The AI war and how to win it | Hacker News Reader

The AI war and how to win it

alexw.substack.com

100 pointshardmaru130 comments

This is a very urgent-sounding article about China's supposed AI superpowers and how they will dominate the battlefield, coincidentally written by the head of an American "AI" company who would likely personally benefit from shifting $150B into AI research, as suggested. Complete with AI-industry standard claims about technology that is "less than 10 years away", and sci-fi about "autonomous, adaptive drone swarms".

War is so much more complicated than the author wants readers to think. Raw technological dominance does not assure victory, as the US found after 20 years in Afghanistan. AI, for all its impressive feats, continues to operate without flexibility or common sense, and foundational concepts like geography or economic influence remain the most important contributors to success in the military sphere.

Even if one takes the continuance of the DoD's aims as an unalloyed good, what areas of its budget would go unserved with 25% redirected to AI research? How do those cuts or reprioritizations affect US military capability? I don't suspect the author has thought about this very much, and yet he feels qualified to write yet another AI/China scare article, which more and more seem designed to terrify politicians who determine funding rather than to convince military experts.

To your point, I literally laughed out loud at this in the article:

Based on the pace of progress with AI technology today, I believe this is less than 10 years away.

How does anyone write things like this and expect to be taken seriously with the track record of AI regarding real world things like autonomous driving?

Surely someone has a "law" that states generalized AI is always 10 years away.

It's usually predicted as 25-50 years away because of:

https://en.wikipedia.org/wiki/Maes%E2%80%93Garreau_law

A 10 year prediction is because this is the CEO of a company saying it. So the lifespan used in the prediction is that of the company and not the person.

> autonomous driving

Autonomous driving is "hard", where hard is mininizing unintentaional death for domestic adoption.

Autonmous murdering is easier, perhaps much easier. Because flying from point A to B without obstacles is simpler than driving, and droning a few thousand innocents due to permissive ROE is forgone. It's a vastly different calculations in terms of allowable deaths and hardware attrition.

Autonomous murdering is easier, perhaps much easier.

Is it really though? One variable that autonomous driving has in its favor is nobody is actively trying to disrupt it.

What happens when people put up nets around key areas? Fishing line? GPS jammers? Lasers to disable cameras? LiDAR blocking? Radar jamming? EMP? Radio jamming? All of this is being studied today and will at least keep pace with AI advancements. For just one example:

https://nationalinterest.org/blog/buzz/chinese-engineers-sho...

Autonomous vehicles don’t have to deal with any of these. Autonomous murder devices most assuredly will because people have a strong incentive to prevent their own murder.

IMO yes, mass attritionable UAVs exacerbates fundemental disadvantage of anti air - hard to defend everywhere vs air power easier to concentrate to create local tactical superiority.

Countermeasures can keep pace but doesn't mean they can match scale. It's typically going to be easier to mass UAVs who can use standoff munitions than to preposition defensive assets, especially land based ones, in sufficient concentrations which can be countered by adding more UAVs during mission planning. Countermeasures for high performance UAVs, essentially unmanned fighters instead of low tier drones will be very difficult. Advanced US/PRC HALE drones fly much higher and potentially faster, there's a reason counters like DE or various form of jamming target their ordnances at lower attitudes, not high flying, fast flying drones themselves which still requires combination of high performing ground and air based interceptors that can hit high. Except UAVs can tactically operate without life preservation in mind, which means much more permissive/aggressive operations = even more difficult requirements for counter-air infra.

Ultimately it's balance between offense / defensive is matter of industrial base, producing and positioning enough hardware to overwhelm opponents. Effects on battleground will be situational, i.e. large powers can already overwhelm smaller powers, and for peer powers it will be theatre specific, i.e. PRC can more easily overwhelm US carrier groups with mass drones because CSGs has finite VLS cells, power generation etc for defense whereas land based aviation can be massed at scale. Alternatively, PRC's very comprehensive anti-air network might be more easily penetrated by aggressive UAV formations using tactics manned missions won't attempt.

You have a valid point. But as a person who has worked in industrial automation, and also the military, yeah automatic murder is probably 50x easier than preventing human bodily harm.

Just having machinery around people is inherently dangerous.

But aside from complex concepts, The mine for example, is potentially one of the first Automated Killing devices, and is notoriously good at it.

Integrating technology into simple killing devices isn't very difficult compared to keeping people safe from harm. And while I agree that there are plenty of good ways to stop automated killing devices, you have to make big sacrifices to stop them, and that's another point to be made about this unique situation.

You either have to outpace your opponents technology, or be willing to fight without some or all of your own, because when you start introducing stuff like what you listed above, you can no longer use your own systems of defense, which then makes you vulnerable to other conventional forms of attack, giving the advantage to the technologically superior team.

In my opinion 'Terminator' might not be 10 years away, but automated killing devices are already in the field, and they aren't being reported on because they are classified. You would not believe the money and effort that goes into recovery of physical assets (working and destroyed) that are in theatre.

How much do you think the Ukrainians (or the Russians, if you're a psychopath) would pay for a fully automated missile that detects the source of incoming artillery, launches in that direction, identifies the gun, and blows it up?

It really doesn't seem much of a stretch from the current state of technology. The only thing missing is an AI-driven targeting system for the last mile. I sincerely hope somebody (on "our" side) is working on it.

I don’t because “their side” will eventually reverse engineer it as they have with all other weapons

Not very different from systems like shotspotter deployed in us cities, modern couter artillery systems

You mean like https://en.m.wikipedia.org/wiki/Counter-battery_radar ?

Ukrainians are using Polish TOPAZ system that does mostly that. Not AI needed (radar systems). Probably radars or satelitte images could be imroved with AI though. I think with somethink like starlink you could use tracking tech for tracking foreign military but US seems to have something like that. Probably also smart to automate analyst like in Russia-Ukraine war you could source info from relevant Telegram channels. Also tracking of social media for any leaks (seems like it’s present already).

What ukrainians needed was their own satelitte that got crowdfounded as US didn’t want to share all intel in real time. What they need now is long range rockets and then Russia will start peace negotiations.

This is a somewhat dated (not as in obsolete, but as in used for decades) and normal part of warfare. Artillery combat generally involves "shoot and scoot" strategy since the units tend to be identified quite quickly, in some cases before their first shells have even landed.

All AI would do is pull the trigger in the end. And I'd always rather that be done by a human. If AI were in charge of nuclear deployments during the Cold War, it's likely we would have seen nuclear annihilation since there were genuine false positives where a nuclear response was only stopped by a human breaking the chain of command and refusing to escalate as he "should have." [1] With AI we needn't worry about such 'discipline problems.'

[1] - https://en.wikipedia.org/wiki/Stanislav_Petrov

Counter battery radar only tells you where the enemy was when they fired, not where they are now. Even with drone surveillance, the drone has to find the target before it gets shot down, and then the information has to be acted on before they move. Also counter battery radars are themselves are vulnerable targets. It doesn't seem implausible to suggest that within a decade there will be missiles or loitering munitions with image detection programs for seeking out enemy artillery (or ships, or tanks, or even infantry) that would render shoot and scoot much less effective. It wouldn't require anything like AGI. There are all sorts of ethical issues with this idea, but the PLA isn't going to care about that. Not really sure how the West is going to handle the fact the PLA will have weapons that much of the western public are against developing in principle. I guess that's another parallel with nuclear weapons.

> "War is so much more complicated than the author wants readers to think. Raw technological dominance does not assure victory, as the US found after 20 years in Afghanistan."

I'm not sure that really proves your point, in that the US' problem in Afghanistan wasn't that their tech dominance wasn't enough - it clearly was, they pretty much had control of the country for years, no?

Where they failed was that in a fairly rare example of a "victory target" for a war, wanted to not just win militarily but also to change the minds/culture of an entire country to the point where they could withdraw and have the country keep acting the way they were enforcing while there.

Sure they failed at that, but had the goal been to take control of Afghanistan and make it a US colony, permanently occupied for the foreseeable future, that would've been doable (at least technically, ignoring US domestic politics etc etc)

Historically most wars have been about seizing land/wealth/people, not about changing a country's culture and then walking away. Surely drones which patrol the skies and kills anyone at the first sign of any sort of resistance (holding a gun, crossing a forbidden line, whatever) would've made holding the country a lot easier from a technical point of view.

Imagine if the US army had, when in Afghanistan, been given the same orders as say UK was giving back when building an empire "make them submit, kill those who don't, that's going to be our land" is there any doubt that their military strength wouldn't have been enough against Afghan rebels who in the real case could hide in mountains until America ran out of time and withdrew (which wouldn't happen if they hoped to own the country)?

So sure I agree with you that war isn't as simple as technical dominance as the reasons for war etc come into play, but since most wars have different reasons to the US' i Afghanistan... wouldn't Putin love drones that could fill Ukranian skies and kill anyone at the first sight of a weapon (while avoiding Russian soldiers)?

WWII was very much about changing the culture, and the US was pretty successful at that in Germany and even Japan (read about Japan in 1930s, it's terrifying).

The USSR was also somehow successful (see Eastern Bloc), fortunately much less so.

Some of this could be interpreted as satire. For example, it's not necessary to improve AI capabilities in order to generate believeable propaganda - nation-states have been doing that ever since the printed press was invented, and before. It certainly reached a fever pitch in the 20th century, with radio, television, print and internet media all being the scene of numerous pitched propaganda battles, with social media being (perhaps) the dominant information-warfare battleground of the 21st century.

As far as the Ukraine drone war, that's not really different from the Afghan drone war, the main difference being the relative lack of energy infrastructure or heavy armor to attack in Afghanistan. Whatever one can say about Russia vs NATO-backed Ukraine, or the USA vs Afghanistan, neither serves much purpose for making predictions about an 'all-out war' between China and the United States.

One glaring omission is the failure to recognize that China is a nuclear power (as its ally North Korea), and it's not hard to imagine a serious naval conflict between the USA and China going nuclear - what better way to eliminate an offshore aircraft carrier battle group, for example? This article mentions nuclear once (in a 1942 context) and barely discusses what a real naval conflict would look like, other than to claim that 'autonomous adaptive drone swarms' would be harder targets to than aircraft carriers. That's a science fiction story line from, say, the Iain M. Banks Culture novels, not a realistic scenario for naval warfare (and what about submarines?).

All in all this reads like a bit of hyperventilation, or maybe a sales pitch put together by the AI warfare divisions of Lockheed and Northrup. It also ignores the rationale for 'having a human in the loop', but then maybe the author hasn't seen the 1980s movie War Games?

This is relevant document.

The ability to automate and make 'good quality' content is decisive in the information space. The world is greatly moved by photos and video of horrible things, and if that can be faked with high fidelity it will make a huge difference. People won't have time for the truth - the 'video' of 'whatever' will make it's impact.

The drone war in Ukraine is completely different from that in Afghanistan. The Baryaktar drones are almost useless, have an average lifespan of just a few weeks. They're almost all destroyed and don't play an important role. FYI many of the Baryaktar videos we saw over the months were released over time, as the units were mostly destroyed. Russia has no equivalent. Small drones are used for recon, target acquisition, dropping small munitions, which is nothing like Afghanistan. And the scale of it as well. We're seeing the firs 'small drone war at scale'.

I suggest China, being China, will try it at real scale, and during the invasion of Taiwan I expect to see swarms of thousands of drones on the offence - overwhelming numbers.

Everyone is aware that 'nuclear weapons exist' and also that AI works to varying degrees of autonomy.

I doubt China would just blow up a bunch of American aircraft carriers with a nuclear weapon, unless they were under existential threat. China would be severely disadvantaged in a nuclear conflict I would have thought, and would try not to go there under any circumstances. North Korea even moreso. As such, non-nuclear approaches remain relevant.

As a naturalized US citizen who was born in China but intends to raise a family as full-bore Americans, it's hard for me to assess the China threat.

I believe China very much wants to assert it's influence in the Asia Pacific + MENA region, but I just don't it trying to conquer the United States directly. It's never really had a history of imperialism like the West has (excluding the Mongol empire that collapsed almost immediately).

That said, a country doesn't need to directly conquer another to be a threat. Still, it's hard to know how serious the threat really is—the post is compelling but also assumes as a forgone conclusion that China has aggressor ambitions.

China was never an imperial force like European colonists, but it always saw itself as the “Middle Kingdom” and the most important civilization on earth. It may be a different paradigm but the right to power or hegemony is still embedded in the culture and ethos.

It's actually a healthy and non-threatening paradigm?

If they think they're the best and have no reason to expand because of that, the only threat is to our egos.

Tribute. China always expected tribute from its neighbors.

The benefits in return for tribute were often greater than the tribute, but it did show a high level of domination of its neighbors for most of its history.

maybe you should tell that to the tibetans, uyghurs, taiwanese, etc.

That's it exactly. They don't tell us how we should treat native Americans or Cuba.

But we have it baked in that we're the arbiter of what every country does globally, and a history of military action putatively justified by those moral concerns.

we have a critical security interest in taiwan remaining independent -- zero moralizing needed. any empire, nation, group or individual human has some sin you can point to, china is not unique, and neither is the US. the US will lose the throne to china in the 21st c if it takes taiwan. there will always be a throne, the only question is who is in it.

That critical interest is short term until we get the Arizona fabs up. Then it'll be back to standard levels of ideology, nationalism and domino theory.

And if it's just about power for its own sake, fine. But the records of military adventurism are in fact starkly different over the last 200 years. Xi might ramp it up and we might ramp it down, but for now the spectre of Chinese militarism reminds me of those two hunter guys from South Park, "it's coming right at us!"

It was - and still wishes to be 'Imperial' in it's direct sphere of influence, which is to say, most of E and S/E Asia.

China is not ever going to invade the USA of course.

What do you think imperialism is other then "exert influence"?

Also,China has invaded India and claims many Indian states as its own.

Plus it invaded Vietnam (and lost)

Extract taxes, i.e., the Roman version?

Considering American aggression in Iraq and Latin America, China is a dove. However, it is true that China has been aggressively expanding borders into half of its neighboring countries. They are also extremely internally aggressive committing genocide against Uyghurs, Tibetans, Falun gong etc.

A counter balance to the American threat is not a bad thing.

> It's never really had a history of imperialism like the West has

Vietnam, Tibet, Mongolia, and Xinjiang might beg to differ and that’s before considering that the leader of China was literally an emperor up until recently. However, if by “imperialism” you mean across oceans and hemispheres then point taken but that isn’t really a virtue as China has never had a navy capable of doing so. It still doesn’t although it sorely wants to.

> the post… assumes as a forgone conclusion that China has aggressor ambitions.

With the advent of wolf warrior diplomacy, the aggressive and illegal construction of military bases in disputed waters all over the South China Sea, the interference in US elections, the rampant IP theft along with constant attempts to steal military secrets, along with threats of total war whenever the US sneezes in the direction of Taiwan… what would you think?

> wolf warrior diplomacy, the aggressive and illegal construction of military bases in disputed waters all over the South China Sea, the interference in US elections, the rampant IP theft along with constant attempts to steal military secrets, along with threats of total war whenever the US sneezes in the direction of Taiwan

Dyed in the wool American here. All of these pale in comparison to the history of our my nation, which has been at war with much of the world for most of its existence. Rampant IP theft is nothing compared to the bombing of civilian infrastructure in Iraq (among other countries in the Middle East) during the 1990s and 2000s, the clusterfuck of atrocities that was the Vietnam War, to say nothing of our extensive support for dictators in Latin America throughout the 20th century. When it comes to interfering with elections (and indeed the results of them), we put China to shame. I love my country, but I can't pretend that our military industrial complex has been anything but a threat to countless other nations and will certainly continue to do so.

Well said, I was thinking the exact same thing. This guy acts like this is a foregone conclusion that we must have war. It's a choice to stand back or to stand up.

And given the aforementioned disparities, we should really consider if giving China it's space in a multi polar world is more in our best interest than going to war to keep the hegemony alive.

> This guy acts like this is a foregone conclusion that we must have war.

> we should really consider if giving China it's space in a multi polar world is more in our best interest than going to war to keep the hegemony alive.

Non-sequiturs right out of CCP talking points.

Not to mention that IP theft in the late 18th into the 19th century was basically the bedrock upon which the US's industrial and economic systems were built...

The Berne and Paris conventions didn't come to be till the late 1800s --which both the US and China are signatories of.

Good points. But that doesn't mean China won't copy this. Because the model for global power has always been ruthless force. One empire after another.

This requires context.

The current global trade network is a consequence of a massive coordination of many nations helmed by the United States during the cold war. That empire is not an empire of one, it is an empire of many. All conflicts the US has been involved in during and since are downstream of a geopolitical chess game aimed at coordinating many many nations against would be hegemonic challengers.

That may seem like a preposterous justification given some of the conflicts the US has been in and the death of the USSR, and there is certainly plenty to criticize about our stupidity and unilateral behavior and corruption. But I don’t think it’s possible to overemphasize just how big World War 2 was or what the lessons were.

An even bigger global dictatorship than Hitler’s helmed by Stalin was very much a possibility. And the only way to stop a nuclear superpower helmed by a ruthless dictator from building a global network and achieving hegemonic power is to build your own first. World War 2 taught the world that dismantling your war machine allows the ruthless to build a juggernaut while you sleep.

The lesson of World War 2 was to keep the war machine in democratic republics dominant at all costs because dictatorships will surpass you militarily if you don’t.

The reality of that is ugly and in many ways the actions of the US are unjust. I think it can and has been made more just over time and can and should be kept in check by as many independent actors as possible that ensure overreach is impossible and bad actions are addressed.

But it is not comparable to China, not just because the CCP is far less powerful externally and far less accountable internally, but because China is not coordinating with the world in ways that benefits the world. China is far more authoritarian than the US, far less sophisticated in it’s propaganda than the USSR, and far more reliant on home base when extending itself. The US managed to build global logistical networks because more countries benefitted from the US global trade network than didn’t and our propaganda about the material well being we can deliver is actually true, albeit with caveats. The scale of US hegemony is impossible to enforce without the majority of participants voluntarily cooperating. In fact one of the populations hurt the most (though obviously not nearly as much as those killed in the wars that enabled the hegemony necessary for global trade, whether ostensibly or in reality) was the domestic middle class of the US itself.

Despite the flaws, and no matter what some claim, the US is the most foreign friendly benevolent global empire in world history, and the prosperity we ushered in is unprecedented. The technology that has lifted the world out of poverty simply would not be possible without global trade, and the amount of global conflict has (or had, considering how disruptive the Ukraine conflict has been) never been lower. The framework about rule of law violations and unjust conflict used to criticize the US most is something the US made possible on a global scale.

Criticism of the US is vital. It is valid. Our flaws need to be reckoned with, and we are currently facing a crisis of purpose that we have played a significant part in creating with our overemphasis on consumerist material well being at the expense of other considerations, which has created domestic decay. That does not mean a world without a corrupt and stupid global policeman optimizing for material prosperity over other forms of prosperity would be better than the one we have, or that any other organization is better suited to guard against would be global tyrants on the level of Stalin.

It is also imperative to get as many good people high up in that hegemonic military machine as possible. If we denigrate our military too much due to its (many) failures, we risk repelling good people who could keep it in check from enduring that grinding machine and steering it for the better.

So tl;dr, I love my country too, and hope we keep it together and make sure we fix and improve what we’ve got, not dismantle it out of a sense of global fairness that history has sadly shown we cannot rely on.

This is a whataboutism. We aren’t discussing US foreign policy we are discussing if China is a threat to the US. Saying “we’re bad too” doesn’t answer that question, it’s a non-sequitur

> All of these pale in comparison to the history of our my nation

You're responding to an argument about why the US sees China as having aggressive ambitions by saying the US is allegedly worse? That is irrelevant to the argument and implicitly accedes to my point. Besides, if the US was the evil empire you allege it is it would just make even more sense that China was seen as an enemy given the recent increase in Chinese international belligerence and influence campaigns.

> bombing of civilian infrastructure in Iraq

I can only find a single credible example of the US deliberately bombing civilians in Iraq and that was the Al Jazeera office in Baghdad and that assumes you believe the Mirror reporting is accurate. If it's true then that's bad for sure, but it's not exactly Xinjian-level genocide either. I was surprised with how few accidental bombings there were given how badly conceived the war was!

> atrocities that was the Vietnam War

Vietnamese are still dying from unexploded ordinance etc from that war, but, the Vietnamese are friendly with the United States while distrustful of China. Why is that? Because they have a very very long history of enmity with China, with the most recent conflict being a Chinese invasion of Vietnam in 1979, whereas the USA is seen as preferable to China. Given how terrible the Vietnam War was this stark difference stands out!

> our extensive support for dictators in Latin America throughout the 20th century

Fair criticism if you don't remember that China props up North Korea, a range of dictatorships and corrupt governments in Africa and is itself a dictatorship over more than a billion people!

> When it comes to interfering with elections (and indeed the results of them), we put China to shame.

This is true and would be a reason for those other countries to have enmity to the US but we haven't been able to influence Chinese elections because they aren't free elections and their leader is a dictator. This hasn't stopped them attempting to make our elections less free which is a symptom of their aggression towards the United States, the whole point of my original response.

> I can't pretend that our military industrial complex has been anything but a threat to countless other nations

If by "nothing but a threat to countless other nations" you mean that it's implied the US Military would fight any other country then you're technically correct. If you mean that all they ever do or ever have done is threaten every other group of humans then that is totally ridiculous. NATO is a voluntary defense pact that the US spearheads that Finland and Sweden are currently in the process of enthusiastically joining. Australia and the UK weren't forced at gunpoint to sign on to AUKUS. Dozens of other countries aren't being forced to cooperate on regional security issues. I'll grant the issue is obviously extremely complex but you're denying all the positives which is nonsense.

The longer Xi Jingping stays in power the less I worry about China on the international stage.

The subleaders will need to do increasingly extreme things in order to show results to the Supreme Leader. That's starting to play out--the Apple shakedown, Chinese police stations in foreign countries, etc. Those things have finally gotten the Western powers to realize that China is a genuine enemy, and they are starting to retaliate.

In addition, now that Russia has basically been dismantled, the defense system in America needs a new boogeyman. Congratulations, China, you're on deck.

> Russia has basically been dismantled

Western press proving day-in day-out how primitive Pravda was.

Oh my god that's so true. It's like they've invented hire order lying.

Russia hasn't been dismantled unfortunately. There's a good chance they end up with more territory.

Russia is definitely taken apart. They may be able to hold that bit of territory with suicidal bodies in trenches, but they are in really bad shape.

They are down to ancient weapons, ancient tanks, their Navy almost entirely defunct.

Most of their 'fast air' remains.

And they are going to have a very difficult time rebuilding with sanctions, and the severe economic problems with talent and resources.

They could 'come back' over a decade but I suggest Europe is becoming ready for that, and without the offer on the table of a nice, juicy, undefended territory, neither Putin nor his successors will have much luck.

It's the ambiguity that's the problem.

> They could 'come back' over a decade

With what replacement population for their current military age population? No, it's now or never for Russia.

They have enough population.

Which will simply become controlled by China--if Russia can even hold onto the territory (this winter will be telling).

Russia is now effectively a pending Chinese vassal state caused by their coming rebuild. No one else is going to support them, and Russia has wiped out a big chunk of their workforce--both young labor as well as older brainpower. So, Russia will have to turn to China for both labor and brainpower at the cost of sovereignty.

The only thing that would prevent Russia from becoming a Chinese vassal is if they somehow wind up with a leader smart enough to pair up with India.

> The only thing that would prevent Russia from becoming a Chinese vassal is if they somehow wind up with a leader smart enough to pair up with India.

Neither China nor India have the military reach to defend Russia. The overarching reason Russia is waging this war is how blood curdingly difficult it is to defend Russia at the best of times with local populations, short (as possible) supply lines, and decades of preparation. Now bring in an Asian power that would have to extend its reach over thousands of miles through tundra and desert with hostile populations along the way. They're not even going to attempt to do that.

The threat is increasing influence across the world.

I wonder how the world would be different had Zheng He sailed his junks east rather than west.

I tend to worry about the opposite problem: what if the United States is successful in deploying artificial intelligence? In the article, the author uses the United States interchangeably with the concept of liberal democracy, which I believe is a fair assumption today. However, I worry that many of the author's calls to soak up more and more information for the purposes of military supremacy will cause irreparable damage to our political system. Artificial intelligence has the tendency to centralize power in the hands of a small number of people, and breakthroughs in AI are likely to exacerbate this trend. In short, the US might win this hypothetical war, but at what cost?

The way I see it, there are three kinds of technologies: A) those which tend to centralize information or control B) those which are essentially neutral and C) those which democratize information and control.

Group A artificial intelligence capabilities are not hard to imagine. Ubiquitous surveillance is a good example. Many current AI technologies are great for surveillance and can dramatically reduce the marginal cost of surveilling a person. For an autocracy this is great. For a liberal democracy, it threatens civil liberties and reduces trust.

Group B technologies are things like self-driving cars, precision agriculture, and AI-facilitated entertainment (e.g. recommender systems, Dall-E). These are things which I don't see as fundamentally helpful or harmful to liberal democracy. Another way to look at these technologies is that both autocracies and democracies should be equally amenable to developing and using them.

I can't think of any Group C technologies and this worries me. These are AI-enabled technologies whose effects are disproportionately democratizing. A democracy should seek these out while an autocracy will ban them.

The consensus algorithms used by cryptocurrency fit into group C.

Smart contracts are programmable incentive mechanisms. I am hoping we figure out how to use them to coordinate productively, rather than for financial schemes.

I feel like it's just to difficult to pin down what area of ai will improve. For example what about an AI that is relatively powerful at brute force decryption(not sure if this violates some thermodynamic principle). Or even just an AI that can easily find exploitable websites. Things like that destroy trust in systems and probably large systems like gov or corporations have more to lose if information is up for grabs. But maybe it's the opposite way around. Hard to tell imo

For Group C, potentially AI that improves critical thinking skills? This relies on the fact that more critical thinking leads to more democracy. It could be that this type of AI will improve education across the board (rather than critical thinking only), which will improve STEM skills in autocratic states as well, so I could see this falling into Group B as well.

Why do you consider the US a good example of liberal democracy, what with its signifiant democratic deficiency problems?

The newspapers, the TV channels and web forums are heavily influenced by advertisers and companies, to the degree that this allows them to censor material they see as undesirable. On top of that you have a two-party system.

Then there's the fact the US is militarily aggressive-- the Iraq war is legally (i.e. UN charter-wise) very similar to the Ukraine war and the US has had secret torture prisons in Poland, Romania and some other eastern European countries.

Is that really liberal democracy for you, enough to make the US the liberal democracy?

USA may be not the most democratic state on Earth (hard to beat Switzerland), but it's undeniably more democratic and free than the current mainland China state.

Yes, I meant my comment in an absolute sense. It was not intended to be a comparison with China, just that I feel that the US is not a good example of a liberal democracy.

I like the US constitutional amendment preventing the government from making laws abridging freedom of speech though. It's, at least in part, stronger than what we have in Europe, so there are ways in which the US has very good laws democracy-wise, but I'm not sure it's quite enough.

I don't think the question of whether the US is in fact a liberal democracy is central to my argument. If you see the US as more autocratic, the same line of reasoning applies: won't AI simply tend to exacerbate all of the problems you've mentioned?

Certainly, and having reread the argument, I can't say that it's unreasonable.

Let's see

Alex is the CEO of Scale

Scale is a DOD Contractor

Scale is an AI company

Alex suggests DOD allocate funds for AI

Here's some news from someone who has seen the AI leaderboard game closely and led teams that won some

Leaderboards like COCO that Alex cites are not a good indicator. They are flooded incremental performance improvements that hardly translate to any real world application

Either he’s trying to be manipulative by using that data, or he doesn’t understand “ai”

>> (from TFA) In the past 5 years, the United States has still not started, let alone operationalized, a major AI capability that could disrupt our current warfighter.

When he misses or conveniently ignores the Skyborg UCAV program from the Air Force [0], I also question the author's knowledge of the current state of military programs.

[0] https://en.m.wikipedia.org/wiki/Skyborg

I am definitely afraid of mass data collection and drones, both for warfare and law enforcement. They lower the cost of coercing and if need be killing, both emotionally and financially.

AI will lower this cost further, and enable fewer and fewer powerful people to threaten or control the rest.

The Chinese government has so far turned much of its technology inward toward the control of its own people, who seem either OK with it or unable to resist. Perhaps they are just early adopters, and one by one all governments will follow their example, first trusting their own judgment more than the will of their constituents, and then using technology to keep it that way.

We may not need a world war 3, or even an external adversary to lose this race. How about we spend more time on defensive technologies for individuals and laws protecting their use?

I've seen a number of various voices speaking up about the possibility of a turn key totalitarian state in the US. Where everything is in place and waiting for some just to turn it on. That is the worst case scenario here.

That said obviously we can resist, but it is worrying.

This article seemingly handwaves away the single most important factor in modern war between major powers: nukes. If two major powers go to war, and it's a real war in the sense of one trying to directly conquer the other's land then that's going to, near immediately, escalate to mutual nuclear annihilation. AI doesn't change this basic and now decades old formula in any way, shape, or fashion.

I'd also add that the article fails to mention that the wargames we keep losing are not outright nation vs nation war, but instead over things like Taiwan. We're trying to win a war at a location 10,000 miles away against a country 100 miles away. Overextension is at the heart of the decline of all great Empires.

Half the kids in US grad-schools are Chinese. The other half are Indian. Neither are particularly fond of the racial-religious notions of 'US hegemony'.

Good luck keeping your 'supremacy'.

I want to live in a world where liberal democracies have the most power over the world. The power of China, Russia, Iran or North Korea should be minimized.

how liberal and democratic, where 1b westerners and their 350m american elites especially get to lord over the remaining 7b, and consume orders of magnitude more resources per person.

Thanks for clarifying that, I never realized the leaders and elites of China, Russia, Iran, and North Korea have nothing but love, generosity, and good intentions towards those 7b. /s

When you're done with the whataboutism, there are plenty of things that could be improved with the system if you choose to engage with that. Your comment history looks like you're more interested in pro-China/anti-West astroturfing, though.

Of course, left unchecked the barbarians will destroy us all! Never mind that you surround them with military outposts, are the two major oceans away from Eurasia, and have been the instigator of every conflict with them. And you have the temerity to claim the West has love, generosity, and good intentions!

I'd love to hear you explain what you think "whataboutism" means and how Western dominance over the global south is actually democratic, but something tells me continuing to engage would be pointless. It's alright, keep thinking of me as a paid shill if it comforts you.

Article is fancy money begging but has some truth in it.

From my layman understanding core defence rests on GPS and hypersonic misiles. China has their own GPS and good satelite tech. US has Starlink, space millitary branch. So that’s enough to prevent all-in war.

The Taiwan war if happens will be more boring than you think. Probably it will be chinese embargo for Taiwan and US. But that’s not profitable unless you secure other core customer (EU). Win would guarantee higher currency value against dollar but you get embargo anyway so again no profit. Also with economic issues means less export.

Drones probably will be important but US has laser tech to fight them. Can invest into better/faster image recognition.

I think communication systems could be improved. You still find that command is often poor (soldiers act autonomically but dont have enough context info - AI could help with broadcasting). Not relevant for Taiwan issue.

Cybersecurity is obvious AI target, because the point is for things to go wrong ;)

Taiwan war would need to be siege or total destruction as assault by sea is not feasible AFAIK. Siege means economic war or stupid attrition war. If chinese would be stupid enough to go for attrition war then US has advantage in the sea and air. In such case air defence would matter in Taiwan and then also logistics, that why Starship delivery is needed tech. Economic is also not that sure as US can block Singapur choke point and then you need to go around Brazil to trade by sea.

If I was Chinese I would invest into energy and food and ignore every other country (as they have most rare-earth). When it comes to computation then photonic computing could be the leap needed.

I’m also not involved in weapons development, but I’m pretty sure by hypersonic missiles, you mean ICBMs. Technically ICBMs are hypersonic and missiles, but when people say hypersonic missiles, they mean something more like a hypersonic cruise missile or glide vehicle. The idea being, they don’t follow a ballistic trajectory and can maneuver a bit, harder to shoot down, and so scarier to our carrier groups.

In the end of course, the final “core defense,” as you put it, for major powers rests on the fact that we have a bunch of nuclear ICBMs, so, a real existential war between us can’t actually end in real victory for either party (the “loser” in a nuclear war just gets to be first in line in Hell, where we all be for destroying civilization).

Anyway, the reason I bring up the distinction is that ICBMs are old tech, relatively simple, while hypersonic missiles are a recent thing, never tested in battle. Less interesting to the US I think, they are mostly for scaring away carriers, so obviously China would like them!

I always wonder how China can "dominate" AI while there still are no noteworthy Chinese compilers, programming languages, chip building tools, software collaboration tools, and more.

It is likely that "dominance" claims are inflated just like China’s GDP [1]. And for people thinking that China will in the future dominate AI, there is the problem that the demographics are unfavourable and that the foreign policy has been extremely hostile in the last 10 years [3] which partially contributed to nobody willing to collaborate on high tech. For example, China has 0 fabs with EUV machines [2] meaning that no chips below 5 nm can be produced domestically.

[1]: https://www.nber.org/digest/aug19/official-statistics-overst....

[2]: https://www.lightreading.com/semiconductorsnetwork-platforms...

[3]: https://youtu.be/y87R3Lp0jd0

There's little reason to develop compilers or programming languages when so many already exist.

China's total population will indeed not be growing, but they'll soon be getting their first third- and fourth-generation engineers. It takes time to establish an engineering culture from nothing, and I think their capability is likely to increase even if their population decreases. There'll be quite many people in China anyway.

When it comes to ML there really are very many publications by Chinese authors, and many are very good. I'd even say that some of their work is underappreciated. I can't judge as carefully when it comes to materials science, but it seems similar there.

I think Chinese industry is very different from that in Europe, the US or Russia-- jumping to something I understand, for example, they're competing on volume in things like synthetic crystals. They're selling what seems like good hydrothermal quartz in dimensions 20x30x30 mm for 56.5 USD. When I contacted Western and Russian companies about that kind of thing, they have neither an interest in selling nor gave me a price. Meanwhile, these guys are selling it on the internet for basically no money. I think they'll only get better at technology and volume is its own thing.

> There's little reason to develop compilers or programming languages when so many already exist.

Ah, so it was probably not necessary in hindsight for Apple to develop LLVM and Swift and for Google to develop Go and Jax? At the time those projects were created, there were already many compilers and programming languages.

> When it comes to ML there really are very many publications by Chinese authors, and many are very good. I'd even say that some of their work is underappreciated.

Why are the publications good?

> I think Chinese industry is very different from that in Europe, the US or Russia--

Why would that be, apart from the one anecdotal example that you give?

>Ah, so it was probably not necessary in hindsight for Apple to develop LLVM and Swift and for Google to develop Go and Jax? At the time those projects were created, there were already many compilers and programming languages.

JAX is great, but it isn't a language as such. It's a kind of generalized tensorflow. I'd instead characterize it as a framework. Go, and probably Rust are languages that were needed before their invention, similarly with Erlang. But not they exist.

Similarly, Google went for Linux when they wanted to develop Android, they didn't decide to make their own OS from scratch.

LLVM I don't know what to say about-- it's probably critical to the companies that need specialized languages for their hardware to have something with the right license and the right modularity, so it's probably an enabler for things like CUDA and there may well be a path to performance through understanding compilers and processors together etc., so compiler understanding is probably quite important, but now that it exists NVIDIA doesn't make their own, they use LLVM.

>Why are the publications good?

I've seen some solve interesting problems, some that contain new mathematics, etc.-- they're good in the same way that ML publications from the US or Germany or any other country are good. There are even some good ones that no one cares about because they're written very badly.

>Why would that be, apart from the one anecdotal example that you give?

They seem to have actual effective competition between firms, even for fun stuff like synthetic crystals and to be trying to reach any market where they can sell them. This competition means that some industrial vitamins (weird things like synthetic crystals with specialized applications, etc) aren't locked up in B2B-only-and-we-have-to-be-partners relationships and can actually be bought for reasonable cost.

If you sliced the crystal they make for jewellery applications you'll probably find that it has worse Q-values than what's made in Europe or Japan, but at least you can buy crystals to slice and fiddle with things yourself.

Can the US taxpayer tolerate (financially, politically) what it will take to get the best and brightest to do AI work for .gov and .mil instead of adtech? All you have to do is look at defense lab salaries to see there is going to be a problem keeping talent around. Some people are certainly willing to take a pay cut to work on national defense, but probably not enough. Perhaps especially not the kind of person who went into adtech in the first place.

The way to short circuit this is to make certain classes of adtech de facto (maybe even de jure) illegal.

This isn't going to happen by recruiting people to work for government labs. It'll happen by contracting the work out to companies that are already doing top AI research (many of which also work in adtech) and through grand challenges that incentivize research teams and startups to make new breakthroughs.

Google, for example, would have gladly accepted Project Maven contracts if its own employees hadn't protested. If China does supersede the U.S. militarily, that may end up looking naive in retrospect.

Ah, yes, the solution is to cede more of the country to corporations!

People who feel strongly that current procurement strategies are the wrong way to go about things should devote as much toil and treasure as they can to seeing their preferred solution materialize. In the mean time, I hope they will allow others to work within our existing system to keep our adversaries from replacing the US with something much worse for the entire world.

With government salary bands what they are, there is nothing they really can do.

I basically agree. I don't think the political mechanisms exist in the short-to-intermediate term to get salaries in government labs high enough. As you point out, the extent to which "civilians" will be willing to participate in these lines of work is an open question (said without judgement, just as an observation). Presumably folks who have opted-in to work at defense contractors will be more willing, but do they have the ability to attract a sufficient amount of talent in this area?

Paycheques are very hard to walk away from but a lot of very bright minds will go into defence right out of school and never look back.

Quick, give me money or China will destroy us. I also have only a high-school degree but I'll pretend I'm an AI expert.

I dunno. I mean, yeah technology will be used in war, and counties ahead in technology will usually be able to design better weapons. But like:

> Technologies including drones, AI-based targeting and imagery intelligence, and Javelin missiles have allowed for a shocking defense of Ukraine against Russia

What is “AI-based targeting and imagery intelligence,” is it being used in Ukraine, and why is it sandwiched between two media-darling weapon systems? This kinda reduces my faith in the honestly of the author here. Does anyone know about weapons development, is he pulling a fast one?

The underestimation of US realpolitik in this article seems off. Russian tactics in Ukraine are not US doctrine. Ref: Desert Storm for an example of how the US vs a standing army can go.

Never mind discounting aircraft carriers, which no other country fields in the way the US does: blue water navy, logistics, TRAINING (caps, because readiness is most important, look at stats between US carrier and anyone else, nearly 10x less error rate), and scale.

If we try to strong man this argument, let's say it is over for carriers. Exocet/Harpoon equivalent missiles have 100% hit rate in this scenario. Does that make it more or less likely for China to be able to take Taiwan?

I'd argue it doesn't matter because China does not have an expeditionary force. Full stop. I'd like to be wrong, but my guess is they are not trained like US. Either missiles can stop ships and Chinese landing forces are stopped by a hail of Harpoon missiles, or they can't, and US conventional warfare sinks them in the East China Sea.

I would guess China has to go for culture victory, somehow convincing Taiwan that they should come back into the fold. Taiwan does not sound like they are interested.

If we think US military is not working on AI drones, I'd offer US invented the best global network of drones with the highest body count. Turkey is doing a great job with their TB2, but that's no predator/reaper, and those are the drones we can see. It's more than one thing I think, and this article glosses over that.

You can say many bad things about the US military, it's bloated, they overspend, etc. However, they have logistics absolutely down. It is the greatest bomb FedEx ever created on Earth. They work tirelessly whether there is a peer or not. Not a bet I would take.

Must funnel ever increasing resources into wasteful and destructive conflict.

How about using AI to figure out how to cooperate at scale?

Determining ways to win a destructive conflict is a mechanism to signal willingness to cooperate, and the value of that cooperation (averting the war in the first place). Pretending they’re mutually exclusive is wrong.

One problem with human governance is we can't seem to figure out how to coordinate without burning excessive amounts of energy signalling status and power at each other. What if we could do better?

What exactly do you imagine as “doing better”, here?

If you never lift weights, and everyone else around you does, don't be surprised when the "cooperation" tilts in their favor. Such is the nature of man.

Is it wise to compare spending within defense budgets between these countries? Much AI research in the US is conducted by private firms. How does the current defense budget compare to historical defense budgets, such as during the early days of the digital computer in the Cold War?

On a related note: yes, Chinese researchers have models that perform certain tasks well. But are those models useful in the contexts the author mentions?

To be sure, I don't disagree that AI research needs funded. I'm just genuinely curious about these points.

Says the guy who sold products to the Chinese company (Baidu) and tried to sell more to china in the early days. Such hypocrisy. http://web.archive.org/web/20190805191959/https://scale.com/

It's funny that OP doesn't realize "the AI war" will be between the AI and all of humanity.

It won't. It'll be a war between the people and entrenched power structures who will use very primitive AIs to crystalize their position at the expense of everyone else. What rich man or government would ever fear riot or revolution when you can use data mining algorithms and precision drones to liquidate even the most minor whiff of dissent at the push of a button? They will eat everyone else with specialized drones as sapient as insects before truly human AI is a thing.

Not super clear to me how useful computer vision algorithms are to the military. The important thing in the private sector is replacing (say) the lab tech who would otherwise be paid minimum wage to look through your data. If you have plenty of low skill labor available, what advantage do you get?

Outright speed, speed of scalability.

If you can submit what you want and have it within moments, it beats a fancy typing pool. That kind of speed alone lets you have a highly trained someone retry different variations and filter more on quality of result.

Scalability. If you need to do it ten times faster. More hardware = done. More people could make it done too but training etc is required.

I’m still confused by ai fear mongering. All the best models and advances come from the USA. Chinese research, while high in volume, never advances state of the art aside from incremental improvement.

Incoming downvotes, but the fact remains most advanced have basically come from google, Fb and openai

Simple, Scale AI is likely fishing for federal contracts now that its traditional enterprise customer set is throttling back on spending

Exactly this. The "AI race" is yet another nuclear arms race, where the American public is terrified into believing in some sort of (nonexistent) capabilities gap, and the only way to avoid utter defeat is to funnel billions, eventually trillions, of dollars to the right companies.

https://www.youtube.com/watch?v=ybSzoLCCX-Y the real question is do we still have a mineshaft gap? hope we figure out how to get along in our only home.

That's simply not true. There are numerous problem domains in AI where Chinese research groups are advancing the SOTA, or at least are well-presented at the forefront. Your comment reads like little more than casual racism to me.

Such as?

No I read the academic research all the time. Most state of the art papers take an existing method and incrementally change the model to get a better result. They aren’t real advances. Anyone that actually publishes and works in the field knows this. Basically all major advances have come from Fb/google

This is such a brazen deflection I had to create an account to point it out. Racism, seriously? No, papers that advance the SOTA, and more importantly introduce powerful novel concepts and techniques, have plenty of ethnic Chinese authors. The Chinese are very numerous in the field, just like in software in general, unsurprisingly.

It's just those authors overwhelmingly work in the United States, and rarely in other allied countries. Perhaps the highest-cited ML paper that can be attributed to Mainland China is still affiliated with Microsoft [1].

The fearmongering is miscalibrated, whether it be about PRC-developed AGI or about Chinese people stealing our tech. To a great extent, they are building it.

Native Chinese development, on the other hand, is exactly like OP says, incremental and constrained to internal-policing-related niches like face recognition (actually that's pretty much it). They also opensource pretty powerful models [2], [3], so that's nice of them.

But nothing seriously game-breaking.

1. https://arxiv.org/abs/1506.01497

2. https://huggingface.co/spaces/THUDM/GLM-130B

3. https://huggingface.co/BAAI/EVA

Although this seems true at first glance, it is likely that state sponsored cutting edge Chinese AI research would not be disseminated publicly. China does also have certain 'advantages' in the ethics and scale of training data collection.

In 2021, Alex Wang of Scale (the author of the linked article) uncritically quoted the following from Eric Schmidt:

"Eric states that the global growth of AI technology is a national security threat for the United States. In March, we believed we were ahead of China’s technological development in the AI sector, but they quickly showed us that we were wrong. By June, they’d demonstrated a model capable of producing human-like text that was comparable to OpenAI’s GPT-3 model in quality."

https://scale.com/blog/Eric-Schmidt-Geopolitics-TransformX-S...

As such, it's important to note that the author views the entire Chinese AI industry as a threat to the United States, and may believe that the growth of AI anywhere in the rest of the world is a threat to the United States.

Is there a reason that I should think of Wang as uniquely qualified to comment on this space? He runs an AI startup, seems knowledgeable about AI in general, but what does he know about war?

Fair question, but I don't think one needs to be very skilled in the arts of war and military tactics to foresee that AI will be important in the future of war. I don't think it's worth throwing our dear Wang's blog out the window for this. The points he makes about AI's involvement in the future of war are very broad and high level. For example, he compares the amount of money that each country is spending on AI in the military and concludes that the US is lagging. It would be interesting if we had someone, like Palmer Luckey, who is actually involved in military AI speak in more detail about it.

Right, they are high level points, which is why I am having a hard time finding something specific to critique or even disagree with.

For example, when I think of the question of “how will AI be used” the obvious answer, which Wang more or less addresses, is targeting systems. Basically, we’re just going to automate all of that and instead of a bunch of captains in the S2 shop cooking up target packets and deciding which one makes the most sense, the AI just does all of this for us.

But what does this actually mean? Can we obsolete some/most/all of the S2 shop and replace it with a container of servers stacked with GPUs to do imagine processing? Take that manpower and add another infantry platoon? A drone operations platoon?

A great theoretician would pose the question this way. Can what’s playing the government make it to level two?

The prudent thing is to invest in the technology even if china isnt.

A strange game. The only winning move is not to play.

Uses the term "very clear" a lot for something with such hand wavy reasoning. Probably not a coincidence.

The future of AI is not US vs. China. The future of AI is to empower a minority to control the majority.

I've been skeptical of AI posing an existential risk.

This article may have changed my mind.

The AI war will be won by the faction having the best datasets, and China has no limits when it comes to collecting data.

The Chinese government is all-knowing and all-seeing. They are also above the law and can do whatever they want.

> China has no limits when it comes to collecting data

This is an overgeneralization when it comes to the sources and types of data used in a battlefield, to map infrastructure, and so forth.

While the Great Firewall and social monitoring do collect much data on the population, thereby decreasing non-state internal intelligence gathering and capability for spying, there is a difference between tactical data from a hot front and identification of spies.

It feels like China is always overstated by the west

Summary: China bad, because.... AI

With all due respect, has the author ever watched WarGames (1983)?

43 days ago I asked this question:

https://news.ycombinator.com/item?id=33215740

This question got voted down everywhere I discussed it on Hacker News, but somehow the submission itself got 11 points. Apparently, it is an offensive question to some silent, pro-censorship part of the population here. Today I'll ask it again, and I hope somebody will respond in a useful way:

QUESTION:

I write software for parallel processors at a hardware/software company you've heard of. I am located in California but many of my coworkers are located in China (roughly half the team).

We are directly affected by the American government's severe new CPU/GPU export restrictions.

It seems to me that America is preparing for a period of cold war, or worse.

It's time to start thinking about contributing to the American war effort by writing high-performance military computing systems.

For example, SIMD particle filters for hypersonic weapons, or low-latency convolutional neural networks for battlefield devices.

So, Hacker News: What company is the best place to do this work? What team?

Does anyone here already work in this field?

If you can tolerate a .gov email address, there are high-performance computing, edge computing, and AI groups at DOE labs. My impression is that they're always starved for talent in these areas due to stiff competition from industry and lifestyle restrictions that come with a security clearance (being a US citizen, reduced international travel, no drugs other than alcohol and nicotine, your work may require on-site). Defense labs are probably similar.

AI is strong enough to decide who wins the economic fight. Prediction: 10% gdp growth year over year as we approahc general ai.

The AI Economy will be a new reality. It may not be a good reality. Will it be your reality?

Reply on news.ycombinator.com