NATO is dangerously unaware that its military edge is slipping
“(...) war comes not when the forces of freedom are strong, but when they are weak. It is then that tyrants are tempted.”—Ronald Reagan
NATO faces its gravest military disadvantage since 1949, as the balance of power has shifted decisively toward its adversaries in the Era of Drone Warfare. The speed and scale of NATO’s relative military decline represents the most dramatic power shift since World War II—and the alliance appears dangerously unaware of its new vulnerability
NATO doctrine is dangerously obsolete
The Pax Americana is coming to its end.
Failure to defeat Houthi Rebels: The US Navy’s struggles against the Houthis’ low-cost drone arsenal demonstrate how even non-state actors can challenge NATO air defense. NATO forces found themselves hard-pressed to counter relatively primitive drone attacks.
Economic logic of drone warfare: In Ukraine, drones account for over 70% of combat kills—a proportion likely to increase. The economics are devastating: $1,000 FPV drones routinely destroy $7 million Abrams tanks. Modern tanks are dangerously outdated and the entire armored warfare needs to be redesigned from the ground up to face the realities of modern drone warfare.
Lack of military experience: NATO lacks both experienced drone operators and institutional knowledge of drone warfare. The alliance has no significant combat experience fighting against and integrating drones at scale.
Lack of NATO anti-drone air defense: NATO’s air defense systems, designed for traditional threats, cannot effectively counter drone swarms. Russian drones have repeatedly penetrated Polish and Western European airspace, with NATO struggling to detect, track, and neutralize these incursions.
Lack of drone production: NATO’s drone capabilities are shockingly inadequate. The UK army possessed fewer than 450 FPV drones as of August 2024, while a single Ukrainian factory produces 40,000 monthly. Both Russia and Ukraine now manufacture 3-5 million drones annually.
A quick Google states that Ukraine produces about 4000 FPV drones a day. Alarmingly, the same Google search revealed that the total procurement of FPV drones at this moment by the British army is 450 FPV drones! Western militaries are completely asleep at the wheel.
The most problematic is
Obsolescence of NATO doctrine: NATO’s military doctrine remains rooted in pre-drone warfare assumptions, creating an embarrassing disconnect where peacetime generals lecture battle-hardened Ukrainian officers who possess actual combat experience against a lethal, drone-equipped adversary.
What A Genuine Drone-Aware Doctrine Actually Looks Like
The best video on the realities of the modern drone-dominated battlefield is here. Unfortunately, it got removed from YouTube for being Russian propaganda.[1]
Dozens of drones per human soldier
Victory requires millions of drones monthly, not hundreds. A dozen or a hundred of drones per human soldier. Doctrine must prioritize massive drone manufacturing capacity, advanced battery technology, and supply chains measured in millions of units. Majority of military budget should eventually go to autonomous platforms.
Tanks, Navy, Infantry doctrine need to be redesigned from the ground up: Tanks need complete redesign—active protection systems, anti-drone cages, complete top armor. Infantry need to be trained differently, moving in smaller units, under constant drone overwatch, skirting in and out of quickly made drone-proof bunkers and using ground drones for logistics to minimize exposure. Navy should transition from few expensive carriers to distributed drone-launching platforms—hundreds of cheap drone carriers, underwater drone deployments, and autonomous loyal wingmen for the naval air-force.
The Transparent Battlefield: Assume constant observation. Every movement is tracked, every concentration targeted within minutes. Forces must operate dispersed, communicating through secure mesh networks, moving constantly. Resources, command, logistics—everything dispersed, redundant, modular.
Electronic Warfare Centrality: Every unit needs electronic warfare capability. But this is temporary: AI-enabled autonomous drones that operate without communication links are coming. Doctrine must prepare for both current jamming-dependent warfare and future jamming-proof autonomous swarms.
China’s Strategic Advantages
[For a more in-depth discussion I recommend my earlier shortform on the military realities of the coming Taiwan crisis.]
Beyond the drone revolution, China is in acscension.
Industrial supremacy: China dominates global production in critical military technologies—controlling the majority of the world’s drone manufacturing, shipbuilding capacity, battery production as well as a host of other critical dual-use technologies.
Technology parity: China has successfully replicated fifth-generation fighter capabilities (notably copying JSF technology through espionage) and is now mass-producing these aircraft at scale. In fact, they have leaped ahead of the US Air Force fielding the first 6th generation fighter prototypes.
Obsolence of US Navy doctrine: Aircraft carriers—long the cornerstone of Western power projection—are increasingly vulnerable to drone swarms and hypersonic missiles, probably rendering the doctrine of carrier battle groups obsolete.
Eroding advantages: Traditional US strengths—submarine superiority, stealth bomber fleets, control of sea routes— also face fairly rapid erosion with China quickly catching up.
FAQ: The Unrecognized Scale of NATO’s Military Disadvantage
Q: Isn’t NATO aware of these problems? Surely military leaders see what’s happening in Ukraine?
A: Yes, NATO officials acknowledge drone warfare’s importance and announce modernization programs regularly. But awareness and action are vastly different things.
They’re ordering thousands of drones when they need millions, updating doctrine paragraphs when they need to rewrite entire manuals. Moreover many modernization programs are very slow or never materialize at all. The official [often optimistic!] timelines for acquiring the most minimal amount of drones for most NATO forces is often on the order of years.
Q: But NATO is increasing defense spending and modernizing forces. Isn’t that enough?
A: NATO’s response resembles rearranging deck chairs on the Titanic. Member states are buying more tanks that drones will destroy, more fighters that can’t engage drone swarms, and investing in traditional platforms that are already obsolete. The UK’s 450 FPV drones versus Ukraine’s millions tells you everything. NATO is spending more money on the wrong things, guided by doctrine written before the drone revolution. It’s not about spending—it’s about spending on what actually matters in modern warfare. Modern drone warfare means one should have dozens of drones for every human soldier. Tanks and other mechanized divisions needs to be redesigned from the ground up.
Q: Why can’t NATO just learn from Ukraine’s experience?
A: Military transformation takes years. Critically, adaptation requires purging peacetime leadership and promoting combat-tested officers.
Peacetime commanders, promoted through political skill rather than battlefield success, psychologically have trouble accepting that much of their expertise may be obsolete. Unlike Russia, which purged ineffective leaders through brutal battlefield selection, NATO’s command structure remains unchanged— the embarrassing reality is that NATO generals who haven’t seen combat against a peer adversary are lecturing battle-tested Ukrainian officers.
Q: Doesn’t NATO still have technological superiority? Better training?
A: These advantages matter less when the fundamental grammar of war has changed. Superior training in obsolete tactics is worthless. Technological superiority in traditional platforms means little when a $1,000 drone defeats a $7 million tank.
- ^
Which is true. It was Russian propaganda. But it is also the best source on what the West’s adversary is actually thinking and doing.
It’s fair to criticize Western militaries for not adapting to drones more quickly, especially on the defense-industrial side. But, it’s also wrong to frame this as a NATO disadvantage in strategic terms.
On the strategic level, you have to keep in mind that:
NATO is a defensive alliance
NATO strategy has always been built from the premise that NATO is weakest (relative to rivals) in ground warfare.
With respect to Russia, a prolonged war of attrition overwhelmingly favors NATO.
On point #1, what we’re seeing in Ukraine is that drones are tactically a defense-dominant weapon—that is, they make it much harder to seize territory. If there’s a way to use them in successful, high-mobility offensive operations, no one has found it. NATO’s major goal has always been to defend European territory against Russian invasion. Given the tactical balance of drones, even if Russia ends up the “drone-advantaged” side, it’s still the case that the basically defensive nature of drones and the defensive aims on the NATO side probably end up favoring NATO.
On #2, NATO has always been built around having an edge in the air and sea (and nowadays space) compensating for relativeness weakness on the ground. Historically, NATO faced substantial disadvantages on the ground relative to the Soviets, and NATO strategy was always about finding ways to use its other edges to compensate for that. Post-Soviet Russia has never had the same kind of advantages on the ground, but this has still been the area of relative Russian strength. With respect to frontline ground forces, even post-Soviet Russia would obviously have the edge in the initial stages of any conflict and that’s been true ever since the end of the Cold War.
Drones don’t do anything to challenge the NATO air/sea/space edge. Cheap drones operate at low speed, low altitude, and short range. Something that can match the speed, altitude, and range of a jet fighter is going to look a lot like (and cost a lot like) a jet fighter even it happens to be unmanned. For the same reasons, drones are a threat to naval vessels operating close to shore, but so are speedboats, and there’s no scenario where NATO is trying to put a carrier group 10 miles off the Russian coastline. The situation in constricted waterways is different, and it’s entirely possible that drones, speedboats, and perhaps speedboat drones will substantially deny naval access to the Bab al Mandab or the Strait of Hormuz. But that’s just not the Russian scenario and the only such constricted naval area there is the Baltic, where NATO controls the land.
On #3, Ukraine is gradually losing the war because Russia has 4x the population and over 10x the GDP. In a Russia-NATO conflict, it’s NATO with the overwhelming population and economic advantages. So, if you bog down into a war of attrition, there’s no doubt NATO wins it. Again, if we see Ukraine-type dynamics, this is something that’s better for NATO than the situation a decade ago because it makes it a lot harder for Russia to seize the Baltics in a blitzkrieg and then extend a nuclear umbrella over them.
Curious if you have a theory/model for why this might be. The standard answer is “Drones sensor-suffuse the battlefield, making it impossible to mass forces and approach the front lines without being spotted early, giving defender many opportunities to hit you with artillery and drones.” But I’m not satisfied with this answer.
I think it’s also that drones can’t hold ground. If drones could do all the jobs of infantry and armor, they wouldn’t be offense-dominant or defense-dominant, but ground drones are not mature enough to dig a trench or engage in a firefight.
The fact that drones can’t hold ground is relevant if you are thinking of them as a replacement for troops, but if instead you think of them as a weapon for troops—a weapon with very long range, very high lethality, that also doubles as excellent scouting… then I don’t see why this would favor the defender?
...Someone elsethread pointed out that the attacker has to move whilst the defender doesn’t, which means that if ranges are long and visibility is high, that’s worse for the attacker probably (assuming they can’t shoot and/or defend as effectively while on the move as they can while not on the move, which is plausible.) This seems like a good answer to me.
Also, drones kinda sorta can hold ground, increasingly as they get better. There are ground drones now, and also quadcopters have developed a tactic of landing on the ground to conserve battery and conceal / take cover. Also quadcopters can fly inside buildings...
My loose model is that increases in lethality tend, all else equal, to favor defenders. Increases in operational mobility tend to favor attackers.
At some point in any attack, you need to cover some ground where you are exposed to enemy fire from defenders with the advantage of a prepared position. The more lethal their weapons, the harder that is and that’s true even if your weapons are also more lethal because they have the advantage of a prepared position. As currently used, drones boost lethality and they’re also uniquely good against armor which has been the offensive weapon of choice on the modern, high-intensity conventional battlefield for attackers. I guess that’s fairly close to the “standard answer” you’re giving above.
The second bit is that drones offer no operational mobility (as currently deployed). To advance an offensive over any substantial distance, you need to move not just weapons systems but also people, platforms, and logistics forward. That is, suppose you’ve either killed your enemies or forced their retreat from some given area, you need to consolidate control over it and be able to either prepare your own defenses against counterattack or establish a forward staging point for the next attack. Mechanized or armored units do that pretty naturally; you’re bringing people and supplies with you as you go and you can immediately use them for the next steps. Drones don’t. If you’ve overrun a position with drones, there is no obvious next step save for finding a way to move people and vehicles forward, who then become vulnerable to enemy drones.
Another piece of it—quite dependent on the present state of technology—is that jamming is effective enough that most drones are limited to the range of however far they can trail a fiber optic cable which is only a few miles. That rules out deep penetration with drones, again making it hard to follow up on any successes.
I think all those points are probably correct and most of the answer.
Another contributing factor is in the single-use nature of kamikazee style drones. I have this strong intuition from boardgames and computer games that kamikazees are defensive. A battle ending in ‘Everything on both sides is dead’ is a kind of succrsful defense
As a toy example, lets say that a force of 100 tanks can defeat an enemy force of 10 tanks with only one lost. (So 99 remain). This allows a focussed army to snowball, maintaining momentum.
But, if 100 drones take 10 casualties to destroy an enemy force of 10 drones. Then any offensive peters out as the larger army errodes down.
I suppose put another way, the power of an army is at best linear in ammo, probably sub-linear. But its often going to be super-linear (approaching quadraric) in units. Drones are more like ammo, they get expended. The attacker likes super-linear, it allows a snowball effect, it gives them something to counter the natural ways defenders are advantaged.
NATO is certainly not “dangerously unaware” that drones are flipping the table of armored warfare. Drones are a huge focus of the new Army Secretary, DoGE, and major defense contractors (particularly Anduril). A few months ago, we had what was sadly not called “Bring Your Drone to Work Day” with all sorts of new prototypes set up in the Pentagon courtyard for us bureaucrats to see and touch and get a real felt sense for what’s new.
But quadcopters aren’t everything! They certainly haven’t allowed Russia to conquer Ukraine, if anything they seem to favor the defender. American military power was already built on drones, in terms of intelligence, electronic warfare, loiter munitions, and even “traditional” precision bombs (as @Hastings pointed out below).
Yes, the incumbent Defense Primes are over-specialized in “exquisite” hardware that’s expensive, technologically advanced, and produced in low numbers. But that also means they’re very incentivized to develop drone countermeasures. Most things they try won’t work. Some likely will. Even early things they’ve tested have helped in Ukraine, it’s just not the case that an Abrams tank is “dangerously outdated”. There are more threats to a heavy tank than there have been in the past, and lighter tanks are a better fit for Ukraine’s geography, but you’d still rather have the tank than not. This is not always the case!
Western militaries are acutely aware of that viral tweet from a while back:
Yes, if given the choice, you should prefer six identical copies of the USS We Built This Yesterday + the ice cream support ship. That doesn’t mean the IJN Golden Kirin, Glorious Harbinger of Eternal Imperial Dawn was useless.
Disclosure: Literally written from the Pentagon. (Off duty, speaking in a personal capacity, waited to type this until after hours, etc.)
I’d be interested to discuss this more sometime. Take the Abrams for example. It costs what, six million dollars?
Who would win in a fight: an Abrams or six million dollars worth of drone troops? (What might this look like? Maybe 15 pickup trucks, each with a driver and a FPV drone operator in them. Each truck is loaded with 100 drones of various kinds.) Let’s be kind to the Abrams and give them a couple drones too, for scouting.
It seems clear to me that the drone force would win, and it wouldn’t be close. It would be like a WW2 battleship going up against 15 WW2 aircraft carriers. Curious if you disagree.
Perhaps you think that the Abrams would lose to the drone force, but that there are some kinds of missions that the Abrams is better for, and therefore the Abrams is not “dangerously outdated?” What would these missions be? Heavy assaults on fortified lines? Nah, probably still better to have the swarm of 1500 drones hitting the fortified line from 20km away. The Abrams isn’t even more survivable than the pickups really, I mean it’s better if you come under artillery fire probably but if you come under attack from FPV drones it probably takes about as many drones to kill the Abrams as to kill all 15 pickups even in ideal conditions, and realistically the pickups will be scattering and hiding and shooting back at the drones and probably it’ll take dozens of drones to get them all and also the men will have abandoned their trucks and hid in basements.
I agree that Abrams tanks aren’t useless. They are like the WW1-era battleships in WW2; still better to have than nothing; useful for shore bombardment maybe and useful as screening ships for the carrier battle groups. But you sure as hell shouldn’t be building any more of them; that would be a waste of precious resources that could be used to build other things—things that aren’t battleships at all, things like carriers and subs and destroyers. Similarly, there should be no new tanks built from now on basically, since every such tank would consume money that could instead be used to buy more useful things.
What drone countermeasures do you think are promising?
I’ll be at Lighthaven next weekend, Friday and Saturday, happy to discuss in person. This isn’t my focus, but I can present some common views. I can’t discuss specific developing tech or countermeasures, and generally don’t know the specifics anyway. Some sort of countermeasure always develops, though how costly and effective it is, how it changes the various warfare niches, remains to be seen.
It’s worth noting that tanks will basically always lose a one-on-one fight to dismounted troops of an equivalent cost-to-equip, given reasonable cover, morale, and equipment. This was true in 1940, in 1970, in 2000, and now. Sending unscreened tank columns alone into battle in anything other than a flat desert is suicidal. Tanks shine in combined arms, but are vulnerable on their own. Combined arms warfare is extremely difficult to coordinate; neither Russia nor Ukraine have been able to pull it off much in recent years, with the initial Kursk offensive as a notable exception. It shouldn’t surprise us that heavy tanks struggle in geography they’re not suited for, used by armies who are unable to use them to best effect. That is not the only relevant scenario.
Ok, interesting- but what would be the relevant scenario, then, in which it’s better to have a tank than the drone force I described?
Maneuver warfare. Combined Arms Offensives. Breakthrough operations against prepared defenses in high-intensity conflicts. Counter-offensives to stop enemy advances (i.e. Kursk).
Here’s some published US Army discussion of this problem. Yes, Armor officers have tanks and are motivated to say they’re the solution to every problem. But they have a point that other countries failing to successfully execute combined arms does not mean that NATO would. There’s some things we’re good at, skills that we’ve invested in disproportionately compared to peer competitors. Joint Operations at all scales (nations, services, combined arms), is top of that list.
https://warontherocks.com/2025/07/steel-in-the-storm-recent-wars-as-guides-for-armor-transformation/
I probably won’t be around then alas but if I am ill try to find you! Thanks!
I guess 5 Abrams and 30 million worth of drones vs 60 million worth of drones might be a better comparison. I think I’d still favour the drones but it’s much less obvious.
Good point. Yeah I would still favor the drones, and by a lot. The abrams would be about as useless as the Yamato, probably moreso even.
My view on where the tanks might win is: there’s a point at which you basically saturate your capability at “whatever drones are good at” while there might be some other job tanks are good at (my vague guess is that this is something like “attacking well defended positions”—they’re fast, take specialized weapons to defeat, and have big guns), and you’re better off having that capability than further saturating your drone capability. But I’ve little in the way of quantitative insight about where saturation might occur, nor how good tanks are at attacking.
A particular point I’m a bit confused about: I’ve often seen people saying: tanks need infantry support to be safe. However, aren’t infantry and tanks both vulnerable to drones?
Shoot them with a bullet. The anti-aircraft guns of WWI and WW2 would be very effective at cutting swaths through drone swarms.
What pickup trucks are you using? How much do the pickups cost? What armour do they have if any?
There’s one drone operator per pickup. These are FPV drones so they’re limited to at most a handful of drones at a time. You can’t swarm the Abrams, and unlike what the videos would have you believe the chance of an individual drone taking out an Abrams is tiny. The tank has plenty of time and opportunity to blow up the pickup.
The drones aren’t as cheap as you believe—the FPV drones with fibre optic used in Ukraine are many thousands of dollars each. Each pickup, operators, and drone is worth many hundreds of thousands and is likely a sitting duck to artillery, tank fire, and yes counter drones, especially when it moves.
The play would be to sneak the truck in under cover of darkness, set up shop somewhere camouflaged, and then use the drones to help defend the current area, and atrite enemy forces. Basically the same thing as is happening now in Ukraine. It helps in a slow grinding war, but doesn’t help you in a manoeuvre war.
Nato doctrine is all about manoeuvrability and air power. Once air superiority is achieved your pickups are sitting ducks. Only individual people can act effectively against air superiority.
The aim of the tank in that situation is rapid movement and firepower, whilst being protected from most attacks. The pickup can easily be blown up by an enemy ATGM, RPG, or drone operator so just isn’t as useful in manoeuvre warfare. The driver can easily be killed by an assault rifle.
Giving individual troops drones is obviously a power multiplier but with current drone technology I don’t think a drone carrier makes much sense—too exposed in a manoeuvre war, and no different to what’s currently going on in a war of attrition.
(Of course all this changes once we can coordinate fully autonomous drones at scale and low price)
For future reference, please put text that is pretty close to LLM output into expandable sections and flag them as such. For a relatively fact-heavy post like this LLM output is great and often helpful, but I don’t think we are doing anyone any service by dressing it up as human writing. This is generally part of LessWrong content policy, and we would have rejected this post if it had come from a new user (this doesn’t mean the core ideas is bad, indeed I find this post useful, but I do really think the attractor of everyone pasting content like this is a much worse attractor than the one we are currently in).
Needless to say, as sovereign ruler of lesswrong I will abide by your judgement. But forgive me for asking a few questions/comments
1. This is not raw or lightly edited LLM output. Eg all facts and overall structure here are based on a handwritten draft.
2. The LLM assistance was about writing flowing, coherent prose which (for me at least) can take a lot of time. Some may take offence at typical LLMisms but I fail to see how this lowers the object-level quality. I could spend hours excising every sign of AI- but this defeats the purpose of using AI to enhance productivity.
3. That said, if the facts were also LLM generated and I handchecked them carefully I fail to see how this would actually lower the overall quality—in fact my best guess is that LLMs are already much much better in many-most domains than many-most people. eg twitter has seen marked improvements in epistemic quality since @ grok is this true happened. The future [and present] of writing and intellectual work is Artificial Intelligence. To claim otherwise seems to be a denial of the reality of the imminent and immanent arrival of a superior machine intelligence.
4. Pragmatically, I find the present guidelines to be unclear. Am I allowed to post AI-assisted writing if I mark it as such? If so—I will just mark everything I write as AI content and let the reader decide if they trust my judgement.
If not—what’s the exact demarcation here?
Very reasonable questions!
As I have learned as a result of dealing with this complaint every day, when being given a draft to make into prose, the AI will add a huge amount of “facts”. Phrasings, logical structure, and all that kind of stuff communicates quite important information (indeed, often more than the facts via the use of qualifiers, or the exact use of logical connectors).
In addition to the point above (the “writing flowing/coherent prose” part very much not actually being surface level), there is simply also an issue of enforcement. The default equilibrium of people pasting LLM output is that nobody is really talking to each other. I can’t tell whether the LLM writing reflects what you actually wanted to say, or is just a random thing it made up. That’s why I recommend putting it into a box.
I agree! LLMs are indeed actually quite great at generating facts. They are also pretty decent at some aspects of writing and communication.
There is no doubt the future of writing and intellectual work is AI. My guess is within a year or two something big will have to change how LessWrong relates to it (just as we had to change within the last year how we relate to it). But for now AI is not yet better than the median LessWrong commenter at the kind of writing on LessWrong, and even if it was at a surface level, there are various other dynamics that make it unlikely therefore the right choice is for LW to be a place where humans post unmarked LLM output as their own.
I mean, you can put all your writing into collapsible sections, but I highly doubt you would get much traction that way. If you mark non-AI writing as AI content that’s also against the moderation rules.
Simply keep the two apart, and try to add prose to explain the connection between them. Feel free to extensively make use of AI, just make sure it’s clear which part is AI, and which part is not. Yes, this means you can’t use AI straightforwardly to write your prose. Such is life. The costs aren’t worth it for LW.
>I mean, you can put all your writing into collapsible sections, but I highly doubt you would get much traction that way. If you mark non-AI writing as AI content that’s also against the moderation rules.
Simply keep the two apart, and try to add prose to explain the connection between them. Feel free to extensively make use of AI, just make sure it’s clear which part is AI, and which part is not. Yes, this means you can’t use AI straightforwardly to write your prose. Such is life. The costs aren’t worth it for LW.
I’m surprised you are taking such a hardline stance on this point. Or perhaps I’m misunderstanding what you are saying.
The primary use-case of AI is not to just post some output with some minor context [though this can be useful]; the primary use-case is to create an AI draft and then go through several iterations and go through hand-editting at the end.
Using AI to draft writing is increasingly default all around the world. Is LessWrong going to be a holdout on allowing this? It seems this is what is implied.
Apart from the present post I am betting a large fraction of LessWrong posts are already written with AI assistance. Some may spent significant time to excise the tell-tale marks of LLM prose which man… feels super silly? But many posts explicitly acknowledge AI-assistance. For myself, I so assume everybody is using of course using AI assistance during writing I don’t even consider it worth mentioning. It amuses me when commenters excitedly point out that I’ve used AI to assist writing as if they’ve caught me in some sort of shameful crime.
It seems that this ubiqituous practice violates either one of
>> You have to mark AI writing as such.
>>If you mark non-AI writing as AI content that’s also against the moderation rules.
unless one retains a ‘one-drop’ rule for AI assistance.
P.S. I didn’t use AI to write these comments but I would if I could. The reason that I don’t, is not even to refrain from angering king habryka- it’s simply that there isn’t a clean in-comment AI interface that I can use [1]. But I’m sure when they I’ll be using it all the time, saving significant time and improving my prose at the same time. My native prose is oft clunky, grammatically questionable, overwrought and undercooked.
I would probably play around with system prompts to give a more distinct style from standard LLMese because admittedly the “It’s not just X it’s a whole Y” can be rather annoying.
[1] maybe such an application already exists. This would be amazing. It can’t be too hard to code. Please let me know if you know any such application exists.
Yep, the stance is relatively hard. I am very confident that the alternative would be a pretty quick collapse of the platform, or it would require some very drastic changes in the voting and attention mechanisms on the site to deal with the giant wave of slop that any other stance would allow.
Making prose flow is not the hard part of writing. I am all in favor of people using AIs to think through their ideas. But I want their attestations to be their personal attestations, not some random thing that speaks from world-models that are not their own, and which confidence levels do not align with the speaker. Again, AI-generated output is totally fine on the site, just don’t use it for things that refer to your personal levels of confidence, unless you really succeeded at making it sound like you, and you stand behind it the way you would stand behind your own words.
We are drowning in this stuff. If you want you can go through the dozen-a-day posts we get obviously written by AI, and proposed we (instead of spending 5-15 mins a day skimming and quickly rejecting them) spend as many hours as it takes to read and evaluate the content and the ideas to figure out which are bogus/slop/crackpot and which have any merit to them. Here’s 12 from the last 12 hours (that’s not all that we got, to be clear): 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12. Interested in you taking a look.
Random thought: maybe it makes sense to allow mostly-LLM-generated posts if the full prompt is provided (maybe itself in collapsible section). Not sure.
I think posts that are just “hey, I thought X was important, here is what an LLM said about it” seems fine. Just don’t pass it off as your own writing.
He’s trying to tend a garden that an invasive species has just been introduced to. It’ll be easier if we use class signaling as described by Scott Alexander to distinguish ourselves.
I’m happy to signal that I’m a low-class individual, mouthpiece of the AI slop, if that helps.
Bio-supremacists such as yourself can then be sure to sneer appropriately.
Claude assisted?
Point #1 was raising alarm bells on my mind, too naive, so thank you for asking the question.
Of course.
I liked and upvoted it, but would’ve (if available) liked sources. I know it’s kind of hard to read the situation on this in detail, and especially to get hard numbers.
Strong-downvoted for being indistinguishable from unedited ChatGPT output.
(That apparently Claude wrote it doesn’t matter.)
I reread this comment and want to correct a misunderstanding. I think this may explain some of the rather vehement anti-llm responses that I found so puzzling earlier.
I used a LLM to write this post, like I advice anyone whose time is valuable to do as saves a lot of time. That said, it is not unedited at all and I spent many hours going through various drafts and iterations.
2. The post wasn’t written by telling gpt “write a piece about natos military edge slopping”. This was based on a handwritten initial draft outlining every technical point. This is based on several years of tracking this topic semi-seriously as an amateur military buff.
I would say it is the testament to the extraordinary quality of 1shot LLM responses that they occasionally show that is now expected that people assume that the above was a simple unedited prompt.
Why walk when you can bike ?
Because it’s against LW policy:
But why listen to me when you could listen to the pocket PhD?
Why do some people think that’s bad? Roughly:
It breaks the social contract of discussion.
On forums like LW/StackExchange/etc, the implicit deal is: “You are reading my thoughts.”
If you post raw model output, readers are actually getting: “Here is a generic sample from a text generator, lightly prompted by me.”
That feels like misrepresentation, especially if not disclosed.
It’s extremely cheap spam.
Human-written comments cost time and attention.
Model-written comments are nearly free and can be produced in unlimited quantity.
If everyone does that, discussion quality drowns in fluent but shallow text. Downvoting “ChatGPT-y” comments is partly a defense mechanism against being flooded.
Low epistemic reliability.
Models confidently hallucinate, oversimplify, or miss key cruxes.
When a human writes, they can be challenged: “Why do you believe that?” and they (usually) have some model of the world behind it. With a raw LLM comment, there often isn’t a stable belief or understanding behind the words—just next-token prediction. That undermines the goal of rigorous reasoning.
Skill atrophy and shallow engagement.
If you mostly outsource your arguing/thinking to a model, you don’t get better at reasoning or writing. From the community’s perspective, you’re contributing less original thought and more “generic internet essay”.
Style + content are often generic.
LLM text has a distinctive “smooth, polite, yet vague” feel. People go to niche forums for idiosyncratic, deeply-thought comments, not for something they could get by clicking “generate” themselves.
Why the “why walk when you can bike?” analogy doesn’t quite fit
Biking vs walking:
Both are you moving under your own power.
Biking just makes you faster/more efficient.
Using raw AI output is more like:
Sending a delivery robot to a meetup in your name and letting it talk for you.
You gave it the address and a topic, but you don’t fully control what it says moment-to-moment.
Using AI as a tool (drafting, brainstorming, checking math, summarizing sources) and then carefully editing, fact-checking, and putting your own reasoning into the result is more like using a bike or calculator.
Dumping unedited Claude/ChatGPT output as a comment and treating it as “your contribution” is what people are objecting to.
So: it’s not that “biking” (using AI tools) is inherently bad; it’s that outsourcing the whole comment to the AI and presenting it as your own thought breaks norms around effort, honesty, and epistemic quality, and communities push back on that.
I understand and verified to the best of my ability the information contained in the post. If a LW moderator want to take action, I welcome their correction.
EDIT: I checked and the post contains about a 1000 words. At a 1 minute per 50 words this would be about 20 min. I have probably spent at least 3 hours on drafting this post, plus additional edits and engagement.
Quite frankly, the concerns raised here seem to be more originating from a ludditist denial and an intuitive dismissal of AI stylistic choices rather than genuine issues with the content—A stubborn attachment to the weakness of the flesh if you will. This post has generated quite a lot of content-level engagement so seemingly the style is only an issue for a loud minority.
The lesson I drew from Israel vs Iran is that stealth just hard-counters drones in a peer conflict. The essential insight is that guided bombs aren’t just similar to small suicide drones, they are drones- with all the advantages- as long as you can get a platform in place to drop them
I think stealth hard counters a technologically backward opponent like Iran. If one’s opponent has very few or very old sensor systems, then stealth can grant the ability to just fly over someone and drop bombs on them, which is pretty great, because bombs are cheap.
But in (say) a peer US-China conflict, drones are also sensing platforms.
In a potential conflict with China, China will have sensors on numerous (non-electric) unmanned drones, on submarine drones, on high-powered manned aircraft further back, on their satellites, all along their coastline and so on. They’ve had decades to prepare while stealth was the obvious number one American strategy; this is a very-well considered response. “Stealth” becomes a much more relative term; the advantage almost certainly looks more like, “Well, we can get N% closer before they see us” than anything transformational.
Anyhow, in a peer conflict, due to all the above, I expect even relatively stealthy US platforms like the B-21 to be using their stealth to haul LRASMs just a little closer before dropping them off; and even with conservative deployment like that, I expect notable losses.
If the US actually tries to drop bombs with them I expect very high losses (in the timeframe of days / weeks) that the US will be completely unable to sustain over time.
It’s worth noting that Israel’s strikes on Iran also involved using their special forces to disable Iranian air defense, using special forces and—of course—drones. So Iran’s already-inferior defense against Israeli bombers was already notably degraded before overflights.
And if anyone’s striking anyone with drones in an unexpected way in a peer conflict, it’s not gonna necessarily be the US doing the striking. Ukraine destroyed a few hundred million dollars of Russian bomber hardware by smuggling a handful of cheap-ass drones into Russia; is China or the US going to have an easier time doing a similar feat? Are anyone’s current air bases well protected to this? Whose airbases have more hardened shelters that prevent such cheap attacks?
A couple notes: israel and russia are extremely comparable in military spending, likewise Ukraine and Iran. In addition, Ukraine and Iran both went hard into drones to counter the disparity, its very noticable that drones basically work against Russia and basically don’t against Israel- but neither conflict provides dramatically more evidence than the other about a war where both sides are well-resourced and nuclear. In particular, the lack of drone based attrition of the israeli airforce is glaring.
Also, I agree that Israel made great use of drones to take out anti-air defenses. However, this use of drones in no way requires manufacturing millions of quadcopters.
Thank you—this is important. Slight disagreement that “Modern drone warfare means one should have dozens of drones for every human soldier.”
The SOTA in drone technology changes so quickly (eg radio to fibre to AI) that any inventory of drones becomes obsolete. This is why the industrial base you mention elsewhere is so critical. Maybe “Modern drone warfare means one should have production capacity to produce dozens of drones per month for every human soldier.”
There are growing concerns about the coherence and effectiveness of Western institutional frameworks. NATO is sometimes called brain dead. look at the situation that the US refuses to aid ukraine in this war. Clearly, causing damage to russia in this war is worth a lot to the US. Instead there are secret meetings with russian officals and weird russian-authored peace deals they try to force on ukraine.
When it comes to china, one comanpy (nvidia) can probably force the us government to sell its primary edge to china just so nvidia can jack up the price of GPUs.
Democracy itself is barely what it used to be with the electorate being essentially illiterate and simply not informed about the facts.
Any citations on electorate being more ‘illiterate’ (in the relevant sense) then, say, 1960.
I have a hard time imagining how it have possibly been any worse than now. I mean look at the presidents that were elected back then. Skimming a newspaper or having it recited to you once every other month is probably better than the information distribution system we have right now.
I don’t agree with everything here but offers some sources: https://jmarriott.substack.com/p/the-dawn-of-the-post-literate-society-aa1
Thanks! Yeah makes sense, something like extremely adversarial / exploitive information environment, coupled with general decay in elite culture (and mass addiction to vertical video, social media).
Drone carriers seem extremely important. Probably half the missiles and all the attack helicopters on every ship should be replaced with drones that cost 1/20th as much. Eg for amphibious assault ships, the America class, which carries 20 helicopters and some F35s, seems completely obsolete given it has no drone infrastructure. The Chinese Type 076 at least has a small catapult for drone launch, but both should really carry at least 50 large jet powered drones, 500 loitering munitions, 500 recon drones, 20,000 FPVs as well as surface and undersea drones.
Just how expensive are missiles and how bad is NATO at producing them? This article claims that the US navy has 10,000 VLS cells but only about 17,000 missiles (mostly the small SM-2 type), meaning the entire fleet couldn’t be reloaded even once after depleting their magazines.
Unfortunately, I basically agree with this post.
However, here are two counterpoints that I take seriously:
(1) Air power. In a few years there will be fighter-jet-sized drones designed for air to air combat, but for now at least, NATO would probably continue to dominate in the skies. This is especially important e.g. for a war over Taiwan.
(2) Space power. Thanks mostly to SpaceX, it seems like the USA would have a huge advantage in launching stuff to orbit. I don’t have a good sense of how big a deal ‘space supremacy’ will be in e.g. a Taiwan war, but I suspect it will be huge (e.g. maybe it’ll allow you to actually hide some of your major force movements, since the enemy won’t have satellites, whereas not only will you be able to see their stuff, and possibly even deploy bombs/drones/etc. from space to strike deep into enemy territory)
I asked Claude to build this table for me, and spot-checked the links a bit:
2024 Data (from McDowell’s Space Activities 2024 report)
2025 Data (partial year, through Q3/Nov 2025)
Key Notes
China’s 2025 payload mass isn’t tracked publicly in aggregate. Based on China’s 70 successful orbital launches through early November 2025 Payloadspace, and assuming similar average payload masses to 2024 (~2.6 tonnes/launch), a rough estimate would be ~180–200 tonnes, but this is my extrapolation.
NATO is dominated by the USA (which is dominated by SpaceX). European NATO members contribute relatively little mass—Western Europe launched only 46 tonnes total in 2024 Planet4589.
SpaceX alone accounts for roughly 85–90% of global payload mass to orbit.
Best source: Jonathan McDowell’s annual reports at planet4589.org/space/papers/
In a full out war, the side with a disadvantage in space would probably try to introduce Kessler syndrome.
Kessler syndrome isn’t permanent though, and it only affects orbits where there is a high density of satellites. If SpaceX can continue launching they would launch to higher orbits with low debris density, and very low orbits where the high surface area : volume ratio of debris means it reenters fast but the satellite needs to use a bit more fuel.
What they can do is deny LEO due to Kessler syndrome, and deny slightly higher orbits by launching 10^9 tiny ball bearings into retrograde orbit, which will cause several unavoidable collisions per year with any satellite passing through that altitude. I think this forces adversaries to retreat into VLEO which is shielded by drag.
Space power means NATO drones larger than a few kilograms will probably be unjammable, because they connect to Starlink. China just had a study and the upshot is you need air superiority and hundreds of large EW drones to jam Starlink, which is difficult to achieve when your own drones are vulnerable to EW. I’m not sure if NATO would actually have enough of an mass advantage in space to shoot down Chinese satellites, but it seems like satellites can be small enough that hiding “major force movements” would be difficult for any party
Thanks Daniel, you bring up some good points.
On the airforce issue I would be inclined to agree with you but the following video is doing the rounds on twitter right now: https://x.com/teortaxesTex/status/1993637089900900540
Not sure what to make of it, but some concerning claims: only 750 out of 2000 jet fighters in US airforce are combat ready. Average number of flights hour per pilot has dropped to 110 a year, where previously anything below 150 would be considered not combat-ready. Chinese fighter pilots average 200 hours a year.
Lots of individual mistakes here, that together serve to severely overstate the case made:
The Houthis were unable to touch US naval power. What the US couldn’t do is defend ships in a very narrow stretch of water from drone, missile, and speedboat attacks. This is a very specific situation, and it’s like saying there’s an end to US ground power because they couldn’t decisively defeat the Taliban. Asymmetric warfare works, more news at 10.
Also note the drones they were using were far from cheap, often costing hundreds of thousands of dollars.
You do see the videos where the tank gets blown up by the (closer to $5000) drone, you don’t see the vast majority of videos where it does nothing. Meanwhile how many people were killed by the Abrams tank, and how useful was the Abrams tank when used correctly to break through enemy lines and regain movement—something drones are not really capable of.
Given that Nato’s doctrine assumes air superiority, which drones barely impact at all, I don’t see how you could possibly draw such a conclusion from the war in Ukraine, where both sides decisively lack air superiority.
The seas are huge, and cheap drones are short range and slow. Enemy ships are difficult to find in a vast empty sea. Ships are extremely difficult to destroy or cripple. Communications are almost certainly jammed. For drones to be useful they need to be:
fast
long range
carry a large payload
autonomous
We call such drones “cruise missiles” and they are extensively deployed in all NATO Navies. If you know a way to make them cheaper, then the DoD will almost certainly be very interested.
Current drone warfare is all about low cost, slow, short range, non-autonomous technology. Autonomous wingmen would by high cost, fast, long range, autonomous. I don’t see how you could possibly draw conclusions about them from current events.
This is basically current NATO doctrine, and has been for years.
This is true, but note that China is investing heavily into this stuff (and aircraft carriers), not autonomous drone swarms.
I have seen zero evidence of this. Indeed hypersonic missiles are similarly vulnerable to air defences as non hypersonic ones, and come with a whole host of problems of their own.
Ok, to state what probably should be obvious but which in practice typically isn’t: If the US does have a giant pile of drones, or contracts for a giant pile of drones, this fact would certainly be classified. And there is a strong incentive, when facing low-end threats that can be dealt with using only publicly-known systems, to deal with them using only publicly-known systems. The historical record includes lots of military systems that were not known to the public until long after their deployment.
Does that mean NATO militaries are on top of things? No. But it does mean that, as civilian outsiders, we should mostly model ourselves as ignorant.
Fairly confident this is not the case with small first-person view drones.
It is not easy to produce in secret. There is no good reason for it. No other NATO country has built them at any scale. China owns most of the supply chain currently. And perhaps most importantly: a stockpile of fpv drones mean nothing if you dont have experienced drone operators to fly them. These would have to be trained en masse, further complicating any secrecy.
Additionally, Western (small) drones are massively overpriced and inferior to Ukrainian and Russian made drones. This is incongruent.
Apparently, even Anduril’s drones are considered inferior and are not deployed much in Ukraine.
One thing that NATO has that Russia and Ukraine both don’t is a really good air force. If the US Air Force had been fighting in Ukraine, it would have quickly established air superiority over Russian forces and be able to bomb them at will, making it impossible to hold territory any more than Saddam Hussein’s army could. (Of course, Putin could have retaliated with nuclear weapons...)
I do not know how good China’s air force is.
Nuclear weapons still beat drones.
I think the best way to express this reality that I’ve heard was from a buddy, who mentioned offhandedly that it is wild we live in a time when if Ukraine and Russia teamed up tomorrow (by magic or whatever, obviously this isn’t a real thing) they would rapidly defeat the standing European militaries. That’s a historical anomaly big enough to choke on.
And, importantly, it wasn’t true at the start of the Russian invasion of Ukraine. At that point, neither of these nations were producing millions of drones annually. Somehow, since going to war, both nations have grown dramatically stronger, in lockstep. It’s a deranged outcome. If you read it in a novel, you’d laugh out loud.
Ukraine + Russia drone manufacturing goes up like a rocket, while France/Germany, which are not presently on fire and doing a draft...their drone production goes from nothing to nothing. It’s wild.
See also Drone Wars Endgame (RusselThor, 2024).
You retracted your request for additional sources but let me say that there is an enormous amount of background sourcing I could have added. I have added a little bit and cpuld his is a bunch of extra work; currently unsure whether there is enough demand for it. If there is, I can do some more digging.
Eg for the china stuff I wrote the following earlier: https://www.lesswrong.com/posts/tDkYdyJSqe3DddtK4/alexander-gietelink-oldenziel-s-shortform?commentId=LQtg5EtEKKpYJJJfP
Penetrating NATO airspace with small Shahed style drones is easy and can cause a lot of damage.
Sufficient damage would lead to massive NATO retaliation though. Russian air defense systems are stretched thin as is. NATO air forces could suppress what’s left and bomb with impunity. Plausibly NATO just bombs the drone factories and some other military targets.
Building enough air defenses to stop massed drone attacks is very costly. Deterrence through retaliation is cheaper than building lots of hard kill and jamming systems.
Russia is poking NATO in various ways but a shooting war would go very badly for them.
Regarding nukes, yes, they’re scary but if Russia starts bombing NATO cities or military targets on a large scale NATO will retaliate regardless of nuclear brinksmanship.
China
The China Taiwan thing depends on China’s goals. They’re not getting TSMC. Taiwan will go scorched earth on that to say nothing of ASML remote killswitches on production equipment. If the goal is to destroy the semiconductor supply chain, why bother? Blockade accomplishes the same thing with slightly less political fallout.
Whether China has an easier or harder time with a land war to take and hold the Island has little to do with the things western powers care about (semiconductor supply chain). Yes, drones help with taking and holding Taiwan and china can make a lot of them.
Aside from Taiwan, I just don’t see China as wanting to start a land war to take a neighboring country.
Longer term
TLDR:Modern cheap drones are not stealthy in audio, thermal or radar so systems already developed by western military suppliers can take down cheap drones at low cost and in large numbers.
Hard kill anti drone systems seem likely to work. Infantry in Ukraine are using shotguns currently but automated gun turrets with fast tracking and full auto cannons firing rimless shotguns rounds should work. Air-burst electronically fused shells also exist and are cheap to produce in quantity for longer range. Working systems exist and can be purchased from western suppliers.
Ukraine is getting some Rheinmetall Skyranger 35s soon
That’s in addition to all the Gun turret + sensors + software solutions like the EOS slinger also being sent to Ukraine.
Tactical rock paper scissors still applies. The anti drone system can’t stop incoming artillery rounds(though with enough range, the spotting UAVs die), direct fire from an enemy tanks or certain kinds of fast missiles. Drones aren’t the only threat on the battlefield.
Limited supply of effective hard kill systems is the issue right now. A good hard kill system will eat drone swarms all day (assuming ammo supply holds out and reloading is fast)
Much longer term
Active protection systems and counter UAS converges to use fast tracking gun mounts and airburst munitions. This is also effective against infantry including infantry behind cover. Perhaps some limited indirect fire capability. Things get even worse for infantry on the future battlefield.
One side gets more bang/buck by spending to develop more cost effective weapons (UAS, UGV, etc.). Live fire exercises to get performance data + lots of simulation becomes the norm. Fewer unknowns like morale. This over determines the winner in potential conflicts.
Large military bases and civilian infrastructure generally remain hard to protect. Defender has to harden every possible target. Retaliation/deterrence is what stops others from just launching long range one way attack drones at cities.
Interceptors might solve this problem but prospect of retaliation is enough.
Terrorism
Weaponised drone technology could proliferate and solve the “deliver payload” problem but control of explosives is doing most of the work already. Terrorists can already plant remote or time detonated bombs. CF:gwern’s excellent Terrorism is not about terror post. Don’t expect huge issues there. High value targets like politicians and CEOs will have to worry about small UAVs that don’t need a large payload in addition to snipers. Mostly no impact on general public.
I basically know nothing about this topic. But I saw this video, and was kinda shocked. Were the drones in the removed video, better than this?
Going down a bit of a rabbit hole now, I’ll link other videos that seem relevant:
- Drone Swarm: Again, idk anything about this topic. But this seemed pretty scary. I expect the product being advertised would be ~easy to counter against.
Have you been following the Ukraine war closely? The reality on the ground in terms of drone warfare is now far beyond that Anduril advertisement. You can find videos if you search for them. It’s quite terrifying.
I haven’t. Where should I search for these? Preferably, raw footage, non-news broadcast?
https://x.com/bayraktar_1love/status/1993614328256422125
this twitter account shares a whole host of Safe-for-work on the ground footage
All this may be valid, but I would think politics and geopolitics also matter enormously, since they should determine who Europe’s enemies are going to be. It’s no secret that parties like AfD in Germany, the National Rally in France, and Reform in the UK have the potential to change Europe’s direction substantially, and that the US, Europe’s military big brother, is thinking about exiting NATO entirely, perhaps handing control of European forces to a German general as a transitional step. You mention that Pax Americana is ending, but focus solely on military issues. Ideology is changing too.