LorenPechtel
5 days ago
Whether or not they succeeded (and their claims generally do not prove to be exaggerated, so I suspect they did) they just shook up the world's militaries. I doubt there's a base in the world that's truly secure against this sort of attack.
And I don't see any reason to suggest AI being involved here. GPS to an initial point, follow a specified bearing and do image recognition. Note the angle at which you are seeing the image, if it doesn't change as you move you're looking at an image, ignore it. This would fall for any reasonably fancy decoy, but that's not going to be enough to protect the facility.
I've been saying this for years: with pretty much everything military a bunch of cheap units will give you far more bang for your buck than the same value of expensive units. It used to be that since units required operators you would have a big problem with personnel (although we did see the Kamikazes in Japan), but now that it's a microchip running it that's no longer a factor.
Nextgrid
5 days ago
> I don't see any reason to suggest AI being involved here. GPS to an initial point, follow a specified bearing and do image recognition
Is there a reason to suspect autonomous flying as opposed to remote control via the internet? Whatever deployed those drones can act as a bridge between the drone's radio link and cellular/public Wi-Fi/Starlink.
LorenPechtel
3 days ago
If they're not autonomous they can be jammed. Given the amount of drone warfare that has been going on I would consider jamming likely.
Also, humans controlling them means a lot of data flying back and forth. IP, bearing, search pattern and image recognition have been around in the world of anti-ship missiles for quite a while.
mrep
5 days ago
> they just shook up the world's militaries
For the good ones, I doubt it.
Israel today has trophy [0] which can detect if an rpg is going to hit its tank and shoot it out of the sky.
(from wiki): The system allegedly relies heavily on high-speed computational technologies. Upon detection of an incoming projectile, the system automatically computes various parameters, such as the approach vector, nature of the threat, time to impact, and angle of approach. The defensive projectiles are launched by two rotating projectile launchers positioned on the sides of the vehicle. These launchers deploy a number of small EFPs (Explosively Formed Penetrators), forming a precise and closely spaced matrix, targeting an area in front of the anti-tank projectile.
And that's one country with 10 million people and a mere $46.5 billion in military spending. And BTW allies generally share tech (we the US suck right now (sorry)) but Trophy is being integrated with multiple allies [1].
Developing automated drone shooting destroyers I think we can do.
EDIT: To add, I bet those drones can be shot down by 1 minigun or shotgun shell which aren't relatively expensive.
Edit 2, the dutch already have a badass automated minigun too [2, 3].
Edit 3: multiple countries have similar systems but they are all mainly for boats. I think we can adapt the to army bases [4].
[0]: https://en.wikipedia.org/wiki/Trophy_(countermeasure)
[1]: https://en.wikipedia.org/wiki/Trophy_(countermeasure)#Intern...
[2]: https://en.wikipedia.org/wiki/Goalkeeper_CIWS
serjester
5 days ago
There’s been plenty of anti missile systems deployed to Ukraine and the tank losses have still been absolutely insane.
Fundamentally it’s not about building anti-missile technology, it’s about doing it cheaply and at scale. That is a much, much tougher problem.
These drones cost less than a thousand bucks, and if each interception cost you 100k (the cost of a trophy shot) you’re going to lose a lot of equipment. It only takes one miss.
Now imagine scaling that system to protect an entire airfield. That seems next to impossible when these drones have a range of 10+ miles and are basically unjammable. You need China level surveillance across your entire county as a bare minimum.
mrep
5 days ago
Check out my edit 2. The Dutch already have an automated minigun. Bullets aren't that expensive relative to those shitty consumer drones which would get torn apart by them.
Also, ukraine doesn't have Trophy or any minigun/shotun defense system that I know of yet.
Having automated miniguns/shotguns near civilian areas definitely creates a challenge but I think our defense budget can handle that.
oliwarner
5 days ago
These naval CIWSs are massive. ~6 tonnes with a minute of ammunition. They cost tens of millions. All of this is probably acceptable for base defense.
But all high rate fire guns have short operation limits before their barrels melt. About 15 seconds before a 10 minute cool-off, maintenance and reload cycle. It's comically easy to saturate that with waves of decoy drones.
mrep
4 days ago
Those CIWS are designed to take out much more hardened targets than shitty drones. I bet you a shotgun shell can easily take out one of those drones.
Upgrade your drone defenses and now your drones cost more...
I don't work in the military so take everything I say with a massive grain of salt but I bet offense and defense scale in cost similarly and NATO + allies have WAY more money than the russians.
LorenPechtel
3 days ago
And note a big weakness of CIWS in port: what's your backstop? When the wave of drones comes in at low altitude over the city the port is at what happens???
a_vanderbilt
5 days ago
An increasing number of drones being used aren't the shitty consumer ones either. They are still relatively cheap, but their guidance systems and payloads are improving dramatically compared to the start of the war. Saturation attacks are still possible on automated kinetic kill systems, and for a price that is sustainable for the attacker.
mrep
4 days ago
> but their guidance systems and payloads are improving dramatically compared to the start of the war
But can they survive a small bullet or a shotgun hit?
jiggawatts
5 days ago
The Goalkeeper CIWS costs something like $3000 per second that it is firing. It's not a cheap system!
It could be scaled down to the size of a normal minigun, but even that is about $50 per second.
Meanwhile drone costs keep falling, Ukraine is well below $500 per drone now.
mrep
4 days ago
$500 for a drone including its bomb that can incapacitate a military aircraft and can survive multiple bullets?
Also, I don't think you need 30 mm bullets to take out a drone. Those bullets are for much more hardened targets.
jiggawatts
2 days ago
All of these weapons fire a 1-2 second burst to make sure they hit the target. Don’t forget that the drones are moving, there’s wind, etc…
Modern CIWS will use radar to track their own outbound bullets and adjust their aim in real time to hit the target.
LorenPechtel
3 days ago
That's built to kill an inbound missile. And it's almost certainly possible to bleed it for less than it's own cost.
What we need is cheap systems that can be deployed in large numbers that fire wimpy short ranged stuff so it doesn't do too much damage when it falls back. Shotguns are good at this, the smaller pellets do not fall back with lethal energy.
tim333
5 days ago
Trophy systems are like $2m and drones are cheap.
mrep
4 days ago
Trophy is designed for fast moving anti tank missiles. These cheap drones can be shot down with a shotgun. I think we can make anti drone systems a little cheaper.
gregw2
3 days ago
My speculative take is that this was not a surprise to the US military and they have been red teaming this sort of thing for years. Remember the strange civilian reported drones near bases, in the US last year (and in prior years?) Drills/Exercises... Alarm about Chinese buying land near US bases?