Hacker News Reader

Comment by rich_sasha | Hacker News Reader

Only an armchair enthusiast, but: these diesel/electric subs are quite slow and have limited range, certainly compared to a carrier group. Even with the new(ish) Air Independent Propulsion, they can stay underwater for a long time - maybe weeks - but can't go very fast. Their limited endurance also means they have to stay close-ish to their bases. The AIP bit often relies on special fuels (hydrogen, ethanol, and in any case, compressed oxygen) meaning they can't refuel at any random port, or generate the fuel at sea, like recharging batteries from diesel power.

They are amazing and great at coastal operations, but I just don't see how they can chase a carrier group around. They could of course lie in wait and pick the right spot/get lucky, but I'm not sure if this is a viable strategy.

View on HN

See PRC recently revealed XXLUUVs (US will be building them too I imagine), large submarine drones, long endurance, no crew, i.e. expendable blue water one way munition launchers. Space ISR means all movement of carrier groups at any given time is essentially known. Group of XXLUUVs can be plotted on different intercept trajectories. They can be spatially distributed to handoff shadowing. The real problem is not even ASW anymore, imagine subsect of XXLUUVs fleet blasting active sonar not trying even to hide to find subsurface, queue meme of that trumpet boy following annoyed girl. i.e. loud drones that weaponize their expendability as trip wire. Imagine conop: we know where your expensive manned surface/subsurface platforms are, we don't care about being sunk, we're always going to be blasting sensors so we know if you're trying to take us out, mean while we're close enough that we can counter fire before your munitions reach us, and since we're in terminal range, we have weapons grade track all ready. If you ignore us, we're just going to keep following you with weapons grade track that we're broadcasting for other platforms, some of which are other XXLUUVs just outside your ASW range that you don't know about. This moves away from technologic subsurface cat and mouse to industrial/attritional, i.e. PRC can crank out more theatre level XXLUUVs than US can SSNs or their own UUVs.

I see this narrative everywhere: that drones are inherently, and always, much cheaper and effective. I think it's maybe more complicated than this.

In some cases, sure, you can strip out 90% of the ability and pay 1% of the price, and that's a great tradeoff. Eliminating the human means cost savings and greater expendability - a significant consideration is losing your trainer operators and not being able to replace them.

But let's take an advanced aerial drone. Not a one-way Shahed drone but a F-35 or F-22 contender. It will still need all or most of the advanced avionics that make F-35s expensive. If this is the expensive bit, then you can't save that much by just taking the human out.

So an advanced submarine will surely still be expensive if it aims for a similar level of capability.

And if not... Sure. You can make it Nx simpler and hope it is at least Nx cheaper. But even then there's the question of how expensive it is to defeat. Ukraine is ingeniously working to defeat Russian drones at ever lower costs, as Russia tries to send ever more and more of them at lower and lower prices. It's not enough to make something cheaply - it comes down to "value" for money. If these super-simple super-cheap Chinese drones can likewise be defeated by cheap weaponry, the impact is limited.

The other thing Ukraine war demonstrates is that there are limits to what you can do with cheap drones. They are fine for harassing each other - terrorist bombings of civilian areas, disrupting oil refining etc. But apart from stunning special operations done by drones by the Ukraine side, the deep and painful strikes are still done by expensive classical rockets - ATACMS, Storm Shadow etc.

> It will still need all or most of the advanced avionics that make F-35s expensive.

Those jets are catch-all do-everything machines, almost entirely because it's so crazy expensive to build jets + the political/contracting games played, that they are forced to jam everything into them. There's lots of efficiency and tactical gains in small niche applications where each platform is specialized for the job.

> Ukraine is ingeniously working to defeat Russian drones at ever lower costs, as Russia tries to send ever more and more of them at lower and lower prices. It's not enough to make something cheaply - it comes down to "value" for money. If these super-simple super-cheap Chinese drones can likewise be defeated by cheap weaponry, the impact is limited.

Even in Ukraine we really haven't seen 'cheap' defenses properly scale up to deal with the drone problem. Ukraine was getting 90% shoot downs in the early days but that's changed in the past year where Russia's tactics adapted. They've have been getting hammered weekly, even deep into their air defences along the polish border.

With Chinese SAMs and extreme range air-to-air missiles all of the fancy F-35 avionics will mostly be just them doing everything trying to survive those encounters far away from the frontline in Taiwan... where they mostly become very expensive missile launch platforms (bombers) and some radar networking. If you remove maximum survivability + narrow into niches those get a lot cheaper.

It's substantially cheaper after skipping personel + operational costs of training and associated maintenance etc. Keeping the human bits trained involves flight hours, platform wear and tear etc. Unmanned = do more with less, i.e. same airframes can focus on useful missions. Or sustain more with much less upkeep. Wasting hours wearing down platforms to maintain crew proficiency is SUBSTANTIAL over lifetime of highend platforms. XXLUUV fleet - skipping out on 1,000s of crew, and can be built to lower standards etc. The AI proliferates experience across all hulls simultaneously. For same acquistion you can acquire 2/3x size fleet, have most in storage while a few do day to day operations with significantly cheaper OPEX while maintaining readiness.

XXLUUVs aren't cheap or lowend, they are highend but PRC shipbuilding advantage over US simply monumental, i.e. they can match capabilities and win attrition game on budget. Like UUVs following surface fleet is basically DARPA ACTUV proposal... TLDR is instead of spending 500k-1M per day (US costs) on fleet size ASW you can use a few 10k per day drones that keep tabs on marked targets. PRC has ship building capabilities to execute this at fraction of cost. Once you remove manning, attrition based strategies become even more potent for PRC industrial base.

UKR demonsrates how shitty deindustrialized powers are at generating fires. PRC is has industrial base to make 30m cars and 20m motorcycles annually. This translates to industrial base that can output 5 digit shaheed tier munitions daily. This basically enough satuate any layered defense US+co can prepsition. That's just lowend. Medium end like cruise missiles PRC can probably do ~1000 a day, see their cruise missile gigafactory. The key difference between RU, is PRC (like US) has C4ISR to make efficiently use munitions. RU is closer to Iran level.

ATACMS and Storm Shadow range are functionally toys in IndoPac, i.e. we're talking about different scale of of highend warfare over much greater distances and magazine exchanges. UKR is frankly schoolyard fight and has no worthwhile lessons for Indopac except it's important to have strong industrial base for attrition game, i.e. RU able to sustain very incompetent exchanges vs entire US+NATO support. Incompetent as in wildly inefficient and constrained because they have shit C4ISR that can't dismantal UKR IADs or logistics insulated on NATO soil. There's no "sancturary" in IndoPac.

The useful lesson we learned in last few years relevant to highend peer to peer fight is basically shit tier missiles can penetrate the most sophisticated ABM in the world (Iran vs Israel), more than carrier groups has magazine depth. XXLUVs basically another layer of massing fires in quantities current surface fleet composition can't survive, but strips out ambiguity around long distance / standoff kill chains by parking satuation sized salvos always in terminal range. Again these are not cheap low end solutions, these are HIGH END solutions that PRC simply can build cheaply at scale.

> Wasting hours wearing down platforms to maintain crew proficiency is SUBSTANTIAL over lifetime of highend platforms.

That's a good point of course. But I guess that also means we have to figure out how to operate these vehicles with 0 crew - not even remote crew, because they would need training too. I guess that's something we can't do yet - if we could, the humans wouldn't be there already.

> XXLUUVs aren't cheap or lowend, they are highend but PRC shipbuilding advantage over US simply monumental, i.e. they can match capabilities and win attrition game on budget.

I don't doubt China's supermacy in either low or high end manufacturing. But how limited are they here by the ability to build lots of hills Vs to fill them with expensive sensors etc? It seems like a stretch to say that because they can make lots and lots of ships they can make lots and lots of sophisticated unmanned subs too. Either way, if China's industrial prowess is so much better than the US's, it sounds like they would beat NATO with or without drones - they are maybe an efficiency improvement, but if they put their heart into building normal subs, they would still out build the US.

I agree industrial capability appears to be key. NATO, vis a vis Russia, apparently knows they can't destroy all their tanks and kill every last soldier, and instead you need to target supply lines and command structures. AFAICT, this has basically informed th last few decades of NATO strategy. I guess the question is, what is the strategy "against" China. Because, you're right, if it's an attritional war then we're screwed.

> I guess the question is, what is the strategy "against" China.

- Eliminating their food calories and energy imports via effective/unrestricted submarine warfare (resulting in mass famine and internal insurrection)

- Strategic nuclear weapons

Autonomy is still prospective but high priority, US+PRC (even PRC) both have manning issues. Less total crew + different crew - cheaper to train mission specialists on software in office than actual exercises.

The problem is NATO subsurface barely exist outside of US, so they're not really relevant. Even surface fleet barely matters, if anything would probably drag US down because NATO would share US unrep. As for PRC industry they have AESA radars in <10k agriculture drones, their cruise missile gigafactory was CNCing turbofan blades, they expanded primary sub shipyard (bohai, huludao) to 24 bays (US has 5), which also does nuclear. When PRC smashes scale, it probably means they have unit costs down and something cooking.

>strategy "against" China

Honestly, I don't think there is one, or a sensible one. The "cope" strategy, is after distributing assets in 1IC, 2IC US+co can create survivable force structure that basically sinks all of PLANavy, which TBH was always the easy part. Somehow this translates to defeating PRC because it delays invasion. That's basically the TLDR of what most rational seems to boil down to. None of it talks about the fact that US+co can sink all of PLAN, but PRC mainland fires complex can basically dismantle US posture in 1st IC without a single ship, and basically lock down and islands (TW,JP,functionally SKR, PH) in perpituity. An umbrella of distributing assets are all the drone hellscape proposals that forgets PRC can outdrone everyone else. Or PRC medium/long range fires can hit/displace US posture 5000km around PRC, which includes CENTCOM/parts of EUCOM (no one ever looks west). Which is basically where thinking stops, i.e. all the strategy assumes because US can preposition hardware, it puts PRC at disproportion vunerability, but more and more we seek PRC capabilities will leak fighting outside of Indopac, including PRC pursuing global strikes, i.e. imagine if every US asset larger than 20 meters can be destroyed in under an hour. We're talking about all surface combatants, support ships, bombers, fuelers, awacs, fixed infra targets aka everything highend - the entire US expeditionary model. IMO is why golden dome exists, because PRC hitting CONUS is rocketry is a few years away from mainstream discussion. This without mentioning CONUS strategic targets. Or that if PRC dismantles US exquisit delivery platforms, they not only win the fires generation game (very hard to destroy sheltered launchers vs carriers, bombers that can be sheltered), or that in a both navy airforce gone scenario PRC has industrial base to reconstitute faster. Nevermind US+co has only fought vs adversaries with less than 50% US industrial output.

In the case of submarines that is necessarily the case though.

The people that go into a submarine aren't normal people, or they weren't back in the day anyway. They are people you would be really unhappy to throw away.

You can think completely differently about these platforms if those people aren't inside them.

> Space ISR means all movement of carrier groups at any given time is essentially known.

Do you have a source for this claim? I have argued with colleagues for years and they all say that satellites cannot track carriers

Track as in find carriers (then cue other assets to fix), not eyes on deck / line of sight / provide weapons grade track at all times. Movement as in we know carrier can only be in this small geographic area.

Nothing that states it as fact, but there's write ups about PRC yaogan IMINT constellations many years ago, that basically said when constellation size reach X, they'll have Y revisitation rates. We're at X & Y, and people in PLA watching, doing napkin math, using PRC commercial sat specs (jilin/gaofen) like swath size to determine that more or less on paper has functionally persistence in at least pacific theatre, i.e. the gaps were insubstantial that predictive algos can find carriers basically at all times, unless carriers can teleport. Combine this this PRC rolling out a lot of SAR in last few years (i.e. like 200-300 more satellites into have launched since initial articles) + wake detection algos, and on paper they should be able to persistently track carriers through cloud coverage. Of course cannot say definitely, it's just mathmematically possible now unlike a few years ago when constellation too small.

Again, none of this will be publically disclosed as "fact", it's more (as customary with analysis is PRC), the supporting evidence is so overwhelming that it cannot be denied. Extrapolate to PRC and US spaceforce rolls out megaconstellations sized IMINT, it's more or less over for carriers. Maybe stealth bombers or submarines.

> They could of course lie in wait and pick the right spot/get lucky, but I'm not sure if this is a viable strategy.

"we have only to be lucky once, you will have to be lucky always"

As a former Navy officer, the best description I ever heard of a conventionally powered submarine is "a mobile minefield." Nuclear is a different story.

Diesels are basically intelligent mines.

Reply on news.ycombinator.com