How Hot is Ukraine Gonna Get?

[dan/california]

10 hours ago, TheVulture said:

The worry is that once western defence companies get involved, and NATO procurement tendencies, you're going to see the $1000 drone made in a shed in Ukraine with an RPG warhead strapped on be replaced with a custom-built state of the art $20,000 drone that does the same job maybe 5% better.

The defense contractors will over charge, that is what they do for a living. What we don't know is how big the improvements will be. I can see it being the above mentioned five percent, I can see them them literally being WORSE, or I can see them being so lethal we move right along into science fiction scenarios.

My greatest worry is that the NATO defense contractor version will scrupulously observe limits on AI/autonomy, and the other sides won't. So the otherwise vastly inferior version being cranked out of Chinese toy factories and garages all over the the global south will be vastly more effective, even with grossly inferior hardware.

[OBJ]

23 minutes ago, cyrano01 said:

Always tricky to pick out the right lessons though, and even harder to do so if they contradict your existing doctrine. The then Chief of the Air Staff(CAS), Sir Cyril Newall, described the Luftwaffe's support of ground operations in Spain as a gross misuse of air-power!

To be fair to Newall he was subsequently CAS during the Battle of Britain so perhaps deserves some of Dowding and Park's reflected glory.

 

Agree, and the only serious reading I did on it was:
https://www.amazon.com/Spanish-Civil-War-Hugh-Thomas/dp/B002ADRJB8
It is a heavy read.

The one thing that stuck with me was the Luftwaffe's affirmation/refinement of their fighter tactics, finger four flight, two plane elements, lead and wingman, loose formation, shooter and defender, which hurt the Allies when they were still flying in close formation three plane 'Vics' with the wingmen out of necessity focused on not running into the lead.

Of course I also read 'For whom the Bell Tolls' but did not take away a lot of tactical insights from that one

Edited January 31 by OBJ

[slysniper]

well, I am waiting to see, mini drones that have programming that will be able to track other drones and be basically like an heat seekling missile.

So basically suicide drones to take out other drones in the sky. smaller, faster and capable of disabiling other drones.

So where does it all go to, really dont know. but I agree with the concept that we are at a point where its the early stages of developement. 

where does this new warefare end at.

who knows?

Edited January 31 by slysniper

[OBJ]

1 hour ago, hcrof said:

It seems that at least some Russians have come to similar conclusions about war as this forum:

https://www.armystandard.ru/news/2024129114-TnO1s.html

No comment really, except that we should not stereotype the Russian general staff as a bunch of drunks and incompetents stuck in the soviet past. They are learning, even if implementation is the hard part, not theory. 

Really good read, thanks for sharing.

Like you said, Russian GS sees everything we've seen. Let's hope our side translates thinking to doing faster than they can. Unfortunately Russia has the pain and benefit of first hand experience, whereas on our side only Ukraine does.

What do you think we here in this forum should be looking for as indications of successful Russian LL implementation?

[Vet 0369]

On 1/30/2024 at 10:38 AM, Bulletpoint said:

Were there any female marines in Vietnam? Since you're using gender neutral "their"..

Of course there were. At that time, they were known as Women Marines or WM. The Marine corps has since dropped the “Women” qualifier and all males and females are simply “Marines.” Women recruits train alongside the male recruits at Paris Island and San Diego, including weapons and combat tactics. However, I don’t believe that women are integrated in combat units yet. The statement still applies that “every Marine is first and foremost a rifleman.” Including Females.

[JonS]

1 hour ago, dan/california said:

My greatest worry is that the NATO defense contractor version will scrupulously observe limits on AI/autonomy, and the other sides won't.

This is bull****. Either you support the western rules based order, or you do not. You cannot start to pick and choose which rules to observe when you think they've become a bit inconvenient.

[dan/california]

19 minutes ago, JonS said:

This is bull****. Either you support the western rules based order, or you do not. You cannot start to pick and choose which rules to observe when you think they've become a bit inconvenient.

What I am trying to say is maybe we shouldn't write a new set of rules that guarantee we lose the next war. Just as a first pass proposal. 

[Vet 0369]

19 hours ago, Kinophile said:

To circle back to the Pacific - the early assault failures lead to rapid development and implementation of solutions (both technical & organizational) in time for the next island. Each assault iterated on the previous until by the end the US could contemplate invading Japan proper, a truly insane proposition. But they thought they could do it and I personally feel they would have eventually, bloodily, "won". 

Not really sure what “early assault failures” you are referring to. The only amphibious assault “failure” that I remember reading about was the initial Japanese amphibious assault on Wake Island shortly after their attack on Pearl Harbor. The Japanese launched an amphibious assault that was defeated by the U.S. Marine Detachment, who not only defeated the only amphibious assault to fail (in the Pacific), but also sank a Japanese destroyer while they vwere doing it.

Yes, the Marine and Army amphibious assaults were very costly for the Marines and Soldiers who made them (especially during the Navy’s island hopping campaign), but NONE of them “failed!”

[Tux]

2 hours ago, billbindc said:

Was just thinking along the same lines when I read this. And that we are in the triplane era, still. 

Earlier still, imo. We haven’t seen an Eindecker, yet!

As much as I have also long been thinking of today’s drone war as analogous to the first year or two of air warfare (certainly in terms of the likely rate at which development will occur), we should note that there are several key differences:

1. Today’s first generation (or maybe gen 1.5) drones are actually apparently very effective ground attackers. It took until the development of PGMs for aircraft to become anywhere near as efficient.  In this sense it makes much more sense to think of drones as munitions than aircraft.

2. The vast majority of development focus seems to be on increasing drones’ offensive potential (again, “munitions”).  I’ve yet to see evidence of anyone trying to field a single type of defensive drone ‘fighter’ (as I’ve mentioned several times I think there’s a good chance we will see a modern-day “Fokker Scourge” when one does appear).

3. The entry barrier to effective drone use is spectacularly low, to the point that it seems to have been privately-bought and operated drones that did all the early running in this war.  This is very much unlike air warfare and may mean that national armed forces will struggle to maintain a significant qualitative edge over commercially-available drones for the foreseeable future.

[Tux]

2 hours ago, dan/california said:

The defense contractors will over charge, that is what they do for a living. What we don't know is how big the improvements will be. I can see it being the above mentioned five percent, I can see them them literally being WORSE, or I can see them being so lethal we move right along into science fiction scenarios.

My greatest worry is that the NATO defense contractor version will scrupulously observe limits on AI/autonomy, and the other sides won't. So the otherwise vastly inferior version being cranked out of Chinese toy factories and garages all over the the global south will be vastly more effective, even with grossly inferior hardware.

What sort of rules are you thinking of?  Surely a drone that doesn’t differentiate between a tank and a family hatchback, for example, would be objectively a much less effective weapon?

[OBJ]

2 minutes ago, Tux said:

What sort of rules are you thinking of?  Surely a drone that doesn’t differentiate between a tank and a family hatchback, for example, would be objectively a much less effective weapon?

Can't speak for Dan, but at least one issue is human in the loop, which defeats the AI advantage of decision cycle time.

[OBJ]

Where is LLF when you need him? Come on man, give us a meme on the Robo Cop ED 209!

[OBJ]

I understand no human in the loop is already built into the Israeli Iron Dome and US Patriot AD systems. Navy Phoenix system too. So at least from a defensive perspective we're getting there.

Edited January 31 by OBJ

[Maciej Zwolinski]

1 hour ago, JonS said:

This is bull****. Either you support the western rules based order, or you do not. You cannot start to pick and choose which rules to observe when you think they've become a bit inconvenient.

That is an extremely unhelpful take on a real problem. Adherents of any kind of order first have to exist, and second have to enjoy decision autonomy for that order to be applied at all. In other words, they cannot be destroyed or subjugated by others. In other words, they have to be capable to win existential wars. That is obvious.

And the history teaches us that adherents of rules which restricted their ability to effectively resist an enemy who fought under a different set of rules usually ditch their rules and look for others. Aztecs after the first battles with the Spaniards stopped trying to take them prisoner to be sacrificed. Plains Indians did not try to count coups on US soldiers. To unrestricted submarine warfare and bombing of cities by Germans the British and Americans reacted with enthusiastic adoption of the same measures and being much better at it. That's just human nature.

So this is not really a problem of supporting a rules based order, or not supporting it. No piece of legislation would be a serious element of a decision process in case of existential war, I am absolutely certain of it.  The problem is 1) whether the advantages of using autonomous system by the "bad guys" can be overcome without resorting to AI on the "good guys" side, and at what cost; and 2)  does adoption of AI systems also on the good guys side create more risks than the disadvantage vs. the enemy. These questions are really worth pondering, although I am not sure if we can add anything on the subject right now.

PS. The second sentence is probably the most untrue statement one can make in the entirety of the social sciences. People pick and choose what rules to observe literally all the time.  I have been working as a lawyer for 20+ years and have seen people do that (or at least trying to do that) during each of the 8+ hours of every business day. The entire judicial system is created because people do that.

 

[JonS]

5 minutes ago, Maciej Zwolinski said:

PS. The second sentence is probably the most untrue statement one can make in the entirety of the social sciences. People pick and choose what rules to observe literally all the time.  I have been working as a lawyer for 20+ years and have seen people do that (or at least trying to do that) during each of the 8+ hours of every business day. The entire judicial system is created because people do that.

Of course they do! But in response society doesn't make murder or speeding legal simply because both continue to occur.

[Maciej Zwolinski]

11 minutes ago, Tux said:

What sort of rules are you thinking of?  Surely a drone that doesn’t differentiate between a tank and a family hatchback, for example, would be objectively a much less effective weapon?

It is not that. A more pertinent example: currently one of the most effective drones is called FPV - first person view, like a videogame. It is steered by a man behind a frontline via some kind of communications link. However, that link is a source of liability because if it can be attacked by ECM, and if the link is broken, the drone will crash.

On the other hand, already now there are automated visual recognition and tracking devices available commercially which could be installed in a drone, which will steer it to the target without the need for the vulnerable link. AFAIK at the current stage of development of electronic warfare, that link is considered to be the only avenue of attack for ECM and without it, an autonomous drone would become completely impervious to ECM. 

The problem? No man in the loop, as OBJ posted.

[dan/california]

19 minutes ago, Tux said:

What sort of rules are you thinking of?  Surely a drone that doesn’t differentiate between a tank and a family hatchback, for example, would be objectively a much less effective weapon?

 

15 minutes ago, OBJ said:

Can't speak for Dan, but at least one issue is human in the loop, which defeats the AI advantage of decision cycle time.

If you look at this video around the four minute mark, the tech to do rather good target ID discrimination at the approximate range and viewpoint of a drone already exists. And it has for a while. The Pentagon dodged the whole autonomy issue on this by saying the man in loop level was handled at the decision to drop the whole bomb. It is now completely practical give every submunition its own custom quadcopter ride. What I am trying say is that the military effectiveness of giving said quadcopters a kill box and a set of targeting priorities is going to be at least an order order of magnitude more effective than making each of them phone home for permission in a high EW environment. I am just saying a we should apply a bit of realism now, and not after we have gotten a couple of heavy Brigades cut into little tiny pieces. At which point we will do it in a panic, badly.

[Maciej Zwolinski]

6 minutes ago, JonS said:

Of course they do! But in response society doesn't make murder or speeding legal simply because both continue to occur.

Certainly. Even thieves want to have their "property" protected against other thieves.  But that is the point of view of the community which creates the rules in the abstract. Once it comes to submitting to those rules in practice, that is a different story, and practically nobody adopts and "all or nothing" approach. Who never broke the local speed limit when being late for work? Or jaywalked in the middle of the night without a car in sight?

It is even more tempting in international law, because it is vague and the penalties for breach of it are mostly puny. The famous ICC temporary ruling in the RSA vs Israel case told Israel not to commit genocide. But Israel considers its actions not a genocide, and goes on with the business as usual. In international law, the ability to spin a good yarn, being strong and wealthy, and having strong and wealthy friends will get a country of the hook always or almost always. And that is how it must remain, unless a global government is created.

But that digresses to far from the topic.

 

[billbindc]

14 minutes ago, Maciej Zwolinski said:

It is not that. A more pertinent example: currently one of the most effective drones is called FPV - first person view, like a videogame. It is steered by a man behind a frontline via some kind of communications link. However, that link is a source of liability because if it can be attacked by ECM, and if the link is broken, the drone will crash.

On the other hand, already now there are automated visual recognition and tracking devices available commercially which could be installed in a drone, which will steer it to the target without the need for the vulnerable link. AFAIK at the current stage of development of electronic warfare, that link is considered to be the only avenue of attack for ECM and without it, an autonomous drone would become completely impervious to ECM. 

The problem? No man in the loop, as OBJ posted.

I've always imagined the next/next step will be an AI targeting system as a backup/dead man switch when ECM cuts the link. 

[Tux]

12 minutes ago, Maciej Zwolinski said:

On the other hand, already now there are automated visual recognition and tracking devices available commercially which could be installed in a drone, which will steer it to the target without the need for the vulnerable link.…. 

The problem? No man in the loop, as OBJ posted.

Noted. However, do the devices in your example reliably guide drones to the correct targets?  If so why would the West supposedly not permit that?  If they don’t then is there any significant benefit gained from their use, vs the added cost each time the drone selects (for example) a civilian target in error?

[Maciej Zwolinski]

Just now, billbindc said:

I've always imagined the next/next step will be an AI targeting system as a backup/dead man switch when ECM cuts the link. 

Which leads the same problem, just in a lower number of cases.

I think the real conundrum is drone swarms, if they develop in the direction of really swarmy swarms. Like a hundred+ units. Nobody will have a 100 operators per single swarm, and they will have to be at least semi-autonomous (wingmen oriented on one human operated drone) from the get-go.

[OBJ]

2 minutes ago, billbindc said:

I've always imagined the next/next step will be an AI targeting system as a backup/dead man switch when ECM cuts the link. 

Still with the disparity in autonomous AI decision cycle time vs human in the loop decision cycle time. Where they are competing autonomous AI is going to win.

If the discussion moves on to both major powers actively engaged in wearing down the others RSC, the side with autonomous systems is going to win. 

[The_Capt]

4 minutes ago, billbindc said:

I've always imagined the next/next step will be an AI targeting system as a backup/dead man switch when ECM cuts the link. 

I suspect we are looking at hybrid system.  Human control up to a release point and then a pre-authorized kill box will full autonomy once in a denied space.  Target discrimination will be a major argument - hell we have problems with humans doing it right now.  It may have to stay at the vehicle/anti-material level - we can program in restricted targeting to military vehicles but individual people is really hard to do.  Is the person carrying a gun?  Are they acting threateningly.  Are they being shady?  Under ROEs a human can engage on all of these, but I doubt we will trust a machine with all this for some time.

[OBJ]

3 minutes ago, The_Capt said:

Target discrimination will be a major argument - hell we have problems with humans doing it right now.

Given what we know about machine learning...the machines are apt to do it better than humans. The machine won't be scared, tired, angry, sorrowful, melancholy, vengeful, hungry, or feeling cold or wet either.

[billbindc]

3 minutes ago, The_Capt said:

I suspect we are looking at hybrid system.  Human control up to a release point and then a pre-authorized kill box will full autonomy once in a denied space.  Target discrimination will be a major argument - hell we have problems with humans doing it right now.  It may have to stay at the vehicle/anti-material level - we can program in restricted targeting to military vehicles but individual people is really hard to do.  Is the person carrying a gun?  Are they acting threateningly.  Are they being shady?  Under ROEs a human can engage on all of these, but I doubt we will trust a machine with all this for some time.

It is without question that autonomous capability...even if unused in low intensity situations...will be built in. The strategic concern will always be there (i.e. what happens if we fight a mass conflict and don't have it while the other guys does) and once the software is written it is a fairly low cost decision to add it on. And it is quite likely that the ability to identify and target will only become more refined over time.