone, much like an armor piercing round, but that is often insufficient to disable the drone. A larger ballistic projectile (think a solid shell or a rock) is much more effective. Alternative energy weapons based on microwaves and SPL also work well.
Oarch · 12h ago
Presumably hitting any rotor, even a small amount, could be enough to bring it down? They're finely balanced things
glitchc · 11h ago
Already tested. Success rate is too low. A great deal of aenergy gets wasted.
Remember, this is about asymmetric warfare. If the number of rounds or amount of energy required costs more than the drone it shoots down, then it's not an effective deterrent. Militaries are looking for single-shot weapons to take down drones. Fire once and move on. It's the only way to deal with a swarm. Think about it for a bit and it will become very obvious.
tguvot · 16h ago
Israel published last week that it (trialed) deployed laser system to shoot down drones year ago at North. It had 40 intercepts or so. Full operational deployment scheduled for the end of this year
glitchc · 13h ago
You mean the demo done recently? The article might be misleading. The IDF tested 20 different systems produced by their military OEMs. It wasn't just lasers, and lasers are far from being the clear winner. Here's the official post, it contains a demonstration video:
Did you read your own article? It was developed last year and one of the weapons tested in February of this year as per the IDF post. The rest is Rafael marketing speak.
tguvot · 11h ago
Did you read the article ? It starts with
>In a historic breakthrough, the IDF on Wednesday announced that an unnamed laser defense system similar to the much celebrated Iron Beam laser system has shot down dozens of aerial threats during the war.
>Already in fall 2024, The Jerusalem Post had learned that the IDF had used laser defense systems in operational situations but was barred from reporting on that at the time.
It was all over Israeli news together with videos of operational intercepts.
it has nothing to do with "february cookoff" of different systems and orders for iron beam were placed in january with deployment by the end of the year
glitchc · 10h ago
> orders for iron beam were placed in january with deployment by the end of the year
Do you have a credible source from the IDF attesting to this?
And do you believe the IDF would conduct a demonsttation after it has already committed to a system?
tguvot · 10h ago
>Do you have a credible source from the IDF attesting to this?
>And do you believe the IDF would conduct a demonsttation after it has already committed to a system?
what demonstration are you talking about ? if you are talking about whatever was published in february, it's mostly for more tactical/mobile use i believe.
there is an issue that north of israel is very hilly, so it's possible to fly drones from lebanon below radar visibility range and then just to get them pop-up 50km away from border. it was major problem last year and the publicized trials a believe concentrated on sourcing systems to solve this issue.
pjc50 · 16h ago
The main defence drones have against both lasers and Bofors-type guns is staying low, such that they are below the horizon or behind ground clutter until as late as possible.
jvanderbot · 19h ago
Now imagine that but with a shotgun shooting bird/buck shot at decent ROF. Way better.
The problem with any point defense system is radiating any energy makes you a big target. So you would want a passive (EO/IR?) or triggered active/passive system.
lenerdenator · 18h ago
12 gauge would be fairly strong medicine, but it has to be close in. If the drone is a stabilized gun platform with the ability to aim decently well (<4 MoA @ 100 yards) that's not going to be a winning battle.
jvanderbot · 17h ago
Taking the discussion back to reality, almost all uses of drones at the moment are via suicide ("wire guided COTS missiles", you might call them), or just plain old recon. There are probably still plenty of grenades dropped from hover as well.
For those uses, there's a fairly decent approach ["missile"] or hover ["bomber"] stage that is probably plenty vulnerable to autonomous PDS via 12 guage medicine.
Tracking / detection could even be passive, partly acoustic, partly EO/IR, with only a small fire control radar if you really want it.
wiseowise · 18h ago
That’s assuming there’s a single drone and not an intelligent swarm that will circle around you.
jvanderbot · 17h ago
Yes, I'm assuming reality.
jajko · 17h ago
Why nobody us building automated shotgun based solutions for anti-air defenses... maybe they need more than 100-150m reliable reach? Having 300 of those covering one airfield may not be ideal, nor cost effective, nor easy to manufacture and deploy and maintain in such numbers... we talk about russia after all, they let most of their strategic bombers arrogantly unprotected on runways with full gas tanks.
I don't think western common folks grok how depraved that country is in terms of doing good work, reward systems for such and corruption on every single level. puttin' built a mafia state and pushed this behavior from top->bottom, and these are side effects. Not some soviet competence and discipline, which wasn't stellar either but light years ahead of current state.
sreekanth850 · 16h ago
India had successfully used its age old L70, Zu23 guns to track and shot down 500+ turkish drones sent by pak. All these legacy weapons are modernised with passive drone tracking, locking, targetting and automated firing system.
https://theprint.in/defence/how-upgraded-l-70-guns-or-origin...
jvanderbot · 14h ago
I really don't care what Russia does, but in a future war USA is more likely to be attacked by asymmetric exploding drones than to be attacking with them. That I _do_ care about. And for that, a PDS system on a few trucks seems kind of useful. If they can be made cheaply and the expensive bits centralized, then a few $1k every 100 yards is kind of reasonable, don't you think?
robotnikman · 14h ago
That's basically what ammo such as AHEAD is. It bursts before reaching the target and sends out shrapnel in a shotgun-like pattern.
transcriptase · 16h ago
They’re parked in the open because of the treaty requiring those type of assets be observable by spy satellites. The USA in turn does the same.
Paradigma11 · 15h ago
Nope. You are talking about the New Start treaty.
1) Russia suspended the treaty 2023, so it is not relevant here.
2) The use of hangars is not prohibited, as can be seen by the climatized hangars that the US keeps their B1 in.
3) Russia has announced plans for such reinforced hangars years ago, but very likely some dacha or yacht had higher priority.
XorNot · 17h ago
We've had acoustic gunshot detection for years at this point though. It's not like a shotgun firing is a quiet event.
jvanderbot · 17h ago
There's a huge difference between firing back at an enemy that is attacking, and spraying radio signals all over the horizon even though nobody is attacking yet. The former won't tell them anything new, but the latter (which I'm talking about), is (was?) considered somewhat dangerous.
When I said "triggered" I meant you would enable it when under attack, at which point it doesn't matter if they know you're there anymore.
gpderetta · 16h ago
Are shotgun pellets supersonic?
ianburrell · 13h ago
One thing I haven't seen explored is using autonomous drones as defense. Like hand sized drones optimized for speed and maneuverability intercepting larger drones. They should be super cheap. They would also be small enough for troops and vehicles to carry one.
maxglute · 13h ago
Anduril Anvil does drone-drone intercepts from a few years ago.
I think you'll probably see mini flak guns, lasers, microwaves, and defensive kamikaze drones as the main defensive tools
bell-cot · 18h ago
Best-case scenario: The expensive laser system just became the most obvious and highest-value target in the area.
theptip · 17h ago
Depends, in this case the strategic bombers are worth more.
bell-cot · 14h ago
Micro-scale, very likely true.
Bigger picture - if knocking the laser defense off-line slashes the unit cost of destroying bombers, then it may be the obvious first move in any competent attack.
jajko · 17h ago
Take 5 drones, sneak from other directions and simply overwhelm the system. If in pair, multiply the attackers via some decoys, it becomes just a statistics game.
I can imagine this protecting some future US bases in same way C-RAM is used. But from what I read from ie Iraq veterans they had it turned off most of the time for the fear of shooting down its own planes. So much for trust in high tech if its too powerful and automated.
Chinese have some systems, but from demo I've seen the laser beam took some serious time to shoot a single missile. Drones are smaller and way more fragile (so also harder to hit) but this ain't Star trek or Star wars.
FirmwareBurner · 19h ago
I'm thinking the same thing. You don't even need to fully destroy the drone, if you manage to damage the camera sensors or the exposed lithium cell it's game over for the drone.
jasonjayr · 19h ago
How close are we to making a system that can have multiple counter measures autonomously deployed?
(1) if prop based, launch something to snare the props
(2) if reflective, pre-launch something to spray black non-reflective paint at it, and followup with laser
(3) if evasive, approach with random manouvers
(4) if unknown, launch everything and see what works, and feed it back to the training data ...
etc, etc.
jandrewrogers · 17h ago
Mirrors are not an effective defense against military lasers. The power levels are too high and dielectric mirrors only work over narrow wavelengths. In the specific case of US military tech, some of the platforms use white lasers such that even dielectric mirrors are pointless.
FirmwareBurner · 16h ago
>Mirrors are not an effective defense against military lasers
So the cartoons lied to me?
gosub100 · 15h ago
I would try a shotgun style subsonic load with an adjustable fin/spoiler system that can be calibrated relative to the range of the target. The projectiles fin would flare out to reduce speed as it approached the target. Then an explosive charge would release a spray of super glue or pancake syrup, or something to gunk up the mechanism or disrupt the airflow on the propeller.
numpad0 · 18h ago
It still takes few seconds per target with technologies available right now. That's likely the reason why an operational anti drone laser turret is not a thing yet.
Retric · 17h ago
> operational anti drone turret is not a thing yet.
I’ve read about a bunch of these systems even if they aren’t in widespread deployment some are still being tested in real world conditions.
“The Silent Hunter has been used by Saudi Arabia to guard against Houthi drones and missiles.”
“During the World Defense Show in Riyadh the February 05, 2024, Poly Technologies announced the first hard-kill engagement of a one-way attack drone.[6]”
numpad0 · 13h ago
Again, it takes few seconds per target with current laser tech. Which means they don't have important simultaneous multi target engagement capability.
The targets, whether it's plumbing pipe rockets or lipo drones, come in at 100-1000 yards/sec, so you don't really have that many seconds per target.
They work in the demo in which you just shoot down the sole target as it fly perpendicular to the machine for both physical and career safety, but when it comes to deploying the thing around your bed, guns make a lot more sense.
Retric · 13h ago
The quotes suggest that system has actually been deployed and used successfully in real world conditions.
Also, your objection doesn’t really fit how drones have been used. Massive highly coordinated drone swarms are extremely unusual, the threat is mostly individual drones or small clusters.
Many very dangerous drones are well under 100y/s aka 200mph.
1000 yards/second aka Mach 2.7 is well beyond the ‘drones’ people are concerned with and into expensive missile territory. Which is where anti missile systems get used.
Maybe not completely, at least not for the camera sensors. If it has good enough initial targeting data - or a good enough last image from the camera - it may be able to find the target by inertial navigation from there (depending on laser range, of course).
FirmwareBurner · 19h ago
> it may be able to find the target by inertial navigation from there
Or miss by a long shot and hit a civilian instead.
fredthestair · 19h ago
It's extremely easy to make surfaces partly reflective and extremely hard to make lasers outside of labs more powerful.
mattashii · 19h ago
More powerful than what? If StyroPyro can build handheld 200W+ LED laser device in his workshop [0], why wouldn't a sufficiently funded military be able to build an anti-drone laser with the same (or higher) power output?
Note that his laser burns through various reflective materials, including mirrors, copper, aluminium, and steel.
Burning stuff at 20 feet is much easier than burning stuff at thousands of feet in distance.
No comments yet
fredthestair · 9h ago
If a drone is 90% reflective it doesn't need a weapon besides your laser.. (If the star wars approach weren't a scam to enter an ever more expensive race as the side with more money than sense.)
Hoping and praying someone in the administration is able to convince Trump to align closer to Ukraine (even if only slightly) because of this.
The Ukrainians pulled off an absolute coup on Sunday. A third of a nuclear-armed country's strategic bomber fleet inoperable for the foreseeable future. Someone at NORAD probably said "they should have sent a poet" while looking at the satellite imagery.
If middle powers like Ukraine can do that to Russia, they can do that to countries like the US. We need to be on their good side.
pjc50 · 16h ago
The US seems to have given up on the concept of being on the "good side" of anyone and retreated to the safety of bullying and threatening to invade Canada (as a "joke"?)
lenerdenator · 15h ago
Most people have given up the concept of being on the "good side".
That's how Trumpism can gain any traction at all. The amount of international engagement Russia had as Putin made himself tsar was embarrassing, and to a person with no scruples if the money is right - like Trump - it just illustrates that the guardrails aren't really there.
timeon · 14h ago
> Most people have given up the concept of being on the "good side".
In US.
lenerdenator · 11h ago
Oh, it's elsewhere, too. There are plenty of illiberal candidates winning elections across the West.
There's also an argument to be made that Europe more or less sold themselves out for cheap natural gas. Nordstream 2 was constructed after the Russians (at that point, under Putin's puppet Medvedev) had invaded Georgia.
eastbound · 14h ago
> Most people have given up the concept of being on the "good side".
“The revolution eats its partisans” is the most accurate description of it. People on “the good side” turn against their peers for not being on the good side enough. To wit, people who turn away don’t generally first notice that the good side isn’t so good; they first notice being bullied by that side, then they reflect on what it means to support the good side’s points of view (spoiler: A crime against humanity).
tgv · 17h ago
> If middle powers like Ukraine can do that to Russia, they can do that to countries like the US
Ukraine borders on Russia, but the US is separated by ocean from serious threats. Attack by UAVs of this sort seems nearly impossible.
wood_spirit · 16h ago
How does the ocean protect America from swarms of short range drones being launched from normal looking shipping containers on trucks being controlled from thousands of miles away?
xeromal · 16h ago
The first volley would work of course but then container ships are gonna be nuked from orbit after that
Paradigma11 · 15h ago
Which would utterly destroy the US economy and do more damage than any drone attack.
Plus imagine if those attackers realize they can ship those containers from Mexico or Canada.
xeromal · 14h ago
Pretty sure it would destroy whoever is on the other side of those shipping containers too though
D-Coder · 14h ago
> whoever is on the other side of those shipping containers
Could take a while to figure that out. It took days/weeks to figure out the 9/11 attacks with some certainty.
xeromal · 14h ago
I'm not sure what that does for whoever initiates the attack. They still eventually get found out and unless they've managed to conquer the US, everyone is now bankrupt.
unrealhoang · 9h ago
> everyone is now bankrupt
A good reason enough for a lot of people, no?
acdha · 16h ago
Don’t we receive millions of containers of cargo annually, not to mention having fewer internal movement restrictions? It certainly doesn’t seem implausible that someone could ship some drones around - the hardest part is avoiding explosives detector, but that’s a hard problem and the defense has the unenviable task of having to get it right millions of times.
wood_spirit · 16h ago
Countries could be forward deploying these assets covertly in deniable ways in prep for future tensions?
acdha · 9h ago
I’d certainly plan for that if my job was physical security. It’s cheap by military standards and it’s not like we don’t have precedent with things like Russian deep-cover operatives. The commodification of the tech could make an adversary more confident in their ability to deny it if found, too, if everyone is using parts from China and open source software.
I think the key deterrent is that the U.S. has production capacity for the important systems and overwhelming capacity to strike back, so a rational foreign state isn’t going to think there’s a way they win by trying it. Terrorist groups might be a different story, so I’m really glad this wasn’t an option during Bush’s big adventure in Iraq and Afghanistan because however bad drone strikes are for military defense, they’re even worse for civilians.
detritus · 14h ago
I've long-wondered about the use of global container logistics for moving something like a small nuclear or chemical/biological weapon and just having it wait in-situ, until it needs to be activated.
I never considered drones, which is even more obvious, in hindsight.
acdha · 9h ago
Yeah, it’s frankly a big terrifying. Nukes are at least plausibly hard to shield from detection but drones don’t have a unique signature and there’s so much dual-use stuff that it seems too plausible that someone could get stuff through customs by claiming it’s civilian and then arming it right before use.
oneshtein · 14h ago
While not a land border, the Bering Strait does include two small islands, Big Diomede (Russian) and Little Diomede (US), which are only 3.8 kilometers apart.
marcusverus · 13h ago
The US contains, at a minimum, 10s of thousands of nationals from every middle power in the US. Sometime adversaries, like Iran and China, have more than a million nationals in the US. Every one of them has perfectly legal access to the technology which was used in this attack.
Mossy9 · 17h ago
Apparently the final tally was 12 planes, but still, an impressive display
technofiend · 17h ago
Depends on whether you mean damaged or destroyed. Ukraine reports 41 damaged, 12 destroyed.
lenerdenator · 16h ago
12 destroyed. Hard to tell if the others were only damaged, and of course, how badly.
The first casualty of war is the truth.
That being said, any sort of materiel loss on weaponry as important as strategic nuclear bombers is a massive problem for Russia. The logistics of repairing them, if possible, is going to be complex.
jajko · 15h ago
Do you know how many they still had operational? Not that many, somewhere between 50 (more realistic) and 100 (rather idealistic if the rest was fully gutted for equipment).
This was a massive damage, not directly interfering with war against Ukraine that much, but overall power projection. Plus a pretty good insult to russian's FSB and GRU services who had no clue, just like today's Crimea bridge blow.
timeon · 13h ago
> russian's FSB and GRU services who had no clue
I do not want to give any credit to russian secret service but as we have seen in 9/11 and 7th October - secret services in any country are sometimes clueless.
aaronbaugher · 13h ago
"Display" being the operative word. Good for PR, and maybe good for NATO in some small way in the future, but it won't change the war or how many Ukrainians are getting killed. A lot of the Ukrainian "wins" in this war have been of that nature, which probably isn't surprising considering they're being planned by outsiders with other motives.
empath75 · 14h ago
I think Ukraine could only have gotten away with this in the current situation where the US is detaching itself from supporting them, because if Russia genuinely thought that the US was behind taking out a large percentage of it's strategic nuclear fleet, the consequences of that are nightmarish to think about.
Those airplanes are one of the things that give Russia a second strike capability, and if they lose that capability, then they are going to be on a hair trigger in a nuclear crisis.
koonsolo · 17h ago
Other countries have hangars or bunkers to store their planes. That Russia doesn't, is really beyond comprehension.
dboreham · 17h ago
Supposedly they had to be outside to comply with START provisions. But US keeps B-52s (and pretty much all fighters) outside. They might be under tin roofs in desert locations to keep cool. Not in bunkers.
aaronbaugher · 13h ago
Yeah, you can find plenty of examples of them sitting right out in the open. They're protected by two oceans, not by bunkers. A similar attack with trucks full of drones would make short work of them just as well.
koonsolo · 11h ago
Can you explain why they have tires on the wings?
TiredOfLife · 3h ago
START doesn't require that. Besides Russia quit START 2 years ago. Also wherever you got that information is 100% a russian disinfo distributor.
Sporktacular · 17h ago
“In the perfect world, the drone should take off, fly, find the target, strike it, and report back on the task,” Burukin says. “That’s where the development is heading.”
That's the problem in a nutshell. A few years back, few would argue against keeping a human in the kill/no-kill decision chain. It just took one war to get pop tech authors writing on it without even a mention of the ethical considerations or autonomous killing machines.
ip26 · 14h ago
I don’t really want to argue this side, but is it that different from a smart bomb or guided missile? A human is in the loop; the human issued the coordinates of the target to the delivery vehicle.
That kind of operation seems extremely different from a stationary turret or patrol robot with standing orders to shoot upon arbitrary targets at any time it decides to.
Nasrudith · 17h ago
It highlights an awkward apparent fact that 'ethics' and 'honor' are luxuries to maintain when one is on top and they will be thrown out the door as soon as they face an actual threat. Not saying that it is right but that it appears to be the predictable response.
I suppose it highlights axiomatically the terribleness of ethics when they must be defined in a might-makes-right manner. All very high minded and complex questions which leave the awkward question unanswered: what are we supposed to do?
BobaFloutist · 8h ago
I mean I think it's relevant that these machines weren't actually used to kill.
MaxPock · 18h ago
Some bases were able to repel the attacks so I don't get what this "defeat electronic warfare" means
p0w3n3d · 14h ago
I'm not sure what weapons will the WW3 be fought with, but the WW4 will be fought with sticks and stones.
dzhiurgis · 10h ago
Is anyone familiar how does EW work? I keep hearing they send common shutdown codes, but what does it mean? Obviously thats an easy target but commercial drones (probably built in), but how does it work for custom ones? Can't the code be random?
Every generation must learn that "the only winning move is not to play."
bamboozled · 19h ago
Oblivion for potentially nuclear armed bombers?
pjmlp · 19h ago
To the drones used in the movie.
notsydonia · 18h ago
The person who devised this was most likely one Illia Vitiuk, head of the SBU cyber-sec department. Before that position, he was an MMA fighter. In a 2023 interview he says he was inspired by "James Bond films and a life of adventure..." Also apparently stood down then reinstated recently over some unexplained transactions, family finance situation.
I really don’t get all the fretting about drones making kill-no-kill decisions.
You’re telling me that you’d rather this was done by a stressed, emotional, tired front-line private?
Reminds me of the debate about a self driving car that might need to mount the curb to avoid hitting a car - and therefore endangering a pedestrian.
It’s not an easy decision but I’d rather a machine made it than a stressed person!
sorcerer-mar · 19h ago
The thing you're missing is that the decision isn't actually made by a machine. It's made by a person much much higher up across a much broader effective surface area.
That is, historically to wage war you had to ultimately convince/coerce millions of people to wage that war for you -- continuously for years -- along with millions more to support the effort politically and economically. In the highly automated world, you literally only need a lot of money (which also tends to concentrate outside of any semblance of democratic control).
worldsayshi · 18h ago
Yeah a single rich person might soon have the ability to, without coercing a soldier, decide who gets to die. He might just need to coerce someone to deliver the drones and put them on an (automated) truck.
Nasrudith · 16h ago
The smart-ass answer is 'a single person already does, and he doesn't need to be rich, just able to build a time-bomb and send a package'. The Unabomber approach has its limits of course, including being hard to reach those who do not open their own mail.
orbital-decay · 19h ago
The amount of autonomy and decision complexity matters a lot. Autonomously making its way into a known area, recognizing the target and self-correcting to it is basically how every modern cruise missile works.
netsharc · 19h ago
And now we know making an ersatz cruise missile in the garage is possible...
pjc50 · 16h ago
So the interesting question is to what extent the Iranian derived Hezbollah "garage rockets" (very poor guidance, CEP is basically "somewhere in Israel", but extremely cheap and require only smugglable parts) can be combined by some power with Ukrainian style precision guidance.
AnimalMuppet · 18h ago
True, though (so far) range, velocity, and payload are far smaller than what we expect for a "cruise missile".
falcor84 · 17h ago
Well, this week proves that range and velocity limitation can be overcome if you ship them near to their destination in small boxes ahead of time. This wasn't really an option with cruise missiles.
This issue is too complex for me to really have a good thought at the moment, but I will say that this reminds me of that classic story of the Soviet who didn't launch the nukes during a threat came in from the US.
The threat ended up being a false alarm, and that human judgement saved a lot of lives. A machine, assuming it would have launched when seeing that signal, would've ended differently.
perihelions · 19h ago
In 1962, the US sent out a valid nuclear launch order with correct launch codes, by mistake.
Imagine an AI system is programmed to poison the well incase of losing scenario. So it takes advantage of Kessler System to make sure no one else has advantage if it doesn't too poisoning the whole of earth orbits for centuries to come in which satellites could reside.
scott_w · 18h ago
The issue is that software making these decisions can't possibly have the full context when making the decision because it was coded long before the decision point occurred.
What does that mean in practice? Have you ever saw a flash in the corner of your eye, looked for a second, then realised it was just a reflection? For that split second, your brain identified a potential threat and was trying to quickly decide whether to hide, fight or run. But underneath that was a nagging of "it's probably nothing," that caused you to delay pulling out your gun and opening fire on a passing car.
A computer programme may or may not have this coded in. A machine designed for an active war-zone, or a car assuming it'll only ever auto-pilot on a motorway, is used in a suburban area. That same programme that's fine in its original intended location suddenly opens fire on a grandma with a shiny stroller, or swerves into a pedestrian because "pedestrians don't walk on motorways."
That's what we worry about.
pjc50 · 16h ago
> Have you ever saw a flash in the corner of your eye, looked for a second, then realised it was just a reflection? For that split second, your brain identified a potential threat and was trying to quickly decide whether to hide, fight or run. But underneath that was a nagging of "it's probably nothing," that caused you to delay pulling out your gun and opening fire on a passing car.
.. or vice-versa. The average war has plenty of stressed people firing semi blindly at half-identified shapes. And if we want to be literal, there's the case of Lee Clegg and the exact circumstances in which it is legal or not to open fire on a passing car and kill a teenage girl.
scott_w · 12h ago
I’m not denying that awful shit like this happens with humans. I’m pointing out the moral hazard in turning it into a preprogrammed system. We’ve seen software bugs empty bank accounts simply because running that faulty system for 5 minutes iterates 100,000 times. Now give the power to decide who lives and dies and it’s only limited by the ammo it holds.
user568439 · 19h ago
The stressed private might still have a bit of empathy and humanity. Meanwhile millions of drones can be programmed (or hacked) to kill millions of people without excluding civilians or anyone
alkonaut · 18h ago
We have had weapons which are autonomous for decades. You launch them consciously then you know that it will find and destroy weapons based on some "intelligence" (A homing missile with a radar you know is likely to hit the thing that reflects the radar waves, whatever that is. There are artillery shells which home in on vehicles and so on). The launch decision by the human means "I'm responsible for this thing hitting and the thing that it finds". The kill/no-kill decision is made at launch time. An AA missile might hit a civilan jet, but there is no way the operator will make a new kill/no-kill decision once it reaches the jet. You made the decision at launch.
That's the same with these drones. The smarter they get, the further away the human goes. Today it might be simple to create autonomous weapons who are instructed to kill vehicles matching various known appearances. That too already exists. The strike on the Russian bombers was reportedly carried out manually, but it would have been pretty easy to have that autonomous, since the targets are huge, stationary, easily recognizable and easy to navigate to in the geography.
If you launch a quadcopter and instruct it to kill any adult human it finds, then that's the same thing. You wouldn't launch it into an area where there is a remote possibility of being any civilians. No difference from firing an artillery shell. If there is a civilian, or a soldier waving a white flag or whatever - there is no cancel button for your artillery shell. The decision to kill whatever is in the other end was made when you fired it. There is literally no difference between firing a million drones and firing a million artillery shells down range. It's your human responsibility and your human consciousness when you make the decision.
I don't think we have had widespread use of autonomous human-targeting drones yet, but it's by no means science fiction today. Just a matter of time. We'll see their use in this conflict.
pjc50 · 16h ago
Don't forget there's a war on right now in which precision munitions are being used to specifically target hospitals full of civilians on the pretext that the enemy is allegedly underneath.
2OEH8eoCRo0 · 18h ago
I think this is such a hot topic around here because it makes sheltered nerds begin to comprehend the gritty reality of warfare.
yaris · 19h ago
My understanding is that a drone will make decisions that are reproducible (same data - same decision), so if anything goes wrong then it should be possible to investigate (to some extent) and fix.
A stressed private is in this sense ”undebuggable” because much more not-easily-reproducible factors influence decisions. Also I’m afraid that stressed and tired privates at war tend to err towards ”just kill them all” because it looks much more like a videogame.
lukan · 19h ago
The data is never the same. Every situation in war (or reality in general) is unique.
watwut · 18h ago
Human soldiers kill civilians pretty much all the time. Then they brag to their friends how cool they are. Drones do not rape, soldiers rape (and yes they rape men too in case someone wants to make it about gender).
All the bombs Russia thrown onto Ukrainian civilians were thrown by human soldiers.
scott_w · 18h ago
> Drones do not rape
Yet. Drones also don't get tired.
wiseowise · 18h ago
They are powered by batteries or gasoline.
maxwell · 16h ago
Neither do the defensive ones.
__MatrixMan__ · 17h ago
When a stressed person mistakenly does something they shouldn't have, they typically stop doing it after only a small handful of mistakes. We get tired, or bored, or we wise up to our mistake.
When a misconfigured computer does something wrong, it frequently does it over and over and over again until it is prevented from doing so by an external intervention.
No tired private is going to mistakenly rampage through a populated area mowing down civilians. But a confused drone swarm might.
amelius · 17h ago
"To err is human, but to really foul things up you need a computer."
Nasrudith · 16h ago
Very good principle but horrifyingly a drugged soldier might. Look up Larium, an anti-malaria drug potentially implicated in wartime rampages. Morally and ethically it raises some weird questions. Sure you can balance the probabilities of death from taking a drug vs not. All drugs have a lethal dose. But most drugs don't kill people who haven't taken them.
_verandaguy · 18h ago
I think this oversimplifies the matter.
At a micro level, sure, you could argue that kill/no-kill decisions are being made by frontline enlisted soldiers, but the stress, the exhaustion, and being able to see your enemy downrange add a degree of humanization and discretion.
Wars are brutal. Some warring factions are incredibly brutal. But taking a human life can't be reduced to a `KILL? [Y/N]` decision made by some MAJ somewhere. Killing enemy soldiers is a last resort (formally, this is the concept of _military necessity_), and normalization of it as anything else is a mistake.
I also think it's a fallacy to compare this to the self-driving car trolley problem; the timeline of decisions, length of the decision chain, and ultimate goal of the decision-making process are _aggressively_ different.
My overarching points are, I guess:
War is a tragedy, obviously, but I don't think there's a way to avoid it right now. In absence of some way to stop war forever (hah), we can't trivialize taking human lives. It's not lost on me that this has been happening for a while at this point, and that it's getting worse. I still oppose it, and I think we all have a responsibility to be more critical of this regime of warfare.
I don't know how to counter the argument of "well, if we don't do it, _someone else_ will, eventually." This is some really fucked up, self-justifying inductive reasoning that can't easily be countered by calling out the moral bankruptcy of the premise. In the past, mutual disarmament treaties have been a down-the-line bandaid for this kind of thought process, but the nuclear rearmament we're seeing in the world right now shows it's not a panacea.
XorNot · 17h ago
You could start by not inventing vague hypotheticals to argue against, and instead engaging with observable, measurable strategic and tactical reality?
War is studied. There are journals, papers and research on war fighting at all possible levels.
In the most recent action by Ukraine you can observe actual reality: what did they attack? Military equipment of the enemy. Why did they attack it? To degrade the enemy's ability to sustain and rotate their forces attacking them. What was it for? Well for one thing it will hopefully considerably reduce their ability to bomb civilian targets.
_verandaguy · 17h ago
In this specific case, I agree that it was fine -- using drones with limited decision-making ability to strike targets like parked aircraft is okay, as long as there's an overwhelming likelihood of the drones not getting false positives from invalid targets.
My response was more aimed at the parent comment to my previous one, which seemed to paint delegating kill/no-kill decisions with a brush of "I don't know why this is such a big deal."
maxwell · 16h ago
If war is tragedy, where is the dramatic irony?
bad_haircut72 · 19h ago
I was once talking to an F35 pilot about what its like and during the conversation he said something, "If its red, its dead" referring to how the computer classifies things as friend/foe for you. I realised it doesnt mean much to have a human in the loop when theyre making their decisions based completely on info the computer serves up anyway.
dralley · 18h ago
Indeed - in a hostile situation you're never going to get an eyeball on most hostile air targets.
JohnBooty · 19h ago
Can we think slightly more broadly?
Targeting a stationary Tu-95 bomber with no protective measures in place is probably the easiest possible identification task for a drone.
A lot of kill/no-kill decisions are more subtle, or involve unknowns, possible nearby civilians, etc.
Or look at this crazy FPV piloting job. You think AI could do this? Pilot maneuvers through an absolute maze of anti-drone nets to hit a moving truck. (essentially SFW; video terminates slightly before impact)
Drone operators are generally not on the front lines, and sometimes they are literally on the other side of the world sitting in a cubicle. They are usually specialists of some sort, not untrained privates.
WillAdams · 14h ago
Drone operators seem to become quite jaded by their safety/removal from the theater of operations.
and reporting of the use of this term by U.S. drone operators.
CamperBob2 · 18h ago
Why couldn't AI do that? It's better than any human at this stuff.
JohnBooty · 13h ago
You could certainly train AI to navigate that particular maze of netting, but I'm far from convinced you could train an AI to navigate a near-infinite variety of novel, hostile measures not present in the training corpus.
It seems trivial to confuse a Tesla's AI. I'm assuming they're fairly near the top of the game when it comes to that, yes?
This sort of intentionally hostile pathological case is of course rare in real-world driving. It will not be rare in warfare.
And a drone has to operate fully in three dimensions, unlike a Tesla which is effectively operating in two dimensions.
An autonomous drone will also have extremely constrained computing resources relative to a Tesla due to size/weight/power constraints.
GuinansEyebrows · 17h ago
Sometimes I feel like the entire AI field just never watched The Terminator.
ThrowawayR2 · 15h ago
Campy, low budget action movies are not a sound basis for forming public policy.
CamperBob2 · 17h ago
Whether it should is a different question. It certainly can.
No comments yet
cheeseomlit · 19h ago
It doesn't take much imagination to see what's so disturbing about it, they're swarms of autonomous killing machines. It can't be bargained with. It can't be reasoned with. It doesn't feel pity or remorse, or fear. And it absolutely will not stop, ever, until you are dead
ARandomerDude · 17h ago
> tired front-line private
It's hard to imagine there is a military in the world that would put something so important in the hands of a private. Yes the drones in this war are relatively low cost but they are also so vital to both sides' war effort that there is inevitably more command and control over their use.
cpascal · 18h ago
Machines have been making kill-or-no-kill decisions for decades, and they were a lot more indeterminant. Heat-seeking missiles largely kill whatever is hot. Proximity fuses in WW2 detonated whenever they got near something. Anti-personal mines kill whenever pressure is applied. A CIWS will target things that get too close to it that don't identify IFF. Naval mines kill if something magnetic is near them.
At the end of the day, it's still humans deploying these weapon systems and accepting the risk that they might cause unintended casualties.
lawlessone · 17h ago
>It’s not an easy decision but I’d rather a machine made it than a stressed person!
A machine also can't be held responsible.
Who do the surviving relatives of the crush pedestrian sue? No longer the driver i guess
SoftTalker · 15h ago
Normally the owner of the vehicle, at least if the driver was using it with their permission. Think about this before you let someone borrow your car.
lawlessone · 13h ago
How can you old the owner responsible for the AI's errors?
SoftTalker · 11h ago
It's their car, they are operating it on the public roads.
sfn42 · 16h ago
The thing that gets me about this autonomous car thing is that it shouldn't really happen. When a car has to swerve to avoid an accident it is typically because the driver was speeding and/or not paying attention.
Autonomous cars wouldn't speed and they are always paying attention. Maybe you could argue that it happened because the other driver was drunk or for some other reason swerved into your lane but this is still an incredibly niche situation and if it does happen just sue that guy. The vast majority of car-to-car accidents happen when both drivers are irresponsible. Autonomous cars will significantly reduce accident rates just by following the rules of traffic and being constantly aware.
The whole who should be sued thing is asinine, just have a national insurance fund for it or whatever. Cheaper than all the accidents it prevents.
pests · 16h ago
I’ve read that self driving cars almost always try to slow down /brake opposed to swerving for the simple reason of removing energy from any potential collision.
There was an article here serveral weeks ago about how in some self driving cars the brakes can be pre-loaded and can be used much quicker than a human could, reaction times aside.
Sporktacular · 17h ago
Yes. The private knows he can be punished for getting it wrong.
bell-cot · 19h ago
To be fair, that's a relatively buried and brief aspect in the article. Which is otherwise heavy on real-world military realities.
Personally, I see anyone fretting about imperfect drones making those decisions as either (1) one of today's 10K, or (2) a performer covering a trending topic. Land mines were deployed at giga-scale over a century ago (WWI). Spring-guns were enough of a problem in England to be outlawed two centuries ago.
lukan · 19h ago
A) land mines are banned in most countries for that reason
B) mines don't move by themself, but stay where you deployed them
A moving autonomous killer drone has a potentially bigger effect if unleashed on the wrong area. Besides, war is usually fought in civilian areas. If you send the drone to the frontline, but its GPS is jammed, it might move somewhere close by.
bell-cot · 18h ago
A) Most countries also have laws/treaties banning invasions. But when those happen to them anyway, most countries start ignoring the bans on landmines. Which are quick/cheap/easy to produce, even with century-old tech.
B) https://en.wikipedia.org/wiki/Naval_mine Not that anybody's doing Environmental Assessments on land, to determine whether the mines might be washed downstream after a heavy rain, or moved downslope by a landslide, or ...
The problem you describe sounds very much like poorly-directed mortar/artillery fire & strategic bombing - which have been killing civilians at scale since at least WWI.
hoseja · 19h ago
This is flawed thinking. Human decisions scale worse - this is a feature.
basisword · 19h ago
>> I really don’t get all the fretting about drones making kill-no-kill decisions.
>> You’re telling me that you’d rather this was done by a stressed, emotional, tired front-line private?
I want somebody that can be held accountable when it goes wrong. With automated systems we don't have that.
eviks · 18h ago
That's not an issue, you can always hold someone accountable, from those who own to those who program.
basisword · 16h ago
We can't even figure this out for self-driving nevermind warfare.
kelipso · 17h ago
Is the person who presses a button that releases an autonomous kill-no-kill drone into a country going to jail when that drone razes an entire town of civilians? I doubt it. “Mistakes happen” etc.
The visceral reaction is not there that would motivate the legal system to send the group that started a cascade of autonomous decisions by a robot to kill a bunch of civilians.
eviks · 17h ago
That depends on those two imaginary countries. Also, are you unaware of the many instances where people faced no accountability even when it wasn't a mistake, but deliberate?
kelipso · 16h ago
This would make the accountability situation much worse.
That said, I do fret we’re staring down a new age of guerrilla warfare. Drones are cheap, widely available, and increasingly autonomous. Their countermeasures are either impractical for communities (AA Cannons or automated firearms) or costly (jammers, interceptors). The programming can be set-and-forget, meaning operations can be staged months ahead of deployment and make it difficult to find or prevent. The autonomy of target termination specifically raises concerns for the immediate future of violent uprisings, coups, and civil wars.
As an engineer, I am fascinated by it all. As a human, I am horrified that we democratized violence on this scale.
State formation tends to track the relative military effectiveness of large highly-trained standing armies vs. small distributed arms making. The Roman Empire collapsed when they ran out of money to pay their legions. The smaller tribes and kingdoms of the Early Middle Ages unified into the larger kingdoms of the High Middle Ages as the longbow and mounted knight gave the advantage again to large, highly trained standing armies. These collapsed into the city-states of the Rennaissance because the gunpowder musket rendered all the armor of the knights useless. Then the nation-state took over as mechanized arms and airplanes became military weapons, and needed the resources of a large territory to produce them.
It's likely that the drone, being both cheap to produce, easy to use, and extremely lethal to existing weapon systems, will produce a similar political revolution. And it seems tailor-made for smaller political units: drones can lay waste to an invading army, but they suck at power projection because their range is only ~10-20 miles. Might we see a return to city-states as the primary form of political organization? Maybe all the arguments about whether Russia vs. the U.S. vs. China will come out on top are moot, because the very concept of a nation-state will disintegrate, and instead we'll have Beijing vs. Shanghai vs. Shenzhen vs. Moscow vs. Kiev vs. the Bay Area vs. NYC vs. Washington DC? Drones are also ideal for defending shipping lanes, so perhaps we'll see a loose confederation of economically-bound city-states, but each having their own culture and social laws.
Guerilla use of drones need off the shelf microcontrollers from somewhere, they aren't fabbing them in their backyard.
It's all about China. They have the ability to cut off drone production for the rest of the world.
If you can buy the parts, then yes, but producing a complete drone all on your own is not that easy I think.
Still easier than a stealth air plane or a cruise missile, though. So your predictions might come true, because I also see most state armies being really slow to adopt to this new reality.
How do you even defend against this in a terrorist use case? When a small drone with a grenade or homemade explosive is so accessible? Any Christmas market in central Europe these days is surrounded by car barriers to prevent mass run-overs, but what do you do when soon someone has the idea of dropping some molotov cocktails from drones in public places? Answering my own question I guess you can already throw one manually without a drone. Securing public places is weird, I'm glad it's not my job.
Also, the US military has been stockpiling kinetic drone countermeasures for about four years now. The idea is you get a hardened, ~11 pound autonomous drone that slams into the target at roughly 90 mph and physically destroys it before returning home. Add on 1-2 year-old US EW technology that now disables autonomous drones (yep, even autonomous drones), and you can establish a very comprehensive defense. The point I'm making here is that the tech is not only possible, but it exists.
Is it perfect? No. Though defense against firearms and explosives today isn't perfect either. Namely because of response times of the countermeasure. So in that sense, we aren't entering a uniquely dangerous situation.
Edit:
I think what will happen is that the first time a UAS is used on civilians, flying drones around population centers will be banned without permit. That way, if a drone is seen flying without a permit, it gets taken down on sight.
And eventually terrorists catch up. (Probably after the Russia-Ukraine war ends and some skilled people from both sides are unemployed)
It's infeasible to blanket every inch of civilian space with kinetic or EW anti-drone systems.
They may become commonplace at mass events (concerts, parades, gatherings, etc.) but will never cover all soft targets.
And unlike the nearest analog in chemical weapons, drones are dual-use, stable, and easily assembled.
The only reason the 1995 Tokyo Subway attacks [0] weren't worse was because of ineptitude.
Someone could be a quarter as intelligent and successfully fly an FPV drone into a target.
[0] https://en.m.wikipedia.org/wiki/Tokyo_subway_sarin_attack#Ch...
Also motivation and incentives. The reason we haven't seen many drone attacks on civilians is that it is far more lucrative to get civilians to buy your product than to kill them, and the companies that actually have the resources to mount a credible drone attack are making a lot more money doing the former.
Terrorism in general has always been far more overhyped than actually a problem - the median number of terrorist deaths per year in the U.S. from 1970-2020 is 4, making your odds of being killed in a terrorist attack significantly lower than being struck by lightning. And the reason is simply that it's deeply irrational. What do you have to gain from killing a random stranger?
But requisite minimum skillset is also a consideration, and that's where drones are dangerous.
We're talking (play videogames and some soldering) instead of (chemistry or biology).
You can't, and you don't. That is why it is called terrorism. Safety and freedom are sometimes antagonistic goals. This is an example. Terrorism is defended against by not changing society despite the terrorism. It is violence with a political goal, if the politics do not change, the terrorism fails. Not every soft target can (nor should be) hardened, there will always be soft targets.
On the same topic it reminds me Splinter Cell: Chaos Theory (from 2005!) where an AI called "Masse Kernels" was automatically creating missions sent to different PMCs, managing war assets, supervising the war effort, coms, and the like... Feels also that's coming, at least in some forms for now.
At the weekend, I was at a small town which had a half a mile long main road blocked off for a market day. They put up bollards to do so.
Chances are, there were 0 persons planning a car attack on it. So there was an element of "We don't think anything's going to happen, but if it were to happen, we're prepared.". A bit like having a fire extinguisher when there's never been a fire.
But would seeing the bollards also have the effect of discouraging the insane people of the idea of driving the car through the crowds the next time a market day is held?
Oppositely, if they didn't put up any barriers, a psychopath seeing this and the realization that cars can be weapons might give them the idea of "I know what I can do for my act of terrorism..."
* Flak/Shrapnel/Birdshot: An excellent last-minute defense if you’re calm enough to line up an accurate shot, but data shows that equipping civilians with these sorts of weapons en masse is a bad idea for safety and well-being. That’s a no-go.
* Nets: Popular for defense, but it’s a matter of time before drones adapt by flying under the nets or changing payload to something to dissolve it. A kamikaze drone could also be enough to destroy an opening for more to swarm. In a civilian context, they’re an excellent deterrent for high-population areas, for now, albeit unsightly.
* Buildings: Safest for now, provided the structure is relatively hardened and the windows are secured. But most civilian structures aren’t guarded against explosions or external attacks, and even those that are require a human to vacate it eventually. Once inside however, there’s more options for stopping an attack - for now - like interior netting, small arms with pellets or buckshot, or even lasers to blind the optical sensors. Impractical for civilian deployment at scale, presently, and highly variable.
* Jammers: Good against piloted drones, but as the article points out, the current crop of dev work is geared towards autonomous slaughterbots instead of human decision-making. Jammers are restricted by most countries and, if left functioning after an attack, could hinder first responders. If left on constantly, would disrupt civilian work. So that’s a no-go.
* LASERS! Probably the best deterrent in the short term for civilians, I would wager. A randomized strobe of a high-powered IR laser could devastate a swarm of drones’ optics, making navigation or target acquisition difficult or impossible. Sticking a piece of protective glass on the sensor would likely nullify it long enough to finish its mission, though.
And that’s what distresses me, ultimately. The future depicted in Slaughterbots or Horizon is rapidly approaching, where autonomous drones can murder with impunity and are affordable enough that any threat actor could get their hands on it. Combined with modern databases of humans - faces, biometrics, profiles, locations, habits, schedules - we’re nearing an era where assassination or murder is a drone away.
That is what horrifies me. And if there’s one thing my time in the defense industry taught me, it’s that nobody is trustworthy with that kind of power. Companies making these absolutely will use them (or condone their use) against dissidents, opposition, regulators, and governments. Pandora’s Box is already open, and I don’t think enough folks appreciate the horrors it will bring.
Worth noting here that the Ukrainian armed forces have already repeatedly deployed drones with the ability to spray pretty impressive amounts of napalm all over their targets from fairly high altitudes. Fittingly, they've been called "Dragon drones", and I wouldn't want to be under any anti-drone net if one of those arrives.
https://www.youtube.com/watch?v=DkiCWiDs2kQ
[0] https://en.wikipedia.org/wiki/Thermite
[1] https://www.youtube.com/watch?v=XttJet5WkTk
I can see this from the other perspective. I've read stories for the past 30 years now about police forces and swat teams abusing people, murdering people because someone filled out the address wrong on a warrant, etc. And I wonder, in the coming years, if those sorts of scenarios will be quite as one-sided as they are today. What will that world look like?
>Answering my own question I guess you can already throw one manually without a drone.
Sure, but can you throw one from 2 miles away, can you throw 30 simultaneously? Can you then instantly escape without much of any evidence of who you are being left behind, not even a blurry traffic cam picture? The scale of the mayhem is a quality all of its own.
Whatever fear Russians have of drones isn't in their offensive use. They're only being used defensively by Ukraine. Similarly, if you're a jackbooted goon, you might well fear their defensive use as I outlined above. But yeh, you have nothing to fear from the defensive use of drones, because you're not attacking anyone... in the coming years, however, the narrative will likely be twisted so that you do come to fear such people as might use them defensively, because the establishment needs you to despise them.
Jammers don't work against optical cable or AI vision controlled drones. That's a big problem today in Ukraine for both sides.
As for defense, first of all it's detection and tracking. Copters and long range gas powered drones are very loud and easily detectable. Ukraine uses a net of cell phones. Several devices with microphones can accurately pinpoint all drone like sources in real time. That's cheap to install miles around important targets. Then we need just fast AI interceptors 'on hold', in the air if can afford. The last part is missing today, but we'll get there soon.
As for danger, etc. Small remote controlled firearms were easily available for decades. Drones _are_ trackable. When one takes off in big city Russians know immediately where. By using radio scanners. All DJI drones, and most others, communicate and simply broadcast their coordinates. This is used in Ukraine to find their operators.
Yeah if we could just all agree that from now on, all warfare is limited to drone-on-drone engagements.
That was an episode of Star Trek.
Two warring cultures simulated war. The computers told each side how many casualties there were, and people reported to the extermination chambers.
This is something I haven't considered before. What's the worst case here? Is it feasible for me to go live on a farm in <country I want to harm>, buy a fleet of DJI drones at flea markets etc, stick something harmful to them, then hide them in the woods.
I can move away, wait a year or two, and then have them fly to the nearest metro area and wreak havoc. This seems to be cheap and relatively straightforward, and hard to detect. What am I missing?
Yes, that approach is inferior to the drone version. You have to hide them inconspicuously, and a bomb sniffing dog could find them. But you can visit a lot of places in a single European country or US state within one day, and unless the country is already on high alert you can hide something for that time span in public. Yet this doesn't happen. Even regular bombings are rare.
The reasons are manifold: In most places getting explosives isn't actually all that easy (unless you go the homemade route) and is a good method to get attention from authorities. But another factor is that there just doesn't seem to be a large interest in doing that kind of complex attack unless there is already an ongoing civil war. Actual terrorism is fairly rare, and the terrorists tend to be not all that sophisticated.
Are these kinds of drone attacks a scary new possibility? Yes, absolutely! Are they likely to happen? Not really. We might see it as a method to assassinate officials (imagine staging drones at a place where you know the US President will hold a speech in a couple months), but I doubt it will play a major role against the general population
Back in the day, if you forgot a bag on a British bus the driver would get it and run after you, so that it wouldn't be a bomb issue taking the whole day.
For long running conflicts (Israel vs Iran for example) I expect we'll see some fascinating and horrifying attacks in the near to medium term. Of course anti-drone tech is also evolving quickly and I expect that to continue so the shelf life of any specific attack will probably continue to be rather short.
Battery degradation, a year or two's worth of leaves and debris accumulating on and around the drones, literally all of the elements affecting them, animals, etc.
Things you can't help: they will discover the remains of the drones, and also their origin. This evidence will eventually lead back to you (unless you have the aid of a enemy nation-state). Not a big deal if you're dying in a suicide attack, but maybe you don't want the extended vacation in the CIA's worst black ops rendition site.
Violence has always been pretty democratic - you've always been able to punch someone or hit them with a rock and the US seems to have more guns than people.
This is something that I think escapes engineers in this line of work - that something they invented will eventually end up (legally or not) in the hands of people with no scruples.
[0] - https://www.msn.com/en-za/news/other/drones-in-africa-are-a-...
Dutch soldier lives have been ruined because they had to be sent to places like Lebanon and Bosnia. Nobody decent deserves that.
It may motivate actual reform in policing because law enforcement will realize that police officers who kill innocent people with no regard for the law are safer in prison than out on the streets with a paid vacation / desk job punishment.
As it is vigilante action against law enforcement in the west is a sure death sentence and probably life long reprisal against your family once you're dead which is what keeps people in line.
If the development of drone technology significantly reduces the risk of that then you're likely to see many more people respond to violent abuses of authority by law enforcement with vigilante action.
Are jammers really that costly?
Also spoofers that could take over a drone - not sure how much encryption is used in most of these off-the-shelf drones, but it would seem like it wouldn't be too difficult to create a Flipper Zero-type device that could spoof the codes used between controller and drone.
[1]: https://ardupilot.org/
[2]: https://www.theatlantic.com/international/archive/2025/06/uk...
[3]: https://github.com/ArduPilot/ardupilot
[4]: https://www.404media.co/ukraines-massive-drone-attack-was-po...
https://gist.github.com/andrew/2f81952f4867d1b200bb
The big difference is they can now run this on the copter instead of being remotely controlled; a 100$ raspberry pi has enough processing power for this, and so does several other off-the-shelf mini computers powered by lithium batteries.
Crazy times.
> ...each of the 117 drones launched had its own pilot.
https://www.bbc.com/news/articles/c1ld7ppre9vo
From Zelenskyy, a previous comedic actor refusing to flee, "I need ammunition, not a ride," to the defense of Snake Island "Russian warship, go fuck yourself," to all the brave women who volunteered, the farmers towing abandoned Russian tanks, the constant drone attacks on residential and commercial areas, the 40,000 stolen Ukrainian children, this most recent attack on Russian air bases...
If this was a movie, I would probably think it was a bit much myself, but this all happened. We witnessed it.
"Glory to Ukraine! Glory to the heroes!"
https://www.rottentomatoes.com/m/bedtime_for_bonzo
Stick the GPS coord, fly there, and once in a geofence look for a shape to crash into doesn't seem impossible given what was possible 10 years ago.
The context (together with the features extracted) was the killer (forgive the pun) feature though - everything else reduced noise, but context increased signal.
My gast remains flabbered that the sort of thing I was working on back then hasn't become commonplace in the interim. The computing power available today, compared to then, and the accuracy we had (I know for a fact at least one of the designs was made into real hardware, it was called RH7, and "RH" stood for "Red Herring" - oh how we laughed) ... It beggars belief that it was just left to digitally rot.
To me the more interesting question is how they managed sending the real-time video feeds and control data. Since the trucks were mobile, I assume it had to be via a bunch of mobile phones signed up to Russian service providers since Starlink doesn't work inside Russia. To reduce latency, I wonder if the phones were connecting to a covert site in Russia which had a high-bandwidth wired link, maybe a front company established for the operation with servers and broadband internet connections.
Compensating for wind drift is a fairly straightforward software problem when you've got a fast processor, a bunch of high-resolution cameras and a laser rangefinder.
https://www.autelrobotics.com/productdetail/evo-lite-enterpr...
https://www.bbc.com/news/articles/c1ld7ppre9vo
"A military operation involves deception."
He could be telling the truth, he could be lying... A drone programmed to automatically boot up , check its location, and if it's at the right coordinates, take off and crash at some other coordinates (the airfield) is more satisfying to "fans" of automated warfare.
For extra fun, add some other code to "look for plane-like objects to crash into", but now you're approaching dangerous territory of "What if a civilian 737 happens to be boarding at this airfield"...
The reports also mention the truck roof opening remotely, one could also use GPS coordinates to trigger this. But doing it manually from a distance, after checking the surveillance cameras that the coast is clear, is more reliable.
I guess they used smartphones and SIM cards with mobile data for the remote communication...
Civilian 737 boarding airfield where Russia keeps strategic nuclear bombings? Russians would shoot them down faster than any drone could get them.
I'm not sure if the Tu-95 is hosted at any joint-use airports, but joint-use themselves airports are not uncommon. Pskov is joint-use, Ukraine launched a smaller-scale attack on some Il-76s there a couple years back. The scenario that an attack on legitimate target aircraft could be happening metres away from civilian aircraft is realistic.
https://en.wikipedia.org/wiki/Malaysia_Airlines_Flight_17
Good old non-AI radar-guided missile launched by human crew of Russians.
https://www.kyivpost.com/post/53784
A static target only needs to be seen once.
…but if GPS is jammed, and there’s only one camera per fleet, how exactly are the other drones supposed to navigate towards the spotted targets unless they’re all equipped with cameras?
Camera drone hovers above target and kamakazi drone intersects the line between camera and target, and drops.
That being said, having all drones equipped with cameras could enable a more robust version of what they’re talking about:
If uplink with human operators is lost, but short-range comms between drones exist, they could use their video feeds to autonomously coordinate amongst themselves.
We've had two years of footage of drones being flown over tanks, and bombs dropped directly down into them.
We have two years of footage from Ukraine, where camera-equipped drones are launched from a several miles away at most, and where there are networks of pilots and support specialists to assemble and launch more drones in case of (frequent) failure.
I don’t think it’s wise to wager the success of a 6-month mission deep in enemy territory on a plan with a single point of failure, especially when the alternative is equipping each drone with < $100 cameras.
But sure, you’re clearly the better thinker.
I've had that happen to the company I work at and we literally have zero AI stuff.
AI gets so much boost from this nonsense. Because now it's about saving our lives.
"ArduPilot can handle tasks like stabilizing a drone in the air while the pilot focuses on moving to their next objective. Pilots can switch them into loitering mode, for example, if they need to step away or perform another task, and it has failsafe modes that keep a drone aloft if signal is lost."
So it is not fully autonomous.
You can also see the careful departure of drones from containers in the videos, without extra panning or yaw. Not quite how a human operator would fly them.
https://xcancel.com/lemonodor/status/1929269307189469624
If the attack was coordinated this way, I assume whoever sold the colo to Ukrainian intelligence thought they were simply setting up yet another server for a shady Russian scam company. Foreign intelligence services often avoid scrutiny by using the same methods as domestic criminals in the target country.
Now in its third generation, the Ghost Dragon has come a long way since 2022. Its original command-and-control-band radio was quickly replaced with a smart frequency-hopping system that constantly scans the available spectrum, looking for bands that aren’t jammed. It allows operators to switch among six radio-frequency bands to maintain control and also send back video even in the face of hostile jamming.
Immediate post-WW2 vintage. The classic design of AA gun.
As far as I understand it from talking about Turkish drones, you mean https://en.wikipedia.org/wiki/Bayraktar_TB2 style, aircraft size drones, rather than the quadcopter size ones? The latter can more easily hide in terrain.
So being stealthy in the radar spectrum is pretty difficult, and I often wonder if stealth planes are mostly a means to transfer money from the state to defense companies.
A lot of the key AA tech that has suddenly become important in the era of drone swarms began proliferating to mid-tier forces around the 1980s (or earlier), and was retired by well-funded armies between then and the 2000s, because compared to SAM systems, it was suitable only for lower, slower, less capable targets.
Turns out, suddenly large numbers of lower, slower, less capable targets are being fielded, and its really expensive to take them on with SAM systems optimized for dealing with modern manned aircraft, cruise missiles, and/or ballistic missiles.
No comments yet
Inside-out SLAM strategies and on-device ML are much more interesting and are starting to trickle into COTS drones. For example, the latest DJI drones all use SLAM for return-to-home even when GPS denied: https://www.facebook.com/reel/440875398703491 , and the latest Matrice 4 enterprise drones also have end-user ML model runtimes that can fine-tune flight plans using user-provided logic.
Inside-out last-second targeting is also very popular in Ukraine, with off-the-shelf "find the nearest car/person in analog video, lock to it on signal lost, and send Betaflight MSP stick commands to hit it" modules readily accessible on Aliexpress.
https://youtu.be/5xN__ozrbpk?si=vuBtFEcOlgerrVwa
I specially apretiate the small mine clearing drones.
Come Back Alive ex. These guys delivered first deep-strike drones https://savelife.in.ua/en/donate-en/
Serhiy Prytula Charity Foundation ex. Bought a famous spy satellite https://prytulafoundation.org/en
KOLO Charity Foundation managed by UA tech community https://www.koloua.com/en/
Razom Ukraine (US based) https://www.razomforukraine.org/
However, the decision-making based on image recognition mentioned in the article is undoubtedly more effective in more changing fields, when the target is moving
[0] https://en.wikipedia.org/wiki/Silent_Hunter_(laser_weapon)
[1] https://www.rtx.com/raytheon/what-we-do/integrated-air-and-m...
https://eos-aus.com/defence/high-energy-laser-weapon/
What they are good at is target tracking, having started out in satellite communication.
Their tracking system paired with a 30mm Bushmaster cannon and proximity ammo is another solution, and there are apparently 160 of them heading for Ukraine to be mounted on M113 and Kozak vehicles.
https://eos-aus.com/defence/counter-drone-systems/slinger/
Having a laser that spreads out to e.g. 30cm radius at 500m is not hard to do if you need an area of effect weapon & can push enough power (ie. your laser is powerful enough, but not so intense that it ionizes the air & blocks itself). Reflections seem like a bigger problem: If the most effective defense includes guys with shotguns &/or there are a lot of unprotected personnel in the area, how do you make sure stray reflections don't end up blinding them?
Remember, this is about asymmetric warfare. If the number of rounds or amount of energy required costs more than the drone it shoots down, then it's not an effective deterrent. Militaries are looking for single-shot weapons to take down drones. Fire once and move on. It's the only way to deal with a swarm. Think about it for a bit and it will become very obvious.
https://www.linkedin.com/posts/israelimod_israel-mod-complet...
https://www.yahoo.com/news/historic-breakthrough-idf-reveals...
>In a historic breakthrough, the IDF on Wednesday announced that an unnamed laser defense system similar to the much celebrated Iron Beam laser system has shot down dozens of aerial threats during the war.
>Already in fall 2024, The Jerusalem Post had learned that the IDF had used laser defense systems in operational situations but was barred from reporting on that at the time.
It was all over Israeli news together with videos of operational intercepts.
it has nothing to do with "february cookoff" of different systems and orders for iron beam were placed in january with deployment by the end of the year
Do you have a credible source from the IDF attesting to this?
And do you believe the IDF would conduct a demonsttation after it has already committed to a system?
https://www.timesofisrael.com/liveblog_entry/laser-intercept...
it been all over Israeli news. not exactly "news"
>And do you believe the IDF would conduct a demonsttation after it has already committed to a system?
what demonstration are you talking about ? if you are talking about whatever was published in february, it's mostly for more tactical/mobile use i believe.
there is an issue that north of israel is very hilly, so it's possible to fly drones from lebanon below radar visibility range and then just to get them pop-up 50km away from border. it was major problem last year and the publicized trials a believe concentrated on sourcing systems to solve this issue.
The problem with any point defense system is radiating any energy makes you a big target. So you would want a passive (EO/IR?) or triggered active/passive system.
For those uses, there's a fairly decent approach ["missile"] or hover ["bomber"] stage that is probably plenty vulnerable to autonomous PDS via 12 guage medicine.
Tracking / detection could even be passive, partly acoustic, partly EO/IR, with only a small fire control radar if you really want it.
I don't think western common folks grok how depraved that country is in terms of doing good work, reward systems for such and corruption on every single level. puttin' built a mafia state and pushed this behavior from top->bottom, and these are side effects. Not some soviet competence and discipline, which wasn't stellar either but light years ahead of current state.
1) Russia suspended the treaty 2023, so it is not relevant here.
2) The use of hangars is not prohibited, as can be seen by the climatized hangars that the US keeps their B1 in.
3) Russia has announced plans for such reinforced hangars years ago, but very likely some dacha or yacht had higher priority.
When I said "triggered" I meant you would enable it when under attack, at which point it doesn't matter if they know you're there anymore.
Bigger picture - if knocking the laser defense off-line slashes the unit cost of destroying bombers, then it may be the obvious first move in any competent attack.
I can imagine this protecting some future US bases in same way C-RAM is used. But from what I read from ie Iraq veterans they had it turned off most of the time for the fear of shooting down its own planes. So much for trust in high tech if its too powerful and automated.
Chinese have some systems, but from demo I've seen the laser beam took some serious time to shoot a single missile. Drones are smaller and way more fragile (so also harder to hit) but this ain't Star trek or Star wars.
(1) if prop based, launch something to snare the props (2) if reflective, pre-launch something to spray black non-reflective paint at it, and followup with laser (3) if evasive, approach with random manouvers (4) if unknown, launch everything and see what works, and feed it back to the training data ... etc, etc.
So the cartoons lied to me?
I’ve read about a bunch of these systems even if they aren’t in widespread deployment some are still being tested in real world conditions.
What about: https://en.wikipedia.org/wiki/Silent_Hunter_(laser_weapon)
“The Silent Hunter has been used by Saudi Arabia to guard against Houthi drones and missiles.”
“During the World Defense Show in Riyadh the February 05, 2024, Poly Technologies announced the first hard-kill engagement of a one-way attack drone.[6]”
The targets, whether it's plumbing pipe rockets or lipo drones, come in at 100-1000 yards/sec, so you don't really have that many seconds per target.
They work in the demo in which you just shoot down the sole target as it fly perpendicular to the machine for both physical and career safety, but when it comes to deploying the thing around your bed, guns make a lot more sense.
Also, your objection doesn’t really fit how drones have been used. Massive highly coordinated drone swarms are extremely unusual, the threat is mostly individual drones or small clusters.
Many very dangerous drones are well under 100y/s aka 200mph.
1000 yards/second aka Mach 2.7 is well beyond the ‘drones’ people are concerned with and into expensive missile territory. Which is where anti missile systems get used.
full blown deployment by end of this year
Or miss by a long shot and hit a civilian instead.
Note that his laser burns through various reflective materials, including mirrors, copper, aluminium, and steel.
[0] https://youtu.be/UBVlL0FNbSE
No comments yet
The Ukrainians pulled off an absolute coup on Sunday. A third of a nuclear-armed country's strategic bomber fleet inoperable for the foreseeable future. Someone at NORAD probably said "they should have sent a poet" while looking at the satellite imagery.
If middle powers like Ukraine can do that to Russia, they can do that to countries like the US. We need to be on their good side.
That's how Trumpism can gain any traction at all. The amount of international engagement Russia had as Putin made himself tsar was embarrassing, and to a person with no scruples if the money is right - like Trump - it just illustrates that the guardrails aren't really there.
In US.
There's also an argument to be made that Europe more or less sold themselves out for cheap natural gas. Nordstream 2 was constructed after the Russians (at that point, under Putin's puppet Medvedev) had invaded Georgia.
“The revolution eats its partisans” is the most accurate description of it. People on “the good side” turn against their peers for not being on the good side enough. To wit, people who turn away don’t generally first notice that the good side isn’t so good; they first notice being bullied by that side, then they reflect on what it means to support the good side’s points of view (spoiler: A crime against humanity).
Ukraine borders on Russia, but the US is separated by ocean from serious threats. Attack by UAVs of this sort seems nearly impossible.
Plus imagine if those attackers realize they can ship those containers from Mexico or Canada.
Could take a while to figure that out. It took days/weeks to figure out the 9/11 attacks with some certainty.
A good reason enough for a lot of people, no?
I think the key deterrent is that the U.S. has production capacity for the important systems and overwhelming capacity to strike back, so a rational foreign state isn’t going to think there’s a way they win by trying it. Terrorist groups might be a different story, so I’m really glad this wasn’t an option during Bush’s big adventure in Iraq and Afghanistan because however bad drone strikes are for military defense, they’re even worse for civilians.
I never considered drones, which is even more obvious, in hindsight.
The first casualty of war is the truth.
That being said, any sort of materiel loss on weaponry as important as strategic nuclear bombers is a massive problem for Russia. The logistics of repairing them, if possible, is going to be complex.
This was a massive damage, not directly interfering with war against Ukraine that much, but overall power projection. Plus a pretty good insult to russian's FSB and GRU services who had no clue, just like today's Crimea bridge blow.
I do not want to give any credit to russian secret service but as we have seen in 9/11 and 7th October - secret services in any country are sometimes clueless.
Those airplanes are one of the things that give Russia a second strike capability, and if they lose that capability, then they are going to be on a hair trigger in a nuclear crisis.
That's the problem in a nutshell. A few years back, few would argue against keeping a human in the kill/no-kill decision chain. It just took one war to get pop tech authors writing on it without even a mention of the ethical considerations or autonomous killing machines.
That kind of operation seems extremely different from a stationary turret or patrol robot with standing orders to shoot upon arbitrary targets at any time it decides to.
I suppose it highlights axiomatically the terribleness of ethics when they must be defined in a might-makes-right manner. All very high minded and complex questions which leave the awkward question unanswered: what are we supposed to do?
EDIT: Apparently the reference wasn't clear.
https://marvelstorybook.fandom.com/wiki/Drone_(Oblivion)
https://www.npr.org/2023/09/06/1196975759/ukraine-cyber-war-...
You’re telling me that you’d rather this was done by a stressed, emotional, tired front-line private?
Reminds me of the debate about a self driving car that might need to mount the curb to avoid hitting a car - and therefore endangering a pedestrian.
It’s not an easy decision but I’d rather a machine made it than a stressed person!
That is, historically to wage war you had to ultimately convince/coerce millions of people to wage that war for you -- continuously for years -- along with millions more to support the effort politically and economically. In the highly automated world, you literally only need a lot of money (which also tends to concentrate outside of any semblance of democratic control).
The threat ended up being a false alarm, and that human judgement saved a lot of lives. A machine, assuming it would have launched when seeing that signal, would've ended differently.
https://www.vice.com/en/article/former-us-airman-tells-un-an... ("Former US Airman Tells UN an Accidental Nuclear War Was Narrowly Avoided in 1962")
https://news.ycombinator.com/item?id=10452983 ("The Okinawa missiles of October 1962 (thebulletin.org)", 71 comments)
What does that mean in practice? Have you ever saw a flash in the corner of your eye, looked for a second, then realised it was just a reflection? For that split second, your brain identified a potential threat and was trying to quickly decide whether to hide, fight or run. But underneath that was a nagging of "it's probably nothing," that caused you to delay pulling out your gun and opening fire on a passing car.
A computer programme may or may not have this coded in. A machine designed for an active war-zone, or a car assuming it'll only ever auto-pilot on a motorway, is used in a suburban area. That same programme that's fine in its original intended location suddenly opens fire on a grandma with a shiny stroller, or swerves into a pedestrian because "pedestrians don't walk on motorways."
That's what we worry about.
.. or vice-versa. The average war has plenty of stressed people firing semi blindly at half-identified shapes. And if we want to be literal, there's the case of Lee Clegg and the exact circumstances in which it is legal or not to open fire on a passing car and kill a teenage girl.
That's the same with these drones. The smarter they get, the further away the human goes. Today it might be simple to create autonomous weapons who are instructed to kill vehicles matching various known appearances. That too already exists. The strike on the Russian bombers was reportedly carried out manually, but it would have been pretty easy to have that autonomous, since the targets are huge, stationary, easily recognizable and easy to navigate to in the geography.
If you launch a quadcopter and instruct it to kill any adult human it finds, then that's the same thing. You wouldn't launch it into an area where there is a remote possibility of being any civilians. No difference from firing an artillery shell. If there is a civilian, or a soldier waving a white flag or whatever - there is no cancel button for your artillery shell. The decision to kill whatever is in the other end was made when you fired it. There is literally no difference between firing a million drones and firing a million artillery shells down range. It's your human responsibility and your human consciousness when you make the decision.
I don't think we have had widespread use of autonomous human-targeting drones yet, but it's by no means science fiction today. Just a matter of time. We'll see their use in this conflict.
All the bombs Russia thrown onto Ukrainian civilians were thrown by human soldiers.
Yet. Drones also don't get tired.
When a misconfigured computer does something wrong, it frequently does it over and over and over again until it is prevented from doing so by an external intervention.
No tired private is going to mistakenly rampage through a populated area mowing down civilians. But a confused drone swarm might.
At a micro level, sure, you could argue that kill/no-kill decisions are being made by frontline enlisted soldiers, but the stress, the exhaustion, and being able to see your enemy downrange add a degree of humanization and discretion.
Wars are brutal. Some warring factions are incredibly brutal. But taking a human life can't be reduced to a `KILL? [Y/N]` decision made by some MAJ somewhere. Killing enemy soldiers is a last resort (formally, this is the concept of _military necessity_), and normalization of it as anything else is a mistake.
I also think it's a fallacy to compare this to the self-driving car trolley problem; the timeline of decisions, length of the decision chain, and ultimate goal of the decision-making process are _aggressively_ different.
My overarching points are, I guess:
War is a tragedy, obviously, but I don't think there's a way to avoid it right now. In absence of some way to stop war forever (hah), we can't trivialize taking human lives. It's not lost on me that this has been happening for a while at this point, and that it's getting worse. I still oppose it, and I think we all have a responsibility to be more critical of this regime of warfare.
I don't know how to counter the argument of "well, if we don't do it, _someone else_ will, eventually." This is some really fucked up, self-justifying inductive reasoning that can't easily be countered by calling out the moral bankruptcy of the premise. In the past, mutual disarmament treaties have been a down-the-line bandaid for this kind of thought process, but the nuclear rearmament we're seeing in the world right now shows it's not a panacea.
War is studied. There are journals, papers and research on war fighting at all possible levels.
In the most recent action by Ukraine you can observe actual reality: what did they attack? Military equipment of the enemy. Why did they attack it? To degrade the enemy's ability to sustain and rotate their forces attacking them. What was it for? Well for one thing it will hopefully considerably reduce their ability to bomb civilian targets.
My response was more aimed at the parent comment to my previous one, which seemed to paint delegating kill/no-kill decisions with a brush of "I don't know why this is such a big deal."
Targeting a stationary Tu-95 bomber with no protective measures in place is probably the easiest possible identification task for a drone.
A lot of kill/no-kill decisions are more subtle, or involve unknowns, possible nearby civilians, etc.
Or look at this crazy FPV piloting job. You think AI could do this? Pilot maneuvers through an absolute maze of anti-drone nets to hit a moving truck. (essentially SFW; video terminates slightly before impact)
https://www.reddit.com/r/CombatFootage/comments/1l2aqxp/ukra...
Drone operators are generally not on the front lines, and sometimes they are literally on the other side of the world sitting in a cubicle. They are usually specialists of some sort, not untrained privates.c.f.,
https://notabugsplat.com/
and reporting of the use of this term by U.S. drone operators.
It seems trivial to confuse a Tesla's AI. I'm assuming they're fairly near the top of the game when it comes to that, yes?
https://www.youtube.com/watch?v=U1MigIJXJx8
This sort of intentionally hostile pathological case is of course rare in real-world driving. It will not be rare in warfare.
And a drone has to operate fully in three dimensions, unlike a Tesla which is effectively operating in two dimensions.
An autonomous drone will also have extremely constrained computing resources relative to a Tesla due to size/weight/power constraints.
No comments yet
It's hard to imagine there is a military in the world that would put something so important in the hands of a private. Yes the drones in this war are relatively low cost but they are also so vital to both sides' war effort that there is inevitably more command and control over their use.
At the end of the day, it's still humans deploying these weapon systems and accepting the risk that they might cause unintended casualties.
A machine also can't be held responsible.
Who do the surviving relatives of the crush pedestrian sue? No longer the driver i guess
Autonomous cars wouldn't speed and they are always paying attention. Maybe you could argue that it happened because the other driver was drunk or for some other reason swerved into your lane but this is still an incredibly niche situation and if it does happen just sue that guy. The vast majority of car-to-car accidents happen when both drivers are irresponsible. Autonomous cars will significantly reduce accident rates just by following the rules of traffic and being constantly aware.
The whole who should be sued thing is asinine, just have a national insurance fund for it or whatever. Cheaper than all the accidents it prevents.
There was an article here serveral weeks ago about how in some self driving cars the brakes can be pre-loaded and can be used much quicker than a human could, reaction times aside.
Personally, I see anyone fretting about imperfect drones making those decisions as either (1) one of today's 10K, or (2) a performer covering a trending topic. Land mines were deployed at giga-scale over a century ago (WWI). Spring-guns were enough of a problem in England to be outlawed two centuries ago.
B) mines don't move by themself, but stay where you deployed them
A moving autonomous killer drone has a potentially bigger effect if unleashed on the wrong area. Besides, war is usually fought in civilian areas. If you send the drone to the frontline, but its GPS is jammed, it might move somewhere close by.
B) https://en.wikipedia.org/wiki/Naval_mine Not that anybody's doing Environmental Assessments on land, to determine whether the mines might be washed downstream after a heavy rain, or moved downslope by a landslide, or ...
The problem you describe sounds very much like poorly-directed mortar/artillery fire & strategic bombing - which have been killing civilians at scale since at least WWI.
>> You’re telling me that you’d rather this was done by a stressed, emotional, tired front-line private?
I want somebody that can be held accountable when it goes wrong. With automated systems we don't have that.
The visceral reaction is not there that would motivate the legal system to send the group that started a cascade of autonomous decisions by a robot to kill a bunch of civilians.