Tux

Isn't all of this a fruitless discussion? We've had stunning Ukrainian successes in 2022. To the point that the discussion here had devolved into a meme show (go back, I know I complained about this...). Ukrainian soldiers were larger than life superheroes who made fun of Russia that sucked at war badly. Anyone who wasn't on board of the hype train was suspected of being a Russian troll. We had bets running how many months it would take for the Russian army to collapse, remember? Oh, and of course Leopards! Western tanks in Ukraine, some were already dreaming of a drive on Moscow.
From that perspective 2023 was a total letdown. So now one camp goes "See, Russia can't be beaten, they are too big, sanctions are failing, they will always come out top because WW2. Crap because its not 1942 and, sorry to all history nerds out there, history simply doesn't repeat itself. (Doesn't mean we can't learn from history though). The other camp goes "See, well that year wasn't much of a success but still, Putin started out with the goal to seize all of Ukraine or destroy it. They didn't, and they no in longer have the means to, so no matter what, they can't win. Also once the next Western Wunderwaffe comes to Ukraine, Ukraine will win because Russians are incapable of learning. And Ukraine has to win because they are the good guys and it is always like that in Hollywood."
We are arguing as if this was a game of Combat Mission, where there is some neutral instance that is going to decide who wins and who loses based on some predefined and unchangeable parameters. In reality, the parameters change over time, the points awarded change too and who wins our loses is something the different parties involved will decide for themselves.
 

This fellow summarized the German-Ukrainian agreement:
 
 
 
 

If one should have free time during tomorrow and/or next Thursday, National Defence University of Finland is hosting a Russia seminar: https://maanpuolustuskorkeakoulu.fi/en/russia-seminar with quite interesting themes, topics and speakers: https://maanpuolustuskorkeakoulu.fi/en/russia-seminar/programme   
Both days are streamed (if not hacked / interrupted by vatniks 🙂😞https://www.youtube.com/@Maanpuolustuskorkeakoulu/streams

Sorry for the length but this is well worth reading: 
 
Less Than Meets the Eye - Parsing Tucker's Putin Interview
JOHN GANZ
I was probably one of the relatively few people that sat through the entire two hour Tucker Carlson interview with Vladimir Putin. To call it an “interview” is not quite right: Carlson essentially allowed Putin to discourse at length and only occasionally tried to prod him in the direction of his own preferred talking points about the war in Ukraine. Any appearance of tension or journalistic effort only occurred because Carlson seemed to have the expectation that Putin would cooperate with his own line and appeared frustrated (“annoyed,” he said in his prefatory remarks) when it immediately turned out Putin seemed to have his own ideas . Essentially, the interview consisted of a melange of multiple, sometimes contradictory, lines of propaganda about the war. But to say that it was “propaganda” also might gave a misleading impression: it suggests that there is a “real” underlying motivation for the war, while the justifications are merely self-serving deceptions for public consumption. But what it actually might reveal is superficiality and incoherence of the case for war itself. Instead, there were a number overlapping and shifting messages to different constituencies. is not a single overarching ideology at play, but rather a succession of “ideologemes,” little snippets of ideology: themes from Russian nationalism, Western far right cultural pessimism, anti-colonialism, and Soviet nostalgia all crop up—even little remnants of Putin’s Marxist-Leninist training appear, like when he talked about the “excessive production capacities” of the West. Putin doubled down on the theme of “denazification”—evidently somewhat to the irritation of the America Firster Carlson —while at the same time offering a revisionist picture of the start of World War II, sympathetic to Hitler’s territorial aims and essentially blaming the war on Polish intransigence, saying “they pushed Hitler to start World War II by attacking them.” This speaks to the awkward position of Russia claiming simultaneously claiming to embody the continuation of the Great Patriotic War’s anti-fascist crusade while being the darling of a far right at home and abroad, which views it as the last remaining hope of “white civilization.” 

This synthetic, “postmodern” quality does not reflect devilishly clever strategy, rather its incoherence directly reflects the fragility and fragmentation of Russia’s entire post-Soviet political project. The Ukrainian sociologist Volodymyr Ischenko writes of “a crisis of hegemony” in the post-Soviet world and that both Putin’s authoritarian, “Bonapartist” rule and its consequent war arise from the same “incapacity of the ruling class to develop sustained political, moral, and intellectual leadership.” His regime is ad hoc: a cobbled together arrangement of veterans of the security services and the rent-seeking oligarchs who accepted Putin’s settlement. Prighozhin’s mutiny made this provisional and brittle nature of “the state” clear. Rather than reflecting position of strength the strongman antics of Putin reveal fundamental political weaknesses and failures. As Ischenko put it in an interview with The New Left Review:

"Putin, like other post-Soviet Caesarist leaders, has ruled through a combination of repression, balance and passive consent legitimated by a narrative of restoring stability after the post-Soviet collapse in the 1990s. But he has not offered any attractive developmental project. Russia’s invasion should be analyzed precisely in this context: lacking sufficient soft power of attraction, the Russian ruling clique has ultimately decided to rely on the hard power of violence, starting from coercive diplomacy in the beginning of 2021, then abandoning diplomacy for military coercion in 2022."
The political fragility and insecurity of the ruling class, its cliquishness and insularity, its inability to shape a single coherent narrative of national development, its preoccupation with finding tactical expedients to avoid the chaos of the 1990s and the humiliations of the collapse are all wedded to the cult of “special services,” from the former KGB officer Putin on down. As early as the 2000s, Dimitri Furman noticed this aspect of the regime, writing in his Imitation Democracy: The Development of Russia’s Post-Soviet Political System, that a growing number of “activities, essential to the maintenance of the system, were in essence ‘secret special operations.’ Rather than rare exceptions, they were fast becoming crucial and lasting dimensions of all political activity.” With that in mind, it’s worth noting Putin’s insistence on calling the war in Ukraine, not a war at all, but a “special military operation” and its simultaneous development of contradictory propaganda campaigns directed at different audiences rather than a single, articulable vision of Russia’s role in the world. Putin can’t escape looking at everything as an “op.” (Not for nothing, this confusion of war, propaganda, and secret police subterfuge along with the subordination of politics to the needs and views of the national security apparatus is something usually associated with totalitarian states.)

In so far as anything approaching a worldview emerges from the interview, it is Putin’s preoccupation with the central role “special services” purportedly play in world affairs, particularly his apparent belief that the United States is not governed by its political leadership but by its national security bureaucracy, which accords with Carlson’s view of a “deep state.” This is less of ideology than Putin’s own déformation professionnelle, one that’s so deeply rooted that he felt the need to bring up Carlson’s onetime attempt to join the CIA. (He even seemed to coyly suggest that Tucker might actually work for the CIA, which I’m sure Carlson found flattering.) 

From the very beginning, Carlson’s generously offered Putin the chance to present the war in defensive terms, asking,

"On February 22nd, 2022, you addressed your country in a nationwide address when the conflict in Ukraine started, and you said that you were acting because you had come to the conclusion that the United States, through NATO, might initiate a “surprise attack on our country”. And to American ears, that sounds paranoid. Tell us why you believe the United States might strike Russia out of the blue. How did you conclude that?"

Instead of taking that route, Putin immediately launched into a nearly half hour disquisition on Russian history, the point of which was to stress the original unity of the Ukrainian and Russian peoples. Carlson averred in his opening remarks that he was “shocked” by this, but Putin has been harping on this theme since before the war. In July 2021, he published his essay “On the Historical Unity of Russians and Ukrainians,” which states “true sovereignty of Ukraine is possible only in partnership with Russia.” Of course, “sovereignty in partnership” is not really sovereignty at all. Despite Putin’s open and lengthy statement of what the Old Bolsheviks would’ve called “Great Russian chauvinism,” Carlson came away from the interview stating, “Russia is not an expansionist power. You’d have to be an idiot to think that.” From both Putin’s rhetoric and his behavior, you’d have to be an idiot to think otherwise. Carlson is just employing the propagandist’s trick of employing abuse and invective when the facts clearly oppose their case. But, as Michael Tracey’s recent Substack post makes clear, Putin’s open statements of Russian grand imperial ambitions are troubling for Westerners otherwise predisposed to be sympathetic and who have spent a great deal of time rationalizing Russia’s actions or presenting them in a defensive light. 

In the minds of the Russian ruling class, there’s really no contradiction between defensive and offensive conceptions of the war: they both involve securing of their system, and in moments of more grandiose transport, their civilization, against Western encroachment. The other overriding theme of Putin’s discourse, connected to the fixation on “special services,” is the characterization of the Maidan as a “coup d’etat.” The fear is that the example of success of Ukraine’s political revolution might spread to Russia itself. This concern on the part of the Russian elite is not new: it has its origins in the collective trauma of the Soviet collapse. More proximately, it dates back to the “Color Revolutions” of the 2000s that toppled Leonid Kuchma in Ukraine, Askar Akayev in Kyrgyzstan, and Eduard Shevardnadze in Georgia. As Furman writes, 

"These men had headed systems highly comparable to Russia’s, if substantially weaker, and their ousters aroused an irrational panic of the kind seen in tsarist circles after the French revolutions, or in Soviet circles in the run-up to the Prague Spring. To acknowledge the naturalness, the predictability of these regimes’ collapsing would mean acknowledging the inevitability of the collapse of Russia’s regime, too – an impossibility. Those in power in Russia thus concluded instead that these revolutions were all the work of Western security services (very much as Soviet leaders had blamed similar forces for unrest in Hungary, Czechoslovakia, and Poland)."

Since that time, Russia’s foreign policy in its “near abroad” has since been fundamentally counter-revolutionary. As Ischenko notes the tempo of revolt had been picking up in the run up to the invasion:

"Such uprisings have been accelerating on Russia’s periphery in recent years, including not just the Euromaidan revolution in Ukraine in 2014 but also the revolutions in Armenia, the third revolution in Kyrgyzstan, the failed 2020 uprising in Belarus, and, most recently, the uprising in Kazakhstan. In the two last cases, Russian support proved crucial to ensure the local regime’s survival. Within Russia itself, the “For Fair Elections” rallies held in 2011 and 2012, as well as later mobilizations inspired by Alexei Navalny, were not insignificant. On the eve of the invasion, labor unrest was on the rise, while polls showed declining trust in Putin and a growing number of people who wanted him to retire. Dangerously, opposition to Putin was higher the younger the respondents were."
Again, the war is a piece of domestic policy as much as it is foreign policy: an attempt to consolidate a regime that feels itself to be vulnerable. The acquiescence of the population and the resilience of the Russian economy in the face of sanctions may prove that it was a successful expedient, at least temporarily. It would be dangerous indeed if Russia’s regime concluded that such “operations” redounded mostly to its benefit. 
 
 
 

Regarding Trump, hi from Latin America (Argentina) land of great football and sometimes populist egomaniacs...
I´ll just paste something I wrote in January of 2017 on another forum:
"Coming from a country with it´s fair share of populists in it´s history, let me tell you a couple of things about what will happen:

First: it´s not about policies, it´s not about facts, now it´s only about power and how to keep it

Second: the only thing that will be rewarded is loyalty, it may or may not align with skills, it doesn't matter any more

Third: it´s not about liberal and conservative, It´s not about democrat or republican, now it´s about trump or anti trump, there will be no middle ground to hide, both coalitions will have members of previous coalitions, those won´t matter much anymore, more republicans will be on Trump´s coalition because he was elected as one of course

Fourth: his only objective is reelection, and after that a third mandate, and a fourth and so on, to this end he will push any policy he deems useful, left wing or right wing, won´t matter, he sometimes will be align with former liberal/democrat ideas, he will take ownership of anything that promotes him, claiming it was his idea all along

Fifth: any source of check and balances will be targeted as traitorous, anti american, etc., this includes the press, other parties, the supreme court, whatever, anyone he sees as a member of the anti trump coalition is now a target, for now of rhetoric only, time will tell how far he will be allowed to go, and don´t think for a second there is something he won't do, or that he has any moral limit, he doesn't

Good luck trying to rationalize his government into anything other than an ego trip

His only weak point is succession, like every populists he hates the idea of giving power up, to anyone, even his children, he won´t groom a successor and get mad every time this point comes up

I sincerely hope current checks and balances work, but I´m not optimistic, lots of people will come under the spell, it´s tough times for anyone that can´t escape facts, reason, science, for those are now the enemies of the US government."
 
Not spot on in every point, but It came close to happen the first time around, now he has evolved and learned, it will be worst.
The paradox is that it was in the hands of Republican Senators to impeach and keep him out of office for good, they failed.
 
 

@The_Capt from over a year ago ...

Guys, c’mon.  If we are going to keep even a veneer of semi-pro analysis/assessment we are going to need to post citations for stuff like this.  I honestly don’t have a problem with Wikipedia deep even.  But we are just tossing stuff out right now.
I have never heard of these sort of ratios, most ratios go 1:1.5-2 against the attacker over time.  In fact in urban combat doctrine still calls for higher attacker advantage ratios - I have never heard “we can attack this city with a numerical disadvantage…huzzah!”
3:1 has been beat up a lot but we still look at these ratios, along with modifiers/multipliers in any planning.

Nadezhdin’s out, although there’s something wonky about the numbers being reported:
“Refusing to give up, Mr Nadezhdin said on social media that he would challenge the decision in Russia's Supreme Court.
"I collected more than 200,000 signatures across Russia. We conducted the collection openly and honestly."
The Central Election Commission said that more than 9,000 signatures submitted by Mr Nadezhdin were invalid.
That left 95,587 names, meaning he was just short of the 100,000 required signatures to register as a candidate, commission member Andrei Shutov said.”

I don’t know whether to be encouraged based on the assumption he’s been rejected because the government think he’s too popular, or discouraged that he only managed to get c.200k signatures in the first place.

I think there is more than enough evidence that Russian officers do learn. I mean, just read @Haiduk s latest posts. Since we all like our historical comparisons, this is like WW1. In the first half year or so casualties were enormous because neither side knew how to deal with the new battlefield realities. Both sides continued to feed bodies into the meet grinders with frontal assaults afterwards. That doesn't mean they didn't learn but that it took time to learn the lessons and to find solutions - and how to deal with what lessons the other side learned.
Sure, Russian officers don't seem to be the most... intellectually flexible type but evidently they do learn. Question is, do they lean fast enough to complete with Ukraine (who apparently have difficulties of there own) As @The_Capt keeps telling us this is about which side leans faster, institutionalizes it and manages to convert those lessons into advantages on the battlefield.

For us is just two sorts of sh...t
Nadezdin: "Russia will not lost SMO. I will not give the comamnd to withdraw without reconcilement of peact treaty. If there is a peace treaty - let we will discuss it. If there is no peace treaty - we will not withdraw, we will dig in on that place, which we hols ad will wait peace treaty"
All difference between "liberal" Nadezhdin and "dictator" Putin is Nadezdin doesn't want furher continuing of the war and conquering of full Ukraine. But he is for Russia must save in own composition already occupied territories and Ukraine have to accept this for the sake of peace. Else Russia will fight, not advancing, but defending own territories by all means. Also Nadezhdin categorically against any reparations from Russia side to Ukraine. 
  
 

Nadezhdin use for us is an indication of the problems Putler is facing.
He was never going to win, but he was needed to make a sham election "look" like an election. 
The fact he has been pulled speaks volumes about Moscows lack of control of the sham.
 

Zaluzhny actually admitted that he didn't deliver when in his Economist interview he called it his own personal mistake that he underestimated Russia's ability to regenerate force and resist the Ukrainian offensive this summer. With the fate of the nation at stake, I think it's pretty clearly Zelensky's role to decide if Zaluzhny should stay on the job and the latter making public statements about what should be done in opposition to the administration he serves isn't making that any easier. What politics there is *should* be in Zelensky's corner. Zaluzhny...like MacArthur as mentioned above...is inserting himself into that arena and will have only himself to blame if that gets him fired. 
Note: I say all of the above while liking and respecting Zaluzhny but generals don't get to call the shots in the kind of government I want to support. 
 

There are several layers to it. In crude terms, it's that Zalushny has a different vision going forward for how to fight the war that includes, inter alia, very large mobilization and an emphasis on drone warfare. He also feels quite comfortable talking outside the chain of command and in public to attempt to make his vision of the war apply going forward. Zelensky has what could be described as a more political take on the war but really it seems like the biggest issue is that he believes that the civilian primacy over the Ukrainian commander should be complete. It should not be a competition, whatever tensions may exist within the relationship and Zaluzhny has to some degree made it one...even if with pretty good intentions. 
I tend to agree with Zalushny's assessment in military terms but I think in the long run Zelensky has the right of it. If Ukraine is really going to reject the "Eurasian" model Putin sells than it has to fully buy into elected, civilian control of the security services writ large.
My 2 hryvnia. 
 

Hmm. I'm really not sold on this dronetopianism.
Granted, that is partly cap-badge tribalism. Mind you, in Commonwealth-style armies - and probably most Western-style armies - drones are likely to end up in fires units (that is, artillery units) anyway, so it'd be a change but not a loss of organisational influence. As an example, look at how rockets were absorbed by the artillery fraternity.
But I don't think tribalism is the only reason influencing my thinking.
To start with, it seems like a version of the old argument "if snipers are so great, we should just train every rifleman to be a sniper!" Yes, snipers are great, but the personal attributes required to be a sniper a rare, they're really expensive and slow to train, and the final product is also of niche utility. A notional 18th Sniper Brigade would either be tactically and operationally useless, or a very very expensive hammer being used to crack grapes. Drone operators seem to suffer from the same limitations, and without good drone operators all you'd have is a ToyWorld's-worth of spare parts sitting behind battalion HQ.
Drones are really good at hitting and destroying the specific point they're aimed at, but appear to be pretty terrible at damaging or even annoying anything else. That's kind-of ok as long as you have really super great - and really super reliable - tac ISR. 'Dumb' artillery is good at damaging the thing it's aimed at, and great at suppressing that as well as everything else in the general vicinity. PGM artillery shifts that seesaw towards damage and away from suppression, but retains both effects. Also, the drills required for a single artillery forward observer to remove an entire grid square from your "worried about" list are fast (ie, minutes), simple and well trained. Trying to do the same thing with drones would be extremely slow (ie, hours), require hundreds of operators, and consume the entire cognitive abilities of at least a bde HQ. Hopefully nothing "interesting" happens while they're busy with that.
Logistically, artillery ammunition is famously "heavy", but it is also very compact and simple. The rounds come packed in geometrically simple tubes, they stack really well, and are insensitive to heat, cold, dry, damp, and being bounced around and generally careless treatment. Drones appear to be light and simple and easy - hell, I can carry two in boxes under my arm, and get a dozen in my car! Well, sure, but how does that scale? When every rifle company is firing off 100 munition-drones per day, and every battalion is burning through a thousand ... where are you putting all the dunnage? Who is assembling them? How many trucks are running about in constant loops to bring them forward from Div HQ? Drones are also kinda fragile. That's partly why they are so comparatively cheap, but what is an acceptable dud-rate for drones?
Drones are definitely a problem for conventional artillery, as a supplement to traditional CB. They're also "competition" for conventional artillery, as a supplement to traditional fire support. And that's the key word: supplement. Not replace.
Incidentally, part of the reason 'mil-spec' kit tends to be so expensive is that wars are rare and peace is normal. Most kit spends most - like, 99.something% - of it's time sitting around waiting for a war. But it isn't really just sitting there, it's being used on a regular basis to support training and exercises, and that kit needs to stand up to that regular use and abuse, and then be ready to transition to war-use at a moments notice. That ruggedness and decades-long reliability costs money. COTS drones offer none of that. I get that the flash-to-bang time for drones in Ukraine right now is probably being measured in weeks at best but, again; wars are rare and peace is normal. For a standing army that is mostly at peace - like, oh, all of NATO for the last three quarters of a century - that $1000 COTS drone is going to have to last a lot longer than a few weeks. It's going to have to support multiple courses and exercises, over many years. Occasionally one will be expended in training but, as with Javelins or VLLAD missiles, that's going to be the exception rather than the rule. The rest of the time it just gets lugged around as make-weight.

It isn’t code in the classical sense.  You have the mathematical construct of a neuron which you code for and then have a vast interconnected web of these.  You then train it where the strength of those interconnections is adjusted during training.  It is analogous to our own brain where as we learn certain neural pathways are strengthened while others are pared away. It is almost a black box.
The rate of improvement in AI both in narrow and broad applications is nothing short of astounding.  I just worry the we are racing to develop the tools of our own demise with autonomous kill drones coupled with LLM AI that are becoming increasingly capable and  closer to sentience.  
I would rather hope for Kurzweil symbiosis instead.

https://v.redd.it/ud6kcx0m8tfc1/DASH_720.mp4?source=fallback
(Ai) image recognition in a lancet prefers hole in a wall more than ifv next to it. 
To add some Ukraine context 

Most things people consider AI right now, from self driving cars to chatGPT is almost as hard to understand as giving someone a brain scan to ask why they like the colour pink. The code is simpler than you would think, but the "thought process" is almost totally opaque. A major area of AI research right now is trying to get it to explain why it did stuff in a way we understand 

The term "AI" is doing a lot of legwork in this discussion and, I think, is being used to refer to different things a lot of the time.
As previously noted, "AI" of sorts is already integrated into weapon systems at the level of smart munitions and augmented feedback to operators of FPVs, for example.  The "AI" I think most people are concerned about/interested in in the context of the next few years is the kind of AI that will be able to meaningfully automate processes which have, to-date, been too complicated or nuanced to take away from human beings:  mainly target selection and prosecution within a defined combat zone.  That's all fine and those are the types of "AI" which will (I think inevitably) be integrated into our next generation fighting systems.
Beyond that there is the kind of AI that starts beying employed against the enemy's AI.  For me, this is where things start to get interesting.  I think at this point AI is at least as heavily employed in deceiving death swarms and Terminators as it is in driving them and that means that warfare will become extremely dynamic: the best way to defeat an AI-driven war machine is to make sure it doesn't recognise you in the first place and there are countless unimagined ways of making that happen.  War, warriors and weapons will only appear recognisable to our eyes for as long as AI doesn't get too good.  Once it does start to get there, we will simply change what they looks like (hold that thought).
And then we start saying things like:
Now please don't misunderstand me; I think that this is an interesting thought and idea to discuss but I also think that, in an effort to scout ahead, it has not-altogether-deliberately strayed a bit off-map.  "AI" does not mean the same "AI", any more.
If we ever get to the point when "wars are fought exclusively by AI systems" or when people are not involved in warfighting then I see that world reflecting one of two possibilities:
People no longer exist.  If they did exist then they would still throw shade, b***hslap each other and get into large scale brawls which would take the sociological place of whatever warfare is now that AI has excluded us from in the future and that would then become the new, real warfare.  In other words if we are ever excluded from "warfare" because AI is just too damned efficient and lethal then that will suddenly solve absolutely nothing and we will go somewhere else and start fighting again, without it.  The people who are unsatisfied by AI-controlled warfare will simply change warfare to be something else entirely.  Or; People get imaginative enough to realise that AI isn't best used to target enemy machines with explosives any more than a nuclear reactor is best used to heat the cavalry's stables.  If AI is in such a derivative state that wars could theoretically be fought by it to the exclusion of actual people then we should find a far better use for Marvin than what convention would currently consider the military domain.  If AI is this powerful it should be working primarily in the information domain, ironing out conflicting certainties (thanks for introducing useful terminology, Capt) at the level of the information people absorb and believe on a day-to-day basis.  In this way AI should be winning wars before we even know they've begun and yes, that means that, as far as we're concerned, AI should be preventing warfare altogether.  To the extent that such a thing may not be possible, AI should work to mitigate whatever level of conflict turns out to be necessary between human beings but that will probably still mean allowing us to do it ourselves in order to make sure something actually gets resolved in the process. Tldr: I think that, if AI advances to the point that it could exclude us from warfare altogether then the political and natural sciences, healthcare and economics will be the fields upon which those wars are won, not the trenches and treelines around Avdiivka.

The term "AI" is doing a lot of legwork in this discussion and, I think, is being used to refer to different things a lot of the time.
As previously noted, "AI" of sorts is already integrated into weapon systems at the level of smart munitions and augmented feedback to operators of FPVs, for example.  The "AI" I think most people are concerned about/interested in in the context of the next few years is the kind of AI that will be able to meaningfully automate processes which have, to-date, been too complicated or nuanced to take away from human beings:  mainly target selection and prosecution within a defined combat zone.  That's all fine and those are the types of "AI" which will (I think inevitably) be integrated into our next generation fighting systems.
Beyond that there is the kind of AI that starts beying employed against the enemy's AI.  For me, this is where things start to get interesting.  I think at this point AI is at least as heavily employed in deceiving death swarms and Terminators as it is in driving them and that means that warfare will become extremely dynamic: the best way to defeat an AI-driven war machine is to make sure it doesn't recognise you in the first place and there are countless unimagined ways of making that happen.  War, warriors and weapons will only appear recognisable to our eyes for as long as AI doesn't get too good.  Once it does start to get there, we will simply change what they looks like (hold that thought).
And then we start saying things like:
Now please don't misunderstand me; I think that this is an interesting thought and idea to discuss but I also think that, in an effort to scout ahead, it has not-altogether-deliberately strayed a bit off-map.  "AI" does not mean the same "AI", any more.
If we ever get to the point when "wars are fought exclusively by AI systems" or when people are not involved in warfighting then I see that world reflecting one of two possibilities:
People no longer exist.  If they did exist then they would still throw shade, b***hslap each other and get into large scale brawls which would take the sociological place of whatever warfare is now that AI has excluded us from in the future and that would then become the new, real warfare.  In other words if we are ever excluded from "warfare" because AI is just too damned efficient and lethal then that will suddenly solve absolutely nothing and we will go somewhere else and start fighting again, without it.  The people who are unsatisfied by AI-controlled warfare will simply change warfare to be something else entirely.  Or; People get imaginative enough to realise that AI isn't best used to target enemy machines with explosives any more than a nuclear reactor is best used to heat the cavalry's stables.  If AI is in such a derivative state that wars could theoretically be fought by it to the exclusion of actual people then we should find a far better use for Marvin than what convention would currently consider the military domain.  If AI is this powerful it should be working primarily in the information domain, ironing out conflicting certainties (thanks for introducing useful terminology, Capt) at the level of the information people absorb and believe on a day-to-day basis.  In this way AI should be winning wars before we even know they've begun and yes, that means that, as far as we're concerned, AI should be preventing warfare altogether.  To the extent that such a thing may not be possible, AI should work to mitigate whatever level of conflict turns out to be necessary between human beings but that will probably still mean allowing us to do it ourselves in order to make sure something actually gets resolved in the process. Tldr: I think that, if AI advances to the point that it could exclude us from warfare altogether then the political and natural sciences, healthcare and economics will be the fields upon which those wars are won, not the trenches and treelines around Avdiivka.

Key word "at some point". Despite we have now heavier "Mamont" drones, carring HE charge, but as I wrote recently drones can't substitute normal artillery shell (until it can lift it up and deliver to the target with proper speed), which could be combine both HE and fragmental effect. Usual FPV carries RPG-7 projectille, whih effective only against vehicles, less on fortifications and almost ineffective on infantry (except direct hit). "Mamont" FPV has good HE effect, so can be used against small infantry groups, blindages and infantry covered in the buildings, but it doesn't cause fragmental effect on the infantry. 
But when you have a lack of shells drones are single ultimate weapon which capable to hit enemy target precisely and dismoral enemy troops. We are already near that moment, when "at some point" the quantity will transfrom to quality.
BTW despite Russians lost 12 vehicles in this attack, they managed to reach own goal - later they forced our troops to abandon this positions by heavy artillery and avaiation strikes. 
Here is one more example of combined elimination of Russian infantry near Avdiivka from 47th mech brigade. A drone films from close distanse DPCIM impact, 25 mm Bardley gun work and FPV strikes. Warning - teared off and burning bodies.
 
 
 

Jeesh, I really need to work on my communication skills.
Tux you have given us a very thoughtful post, and for the small part of it I played a role in... 
My thoughts were:
1. It's unlikely humans will ever cede all of warfighting to AI
2. If they do, it will be humans intentionally ceding warfighting to AI
I think we have consensus, majority agreement, humans have already ceded parts of warfighting to AI, i.e. autonomous AD. My comment was based on my view the trend of handing off decision making and execution to AI will continue in ever more warfighting functions, as in the fields you list of natural sciences, healthcare and economics. That's not to say the entire field will be ceded to AI.

To be fair, it is objectively difficult to distinguish between a Russian airfield and a construction waste dump.

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.

Maybe, your thoughts on the comparison between the machine learning data base set for high intensity major power conflict targeting vs that for driverless vehicles?

Getting into part of what I do for a living here and actually, the job driverless car sensors and software must handle is literally the opposite of an autonomous suicide drone. The former must navigate every highly complex driven environment and avoid hitting anything. An autonomous suicide drone can be geofenced and must just hit the likeliest right thing most of the time. It's an order of magnitude easier.