chrisl

But reusable means the drone has to return, so it needs more battery for the same range, meaning higher weight (or smaller payload). So one way has its place.
I think both (reusable) bomber-like drones and (one way) missile-like drone have been seen, including bomber-like drone carrying missile-like drones.

Either someone has made a real tech breakthrough, which is not impossible, or someone has found a pet congresscritter to slip something in the budget for there brother in law. Publicly known rail gun tech is not ready for this. I did see something about the Japanese doing a CIWS system with a 40mm rail gun, but 155? And if it works there no reason not to shoot it at land targets...
Edit: I think it is a misleading tweet, the Army is using some combination of a very high energy propellant, and the the emerging ram jet shell technology to boost the PROJECTILE developed for the navy rail gun program. They haven't lost their minds and tried to put an actual rail gun on wheels.

They cannot drop arty precisely on your head if you take down their observer drone. Well, taking down the observer drone reduces the effectiveness of even FPV drones. However, RU is just unable of taking UKR observation drones down, due to sheer stupidity rather than a lack of technology.
 
Yes. State program (1,000,000 drones program) and private production enhanced by unified military purchasing program really works. That's why I am laughing at the lie that UKR is the most corrupt country.  

I think that counter-drone, drones are a must. Sure invest in PD but without layered flying defence any ground force is done.  And even if you can APS everything, one has to protect logistics lines.  And then if one can protect logistics lines, one has to try and protect from ISR.
My advice is to go lighter but frankly after watching Russian soldiers getting hunted down in a drainpipe 15 grid squares away, I am not even sure that will work.

Another problem (possibly my biggest problem) with APS is how fragile the radars are - both in the sense of physical damage but also presumably in the maintenance required. 
One illustrative example is the fact that they don't bother putting lots of shots in the APS because by the time you have blown up one or two incoming missiles your radars are likely wrecked anyway. This can be exploited by sending in (just) 2 drones: one claymore type anti-personnel drone that explodes just outside of the intercept range to damage the radar, then the second HEAT one can attack without risk of interception - kind of like a tandem charge attack. 

If the pro armor/APS crowd is so convinced that current APS is effective then all it has to do to prove the point is ship a couple of brigades worth of equipped vehicles to Ukraine. While not cheap, this actually makes sense from a real world testing standpoint, before we bet the future of NATO land forces on the concept. If the vehicles so equipped wash the Russian blood off of their tracks in the Sea of Azov, well we have an answer, now don't we? Or at least some real data.

Even in some of the early videos of drone operators they were deploying their antennas far from where the operators were located.  Because the antenna does provide a return address.  
But you're right that antennas are cheap enough that they can just leave a bunch scattered around and o from antenna to antenna.  You wouldn't want to leave an antenna and go back to it - someone probably already has the targeting worked out for your location by the time you get back.  But you can have assistant drone crews that go some time after the drivers bug out (but not right away, in case there's artillery or counter-drone on the way) and relocate the antennas.

I challenge this without further data. Looking it online the operational history notes perhaps two dozen cases over about ten years.  Even if we assume it is twice that, that is pretty sporadic.  And then there has really been not been much on the details of the threat.  How many of these were advanced ATGM systems, and how many were older? The fact that the RA is using 30 year old tanks keeps getting brought up as rational, how old are those Hamas and Hez AT systems?
The one area the IDF does have the high ground on is urban terrain, I would concede that.
My problem with Trophy and all APS is that trying to prove a defensive system works because "we want it to" is really dangerous. If APS were being used as much as FPVs in this war we would have a much better dataset to pull from.  And a contemporary environment around them. To extrapolate from IDF small war/COIN operations to peer high intensity long duration combat is introducing too many error points.
I would argue that the exact same issues surround APS - "pinning down precisely how effective and finding out in which situations they aren't as effective to inform how to possibly counter them."  However, some have leapt past this and made some dangerous assumptions.
Let me underline something here - FPV/UAS dominance is not a good thing. When military theories break stuff like the Fall of France '40 happen. The only way to make it worse is to start wishing away these impacts and hiding in the sand. Grabbing IDF employment of APS and dragging it into this war simply makes no sense.  APS might work, we are going to see more of them because that is all we have right now. But wrapping ourselves in some warm blanket of "it will all be ok" is a bad idea.
Anyone who has been in combat will tell you to "become the mouse" - Herbert was onto something there.  The mouse is always afraid and aware. Never sticks its neck out unless it has to and never assumes it is safe.  That is what makes combat so exhausting. You never really sleep. Everything is a threat, all the time. You are looking everywhere at once. And through all that you still have to get the job done. There is no such thing as "paranoia" in combat; everything really is out to kill you. "God I hope this all works" is pretty much the mantra soldiers live by. This is what the thousand yard stare is all about...exhaustion: when a soldier just stops caring and looks at nothing. Time to rotate them off the line at that point.
So we need to stop cherry picking our data here and look at it all.  Right now we know EW works...to a point, but it is not enough to stop hundreds of strikes per week. Cope cages and barns don't really work. Guns shooting into the air do not work. APS has not been seen on this battlefield by either side really. ATGMs still work. Artillery definitely works. FPVs/drones are definitely working. ISR is working. Manoeuvre is not working. Air works but is stand-off.  Cyber is out there but has not been decisive in any measurable way. Infantry still work, but they are stressed and their vehicles do not survive long.  Tanks do not work as intended.  They are lobbing shells and coming forward to snipe but in their main role as the steel tip of manoeuvre they do not work. Logistics works but I suspect it has made some major shifts.
This battlefield has not gone static because the Russians or Ukrainians have forgotten how to employ combined arms.  it has gone static because the way we used to do combined arms does not work there. So what do they do now?

The math supports that Ukraine has moved from a traditional (and heavily soviet influenced) artillery-based doctrine to a drone based special delivery doctrine.   Largely because they quite literally used up all the 152 west of the Russian border and most the 155 that anybody was willing to spare.  
They're replacing it with plans to build at least 1M drones in 2024.  While few of the drones have more explosive power than an RPG7 or maybe four M77 bomblets, the precision of delivery makes up for it.
Some statistics that showed up some large number of pages ago suggested that it takes fewer than 10 drones (and likely fewer than 5) to produce 1 russian casualty.  And the drone drivers get more and more experienced because they're at much lower risk than if they were in a trench at the front edge, or even crewing a gun that needs a clear path to shoot.  
So at 100K drones/month, and an efficiency somewhere between 2 and 10 drones/casualty, they can inflict somewhere between 10K and 50K casualties/month on Russia. Even at the low end, drones are causing a significant number of russian casualties.  It's also consistent with the transition we've been seeing of fewer Ukrainians in trenches defending against assaults, or clearing russians out of trenches, and more Russians getting blown up in convoys in drone attacks far from the front lines.  If drones were costly to Ukraine, we wouldn't be seeing lots of videos of drones chasing individuals.  Or a drone being used to destroy a single abandoned PKM.  Or sent as self propelled grenades to clear out blind holes in trenches.
They have a *way* smaller logistical tail than a similar amount of combat power in artillery, tanks, or aircraft.  A suitcase full of drones that weighs less than a single 155 shell (not including propellant) has combat power comparable to somewhere between 10 and 50 conventional 155 shells.  And the drones just need to travel with a battery charger and some coax to run their antenna far away, rather than an M777 and a truck load of propellant bags.
Until someone comes up with effective counter drone and gets it into the field, drones are going to continue to make traditional maneuver elements obsolete at a rapid rate.
 

Also been discussed ad nauseam.  Russia actually tried and failed before the invasion to blind NATO by destroying one of their own in order to create enough debris to damage NATO satellites.  It didn't make any sense until a little research showed that Russia sucks at payloads and was launching satellites that dropped film capsules as late as ~2015.  The first electro-optic satellite the US launched was in 1976.  So Russia is 40 years behind, and probably has nothing functional.  China may share information with them, but if they are, it's not very quickly.
But asat weapons?  Sure, if you only need to take down one satellite.  If you only need to take down one you can even send a space truck and steal it.  But if you need to take down 50 or 300 it gets a lot harder. Planet has ~20 that do 50 cm on the ground, and ~100 that do ~3 m on the ground.  And they're just one of many companies.  There are at least a couple companies with SAR constellations.  And the business is just starting to wind up - everybody's been working on making launch cheap without planning what to launch.  We're only just starting to see people developing stuff to put on them.  And then there are governments.  Make a habit of taking out a government's satellites and they'll find way to make sure your launches never reach orbit to do the next one.

I have ignored the guy.  He declares in one post that artillery is still doing most the work.  And when I raise the fact that Ukraine did not have artillery and were using drones as an ad hoc replacement…now  artillery was “but one system”.
We know that the UA was in dire straights last winter. Short on just about everything due to hold ups in US Congress. What they did have was commercial FPV drones with RPG warheads, and undermanned infantry. The RA after Adiivka pushed hard and continuously along many fronts.  The daily losses for the RA were the highest in the war. I suspect they were hoping for a breakthrough. We know the reliance on FPVs was heavy because 1) the UA had little else for ranged fires and 2) the RA was reporting the threat and began a series of adaptations to try and counter.
The result: the RA were held back. The combination of C4ISR and unmanned systems superiority held - but it was a close run thing. The UA had to cede ground.
As to ArmouredTopHat - I call “agenda”.  The pattern of hijacking, wall-eyed assessment, cherry picking and now the time honoured “righteous indignation” along with some bizarre morale fall back position is a clear sign. This poster is not here to learn anything.  He is here to promote some sort of agenda. In a first for this thread, it does not appear to be political, which is nice for once.  I don’t know if he works for industry, or is going to try and build cred and then go political. Or maybe it is simply personal because, reasons.
Either way, I have added him to my ignore list for the simple reason that he has said everything he is going to say. Every post is a variation on repeated themes. Every angle coming from the same direction. This is not good counter-thinking or critical analysis - we have others like holoween for that. This is promotion of a position without evidence and countering every other one…because they are other ones.  It distracts from the main point of this thread which is an open-eyed, open source analysis of the war.
Now, I do not advocate banning - and I know Steve wouldn’t based on this alone - but I for one am turning him off and getting back to the war at hand.
[Edit:  and of course evidence of FPVs last winter because I would not want to “insult” :

https://www.forbes.com/sites/davidhambling/2024/04/24/why-is-russia-losing-the-fpv-drone-war/
https://en.defence-ua.com/analysis/detailed_fpv_drone_usage_statistics_show_russias_starting_to_outpace_ukraine-9361.html (this one is interesting as the RA got in the game as well).
https://euromaidanpress.com/2024/03/22/both-ukraine-and-russia-rapidly-increase-the-use-of-fpv-drones-challenging-conventional-trench-fortifications/
https://www.rusi.org/explore-our-research/publications/commentary/weathering-storm-western-security-assistance-defensive-ukraine

I am sure the big space powers (US, China) can shoot down satellites all day, possibly even faster than they can be replaced. The issue is that debris cloud starts to destroy everyone else's satellites too so they will be really upsetting their allies if they do it. Even China needs to consider what the EU would think if they accidentally destroyed a bunch of European sats in a Pacific war. 
So by the time you are destroying satellites it is effectively WW3. 

And Ukraines drone supply is good enough that they send a second one to dispatch people that are already obviously wounded. Because apparently zinc coffins are the only way to reason with the Russians. By the way what is world price of zinc doing....?
 

The math supports that Ukraine has moved from a traditional (and heavily soviet influenced) artillery-based doctrine to a drone based special delivery doctrine.   Largely because they quite literally used up all the 152 west of the Russian border and most the 155 that anybody was willing to spare.  
They're replacing it with plans to build at least 1M drones in 2024.  While few of the drones have more explosive power than an RPG7 or maybe four M77 bomblets, the precision of delivery makes up for it.
Some statistics that showed up some large number of pages ago suggested that it takes fewer than 10 drones (and likely fewer than 5) to produce 1 russian casualty.  And the drone drivers get more and more experienced because they're at much lower risk than if they were in a trench at the front edge, or even crewing a gun that needs a clear path to shoot.  
So at 100K drones/month, and an efficiency somewhere between 2 and 10 drones/casualty, they can inflict somewhere between 10K and 50K casualties/month on Russia. Even at the low end, drones are causing a significant number of russian casualties.  It's also consistent with the transition we've been seeing of fewer Ukrainians in trenches defending against assaults, or clearing russians out of trenches, and more Russians getting blown up in convoys in drone attacks far from the front lines.  If drones were costly to Ukraine, we wouldn't be seeing lots of videos of drones chasing individuals.  Or a drone being used to destroy a single abandoned PKM.  Or sent as self propelled grenades to clear out blind holes in trenches.
They have a *way* smaller logistical tail than a similar amount of combat power in artillery, tanks, or aircraft.  A suitcase full of drones that weighs less than a single 155 shell (not including propellant) has combat power comparable to somewhere between 10 and 50 conventional 155 shells.  And the drones just need to travel with a battery charger and some coax to run their antenna far away, rather than an M777 and a truck load of propellant bags.
Until someone comes up with effective counter drone and gets it into the field, drones are going to continue to make traditional maneuver elements obsolete at a rapid rate.
 

The math supports that Ukraine has moved from a traditional (and heavily soviet influenced) artillery-based doctrine to a drone based special delivery doctrine.   Largely because they quite literally used up all the 152 west of the Russian border and most the 155 that anybody was willing to spare.  
They're replacing it with plans to build at least 1M drones in 2024.  While few of the drones have more explosive power than an RPG7 or maybe four M77 bomblets, the precision of delivery makes up for it.
Some statistics that showed up some large number of pages ago suggested that it takes fewer than 10 drones (and likely fewer than 5) to produce 1 russian casualty.  And the drone drivers get more and more experienced because they're at much lower risk than if they were in a trench at the front edge, or even crewing a gun that needs a clear path to shoot.  
So at 100K drones/month, and an efficiency somewhere between 2 and 10 drones/casualty, they can inflict somewhere between 10K and 50K casualties/month on Russia. Even at the low end, drones are causing a significant number of russian casualties.  It's also consistent with the transition we've been seeing of fewer Ukrainians in trenches defending against assaults, or clearing russians out of trenches, and more Russians getting blown up in convoys in drone attacks far from the front lines.  If drones were costly to Ukraine, we wouldn't be seeing lots of videos of drones chasing individuals.  Or a drone being used to destroy a single abandoned PKM.  Or sent as self propelled grenades to clear out blind holes in trenches.
They have a *way* smaller logistical tail than a similar amount of combat power in artillery, tanks, or aircraft.  A suitcase full of drones that weighs less than a single 155 shell (not including propellant) has combat power comparable to somewhere between 10 and 50 conventional 155 shells.  And the drones just need to travel with a battery charger and some coax to run their antenna far away, rather than an M777 and a truck load of propellant bags.
Until someone comes up with effective counter drone and gets it into the field, drones are going to continue to make traditional maneuver elements obsolete at a rapid rate.
 

The math supports that Ukraine has moved from a traditional (and heavily soviet influenced) artillery-based doctrine to a drone based special delivery doctrine.   Largely because they quite literally used up all the 152 west of the Russian border and most the 155 that anybody was willing to spare.  
They're replacing it with plans to build at least 1M drones in 2024.  While few of the drones have more explosive power than an RPG7 or maybe four M77 bomblets, the precision of delivery makes up for it.
Some statistics that showed up some large number of pages ago suggested that it takes fewer than 10 drones (and likely fewer than 5) to produce 1 russian casualty.  And the drone drivers get more and more experienced because they're at much lower risk than if they were in a trench at the front edge, or even crewing a gun that needs a clear path to shoot.  
So at 100K drones/month, and an efficiency somewhere between 2 and 10 drones/casualty, they can inflict somewhere between 10K and 50K casualties/month on Russia. Even at the low end, drones are causing a significant number of russian casualties.  It's also consistent with the transition we've been seeing of fewer Ukrainians in trenches defending against assaults, or clearing russians out of trenches, and more Russians getting blown up in convoys in drone attacks far from the front lines.  If drones were costly to Ukraine, we wouldn't be seeing lots of videos of drones chasing individuals.  Or a drone being used to destroy a single abandoned PKM.  Or sent as self propelled grenades to clear out blind holes in trenches.
They have a *way* smaller logistical tail than a similar amount of combat power in artillery, tanks, or aircraft.  A suitcase full of drones that weighs less than a single 155 shell (not including propellant) has combat power comparable to somewhere between 10 and 50 conventional 155 shells.  And the drones just need to travel with a battery charger and some coax to run their antenna far away, rather than an M777 and a truck load of propellant bags.
Until someone comes up with effective counter drone and gets it into the field, drones are going to continue to make traditional maneuver elements obsolete at a rapid rate.
 

The math supports that Ukraine has moved from a traditional (and heavily soviet influenced) artillery-based doctrine to a drone based special delivery doctrine.   Largely because they quite literally used up all the 152 west of the Russian border and most the 155 that anybody was willing to spare.  
They're replacing it with plans to build at least 1M drones in 2024.  While few of the drones have more explosive power than an RPG7 or maybe four M77 bomblets, the precision of delivery makes up for it.
Some statistics that showed up some large number of pages ago suggested that it takes fewer than 10 drones (and likely fewer than 5) to produce 1 russian casualty.  And the drone drivers get more and more experienced because they're at much lower risk than if they were in a trench at the front edge, or even crewing a gun that needs a clear path to shoot.  
So at 100K drones/month, and an efficiency somewhere between 2 and 10 drones/casualty, they can inflict somewhere between 10K and 50K casualties/month on Russia. Even at the low end, drones are causing a significant number of russian casualties.  It's also consistent with the transition we've been seeing of fewer Ukrainians in trenches defending against assaults, or clearing russians out of trenches, and more Russians getting blown up in convoys in drone attacks far from the front lines.  If drones were costly to Ukraine, we wouldn't be seeing lots of videos of drones chasing individuals.  Or a drone being used to destroy a single abandoned PKM.  Or sent as self propelled grenades to clear out blind holes in trenches.
They have a *way* smaller logistical tail than a similar amount of combat power in artillery, tanks, or aircraft.  A suitcase full of drones that weighs less than a single 155 shell (not including propellant) has combat power comparable to somewhere between 10 and 50 conventional 155 shells.  And the drones just need to travel with a battery charger and some coax to run their antenna far away, rather than an M777 and a truck load of propellant bags.
Until someone comes up with effective counter drone and gets it into the field, drones are going to continue to make traditional maneuver elements obsolete at a rapid rate.
 

It’s what keeps us grounded.

The UA created an Unmanned Forces Branch, this is basically an unmanned service - https://www.janes.com/osint-insights/defence-news/air/ukraine-conflict-ukraine-establishes-worlds-first-unmanned-force
What we are very likely seeing are the early days of them professionalizing these capabilities. So this force is just the beginning. Clearly these guys are elite based on their EW avoidance. 

Now enough of this “wait a minute” BS. The questions are 1) can the UA upscale this sort of performance and 2) at scale can it break the deadlock?  In this war.  We can leave the future to magical wizard 30mm guns and fairy dust, but in this war right here, will we see the UA break the deadlock if they can do this at higher scales?  Can they kick corrosive warfare back into gear? Because if they can, we might still have a ballgame.

C’mon….seriously? You have held up a couple dozen Trophy uses by the IDF as “well proven operational effectiveness” and a few corporate videos as “proof of technological counters”.  We have seen stats of hundreds of vehicle strikes, and more than a few reports like this one. The UA really had nothing but FPVs to hold off the RA last winter, and they did. The RA has done more chicken cage building and weird add ons than we can count. And vehicles - including tanks - are being held back because to advance is nearly suicidal…and we have video of that as well.  Then we get a pretty vivid report of what is effectively human hunting at 15kms -just stop and think about the repercussions of that.  And you are still “well wait a minute”?  Hell, why not cry “fake news”?
Seriously, how much proof do you need? This is beyond critical thinking, this is sticking a head in the sand and living in denial. 
This is not a “blip” it is a sea change as far as warfare is concerned.

Agreed. The operators may already be doing what I would suggest: broadcast from antennas that are on nice long cords so anyone successfully following the signals will just blow up some gear. 
Heck they could even have multiple antennas and if one gets hit pull out the next one. Or better yet when you bug out just leave it and setup in another location with antenna #2. Later send someone back for antenna #1, if its still there great use it for the next new location. You can even have it pre setup in the new location. The operators are only down for their travel time.

Welcome the future.
Able to chase individual soldiers into a drain pipe 15kms away.  At this rate I am not sure if Infantry are still viable, let alone freakin tanks.

I don't care if it can "reason".  I'd be happy if it can show me on an N-dimensional plot how it parametrized and clustered the input data so I can do the equivalent of slapping it on the nose with a rolled up newspaper and say "No! Don't do that to that cluster!"

Sources is an overloaded word.  I didn't mean references, I mean the actual data it used from the inputs in the reasoning and the pattern it actually followed (not a pattern that it produces because there are a bunch of examples in the training data that give both the logic and the explanation. It's popping out a pattern that "explains the reasoning", but dollars to donuts it didn't actually follow that reasoning in solving the problem - it's just the standard pattern humans use for that problem, so it spat it back.
That's because their "reasoning" isn't reasoning, it's spitting back an interpolation of a best fit based on the input set given some overlap of the novel inputs with the training set.  It dies at the edges for the same reason string theory can't predict anything and I can fit an elephant with a function that has 10^6 free parameters but that function can't describe anything outside the elephant.  High order multiparameter fits get really ugly outside the range of the data they're fit to.
ChatGPT answered that correctly because that problem is solved countless times in its training set with very similar wording.  That's exactly an example of a physics 101/102/201/202 problem that appears repeatedly in problem sets and with complete solutions all over the internet.
Here's a simple one that doesn't appear all over the internet.  I asked it three different ways and it got it wrong all three times.  The correct answer is arccos(10/17).  IIRC, if the spheres slide then it's arccos(2/3).  The radii of the spheres don't matter.
there's no reasoning in there.  It's telling a story like a Phys 101 student but doesn't have even a simple physical model. It's following a standard pattern and filling in the blanks, but it's nice sentences so it looks like reasoning.  Fancy Mad Libs.
I've stabbed a few neurons in dishes with pointy sticks, and even tried to train them to do absurdly simple things, and I've never been entirely convinced that the CS model of neural computing is as close to actual neural computing as people seem to act that it is.  People have something like 10^12 neurons and ~10^14 interconnects (synapses).  That's an enormous number of free parameters, and is used to fit a far, far smaller input data set than any of the chatbots have, but the chatbots still kind of suck, but sound good doing it. Given the relative number of free parameters in humans, you'd expect them to go off the rails far more often than they alraedy do (but it maybe explains conspiracy theorists...)
Humans develop models that are much more complex than simply heuristics plus learning some algebraic formulae.  And with far less training data than the Chatbots.  I've read way more of the internet than I care to think about, a ton of books, and done a ton of problems. But I've ingested only a tiny fraction of the input data that your average Chatbot has, but I get annoyed by the second paragraph when they start to BS me like a 1st year undergrad who doesn't know the answer to something really that they probably read the day before.  There's no model of the universe in there.  So you can expand the context arbitrarily, and all you're going to get is more complicated wrong answers if there's not a physical model.
 
 
 
 

"Explainability" is a thing in AI.  Or more appropriately, Machine Learning.
When I use ML to act as a "remote graduate student" it evaluates a bunch of properties of what it's looking at that are based on real physical models, and then compares what it sees for any data set to other real human-evaluated data sets.  It then reports that "this new thing is x% like set 1, y% like set 2, etc" to make clear why it's making some recommendation.  What most of the AI chat bots (and apparently go bots) isn't this.
What a lot of the AIs that take enormous training data sets are doing is just picking the most likely move/phrase/sentence/whatever based based on a distillation of some enormous number of games.  Go is a combinatoric nightmare, so the computer has to constrain how far it looks around any point, so it makes sense that a big non-local surround would fake it out.  So it has no model, or an incomplete model.  All the ChatAIs are basically just "fits" to all the garbage typed into the internet, so they come up with some spectacularly bad "hallucinations".  It's because there's no actual model of reality behind them.