Deadly and Imminent; The Pentagon's Mad Dash for Silicon Valley's AI Weapons | Hacker News Reader

Deadly and Imminent; The Pentagon's Mad Dash for Silicon Valley's AI Weapons

citizen.org

79 pointscempaka68 comments

The Replicator program was supposed to result in cheaper drones using commercial-grade technology, but the price isn't coming down to Ukraine/Russia levels and it's taking too long.[1]

If "AI" got into the Replicator program, someone is gold-plating it. It's supposed to be about US$500 to US$1000 drones. What's coming out are US$20,000 drones.

[1] https://www.economist.com/united-states/2025/01/05/the-us-ar...

Between single use aluminum/lithium-air batteries, a proper sensor suite, and the performance of that F1 follow drone I think a proper $20,000 drone would be much more useful for the US assuming the costs are under control. They could conceivably get significantly better range than artillery and use AI (hand waving a lot here) for target selection to make up for the smaller warhead and 4x cost.

With the right combination of hardware and software, drones could become an incredible general purpose weapons including anti-tank, loitering and mine-like anti-personnel, RF missile, and so on.

$500-1000 drones might be useful in saturation attacks but most of that cost will be eaten up by the warhead. The drones won’t be very useful unless they’re hunter seeker style with small charges and a lot of intelligence (just like the $20k drone) for target selection. I don’t think US doctrine can support them otherwise (the vehicle to deploy them will cost 1000-10000x anyway). These classes of drones work much better for a country like Ukraine in a total war on home soil.

>might be useful in saturation attacks

with deployment of cheap point defense systems - like cheap radar (Starlink phased array is $500 or so) plus shotgun (Russians currently do "soldier with shotgun") - on every tank, IFV, truck, etc. only stealth (not an option in those price ranges really) or saturation will be the way to go. Single $20K drone would be just wasted by such point defense. To the point defense add automated predator/interceptor drones patrolling against the attack drones. Thus saturation with large autonomous hives with drones of various types - coordinators, predators, and actual attack drones.

Those point defenses might work against small slow moving $500-1000 drones, but not one moving 200+ mph with an evasive trajectory (hence why I mentioned the recent F1 drone). Add a small rocket motor to the $20k drone for final approach and it’d probably top out at 400-500mph and it can use propellers to “jump around” randomly in its flight path so targeting it is nigh impossible. That would require something closer to a proper missile defense system or CIWS and tanks/APCs/etc would not have those anyway.

The whole problem with cheap drone swarms is that their range is shit and whatever you use to deploy them will just inherit the existing weaknesses - whether that’s cost, logistics, or payload.

I think saturation attacks are only going to be useful against static targets with large numbers of squishy meatbag targets as a cheaper way of taking out soldiers as measured in $/life, where a (maybe very expensive) plane drops a payload of small hunter seeker drones that target every human-looking IR source in the vicinity. I think that could be a lot cheaper per kill than dropping hundred pound bombs.

>not one moving 200+ mph with an evasive trajectory

that is slow. Well under RPG speeds that current active defense systems (like Trophy) easily catch. And evasion on such a slow speed isn't a real evasion - at best 1m shift in 0.01 seconds - even cheap radar would easily track and would take it out with shotgun or machine gun.

> proper missile defense system or CIWS

that resolves, tracks and targets 30 cm diameter missile moving at transsonic speed at 2-4km. The point defense i'm talking about would need to resolve and target a 30 cm target at 10-20m distance (the tank or APC can take a drone warhead explosion at 5m distance) which is orders or magnitude cheaper/easier (you can even put a sub-$100 car radar or LIDAR onto a rotating plate similar like on Waymo car and whenever it picks something moving toward you at a distance less than 20m - just shoot in that general direction. Yes, infantry hates being around tanks with active defense in combat as shrapnel from the ADS hits the infantry, frequent story with USSR army back then in Afghanistan for example)

>Add a small rocket motor to the $20k drone for final approach and it’d probably top out at 400-500mph and it can use propellers to “jump around” randomly in its flight path so targeting it is nigh impossible.

Now you're trying to invent a jumping around RPG or Javelin missile. I agree that such development will happen too (though "jumping" at 200m/s is petty hard), yet it will not be able to always jump out of the shotgun grape shot cloud at that 10m distance. So, yes, there will be race between attack and defense, and my point is that single drone approach will be left behind pretty much from the start.

> cheap drone swarms is that their range is shit

50-80km communication and flying range is what we have already today. Communication - while staying cheap - can be extended further by retranslator drones (and for autonomous/AI drones the communication is much less of an issue - for example no need for high quality video, if at all), and electric drone-plane (not quad) can do up to 200km even today.

>I think saturation attacks are only going to be useful against static targets

I see quite opposite. Modern air-defense shoots down all single targets with high probability (except for stealth, which like $4M for one Storm Shadow stealth cruise missile). On the other hand - Pantsir-S1 can target simultaneously only 4 targets. A battery of S-300/400 can shoot only 12 missiles before reloading. Trophy and the likes easily shoots down one incoming targets and takes half-a-second or similar to get ready for the next target. And improving the simultaneous capabilities is costly. Where is saturation (horizontal scaling :) is an easy and cheap way to overwhelm/defeat such systems.

Pretty much this - the only way the entire premise works is if you have a huge number of cheap commodity hardware that's going for the 0.1% of a 99.9% succes rate defense system.

Defense companies won't like it, because commodity is synonymous with "low margin", so it's not surprising you end up with something that's two orders of mag higher price point and a buzzworld salad ("AI enhanced" because we have added some ML-based damping to a kalman filter to charge more money) to go with it

A 99.9% successful defense system is unlikely. CWIS is nowhere near that good. Overwhelming a defense with an attack from multiple directions has worked. Read up on the sinking of the Moskva[1] and Russia's current efforts at drone defense.[2]

[1] https://en.wikipedia.org/wiki/Sinking_of_the_Moskva

[2] https://www.forbes.com/sites/vikrammittal/2025/01/24/russia-...

Ukraine is currently manages on some days to send 100-200 drones into several regions in Russia. The drones are slow and small, like with 50kg warheads. Most of these drones are shot down (on various video it seems usually by smaller missiles, like Pantsir's), though a few usually do hit their targets. I think Ukraine is, while slowly, still progressively going along the direction and would come to the strategy similar like what i suggested https://news.ycombinator.com/item?id=42529638

> These classes of drones work much better for a country like Ukraine in a total war on home soil.

Who's to say that's not what Replicator is aiming for? Sans the "foreign enemy" part, perhaps

Sometimes I think it would be so easy to make a disruptive defense company that undercuts the traditional ones massively. Sure, you have Palantir and Anduril, but there's room for so much more.

The "traditional" defense companies typically hire rest-and-vest talent, rewarding mediocrity over actual innovation. I have had many bright peers leave the defense industry for big tech and HFT, because the defense industry's backwards culture is very much about hammering down the nail sticking out and "sticking to what works" (outdated technologies, scamming the government for 20x the production cost.)

It would almost be too easy to hire some SV talent and make way better products for a tenth of the price. I guess my only qualm is that I don't want to make literal weapons.

This is the right answer. Further, its not an open market - every OEM with scale demands a closed stack that isn’t interoperable with other computing systems by default.

It’s specifically because there is no domestic robot production capacity that can meet the requirements. I’ve been buying mostly Chinese robots cause nothing domestic can compete.

Neither BD nor pickle nor cobot etc can actually do what unitree and DJI have done at scale - it’s just a fact.

The root of this is that PE and Venture aren’t investing in US manufacturing - full stop. So they wait around for the govt to invest in infrastructure and then follow on sheepishly.

I’m not seeing the same kind of “private” investment in massive robotic projects so until that happens we’re going to lag on SOTA with US robots

Defense contractors used to be the stodgy old companies that most in the tech industry looked down on. It's been interesting to watch the slow shift in the valley towards widespread acceptance and even wholehearted support for the military industrial complex.

I think quite a bit of work is done in the old guards lands around DC still. Offices are still being built due to SCIF requirements obviating any real work from home movement. Datacenters are going up alongside them. It is a bizzaro world in Northern Virginia and other such parts around DC. A spotless 1990s idea of a suburban utopia surrounded by data centers simulating nuclear armageddon and defense contractors offices designing that armageddon for a profit in the meanwhile, thoroughly intermixed with the strip malls and identical tract housing and neighborhood trails full of joggers and strollers.

Its not support. They quite literally are the military industrial complex

Silicon Valley's defense industry was the *first* tech sector there, and (arguably) the primary reason the rest of tech grew up in that specific place, as an offshoot of it.

https://steveblank.com/secret-history/ ("The Secret History of Silicon Valley")

https://hn.algolia.com/?query=%22Secret%22%20History%20of%20...

Defense companies are documentation-driven and procedure-driven ("stodgy") because they are government contractors, and there have actually always been a lot of them in SV. There is also a precedent in the tech company idea as a sales and marketing tool aimed at the government - in Theranos of all examples.

> they are government contractors

It's not just a kind of subculture though, there are practical reasons: "Defense" products have often meant large batches of durable complicated equipment (e.g. tanks, planes) where your clients plan to have them ready for service for the next X decades and really do require their own copies of good repair-manuals that a non-employee can use, as well as some confidence that the supply-chain is both secure now and can also continue to exist for those X decades.

Even if all you're doing is providing the software for the Big Mech Suit, you can't dodge those requirements since it's part of the same set of product-needs.

This is a stark contrast to, say, an "Uber for Cats on the Blockchain with AI" SaaS site the lives on ad-revenue and no end-user support.

They're not so great at providing the repair manuals this century, in fact a lot of the contracts say that only the contractor can repair the equipment. A lot of helicopter trips are made to aircraft carriers as a result.

It's pretty clear that we need not just a law but an actual constitutional amendment requiring them to go back to the old way.

> It's been interesting to watch the slow shift in the valley towards widespread acceptance and even wholehearted support for the military industrial complex.

Damn and here I thought the valley (and valley culture) was the beating heart of the MIC this entire time.

There was a period where a large enough number of people had serious objections to the entire idea that having Raytheon or LLNL on your resume wasn't much different than having a porn or crypto company on it today. Executives would go to hippie spiritual retreats like Esalen and pay lip service to the idea of positive social transformation, invite random spiritual leaders to conferences, and monologue about their deep commitments to hippie ideals.

Much of that was BS of course, but I'm just commenting on the shift.

They are still those stodgy old companies and SV still looks down on them for good reason. If you work at one you'll quickly realize how hard it is to get anything done because no one gives a fuck at those companies. It's pure rest and vest.

Now the SV-type defense companies - those are a whole different ball games.

Ukraine probably made a difference. It's been a while since there's been that sort of war where the western 'defence' industry is actually defending against an attacker as opposed to say bombing Iraq.

I fear it will only accelerate as all the other hype+money trains continue to peter out.

So I was just thinking, and hear me out, since the budget is all coming from us why can't it be spent on something we like or want?

Technically it is. If you don't like what the money is spent on, "write to your representative"[0] But unless you can start a movement that gets large groups of people to agree about the same things, probably won't get too far. Also, vote. Or better yet, run for your areas house seat and make changes from the inside, if you have the patience

[0] https://www.house.gov/representatives

While pretty much everybody would say no to this if you explained it to them, between nobody wanting to spend a lot of time keeping track of everything that's going on, and the stewards of res publica not exactly feeling excited about the prospects of closer attention, the connection between the legitimate will of the public and the actions at the other end of the chain of command can get pretty tenuous.

> the connection between the legitimate will of the public and the actions at the other end of the chain of command can get pretty tenuous.

This is exactly why I can't imagine an open internet was ever going to survive.

For the same reason that sheep don't decide what kind of sweater gets made from their wool.

I don't think that's a productive attitude. If you were someone's manager, and they had learned that they could get away with a lot under you, there's only so much time you could spend complaining about human nature. Really, it was your responsibility from the beginning. We're like those restaurant owners from Kitchen Nightmares who show up once a week to talk to our own bartender and don't know that the walk-in fridge is 75 degrees F and that there's raw chicken on the ground.

> We're like those restaurant owners from Kitchen Nightmares who show up once a week to talk to our own bartender and don't know that the walk-in fridge is 75 degrees F and that there's raw chicken on the ground.

Who's "we" in this sentence? Plenty of us have known about the "raw chicken on the ground" for years and actively try to do something about it. A better question for this discussion is why have you been looking away for so long, and what are you going to do now that you see the problem?

I don't know the answer to that question generally, but that's the only sense that is productive.

Well, that's sort of the thing you're missing... the vast majority of people are not the "manager" or "owner" in this situation. But if they were to take responsibility, it would mean they'd have to replace the current managers who are. So why doesn't that happen? Simply because, and this may be shocking, but some people don't want that to happen, and rather likes how the current money is being spent, and likes the current managers. The core of politics (and its problems) is a matter of disagreements, not incompetence.

We already have the responsibility for letting this stuff go on, that's the problem. I am also very skeptical that anywhere near a majority would support it if they were aware of it and understood.

Even if the majority of people are in agreement of "we dislike the current people running the government", you still have the problem of agreeing on what to do about it. Some people may want to overthrow current government norms and democracy. Other people don't ever want that, and want to uphold formal structures as is. Some people want to replace leader A with leader B, but others want leader C. Some people want abortion rights, others don't want others to even have that right. Some think topic X is a priority issue, whereas Y is a non-issue, and on and on...

All said, politics is always ultimately a matter of disagreement. But you're right in that, if we only could be in complete agreement, things would be resolved rather quickly as there would be no opposition.

Making it about the people is how you get disagreement out of agreement. :-)

Who says we're in agreement?

The current model of “democracy” in the US —and indeed in most countries— is very arm’s length.

In countries with a model closer to direct democracy, Switzerland being the best example I know of, it might be possible to hold a vote about a specific policy (e.g. we don’t want our army to develop/use autonomous AI hunter-killer drones) but micromanagement of multiple policy decisions is probably both difficult and maybe unwise. Defence topics might be especially challenging given the background knowledge required to make a rational decision may often be classified and/or otherwise unavailable to the general public.

You'd better be happy with the choices you're offered or you're actually a fascist, whether you knew it already or not.

How much do you like or want not being at war? Because si vis pacem, para bellum.

I can’t wait for 2040 when these weapons make their way into the hands of the police in the US through the surplus programs that currently get them MRAPs.

The U.S. does not have a choice not to develop advanced drone technology. It is an asymmetric weapon in terms of costs to damage dealt. The enemies already have it. It's likely that the only counter to a drone swarm attack is a drone swarm defense.

That's really a superficial issue here, because once the program budget has grown to fighter jet levels the asymmetric cost discussion is over and all that's left is taking a potentially culpable human being out of the equation, and the likely more influential aspect of military purchasers wanting to be impressed by vendors just like any other purchasers anywhere else.

Drones don't need to be anywhere near the size of fighter jets to carry the same payload. But meanwhile a single fighter jet is going to do nothing against a swarm of 100 drones that are 1/10th the size and 1/100th the cost since there's no need to train a pilot for 10 years and they can forego many safety systems designed for human pilots.

Then there's a whole range of medium to low speed drones which again just overwhelm an infantry force. Or attach C4 and fly them kamikaze into ships. How do you guarantee you can shoot down 1000 drones in time when only one of them needs to succeed to do massive damage.

> How do you guarantee you can shoot down 1000 drones in time when only one of them needs to succeed to do massive damage.

A million rounds a minute takes out a drone swarm pretty well, a package of C4 alone will deflect off warship grade armour ...

Drones and drone swarms are an issue, sure, but less so in some scenarios.

* https://www.youtube.com/watch?v=t7faoyoxMF8

* https://www.youtube.com/watch?v=wKlnMwuCZso

> A million rounds a minute

Dragonfire is quicker, cheaper per kill, and safer in urban places then spray and pray kinetic weapons.

https://en.wikipedia.org/wiki/Dragonfire_(weapon)

There are multiple approaches.

How fast can Dragonfire recharge, how quickly can it target and kill 1,000 drones?

> spray and pray kinetic weapons.

You understand, I hope, that targets can be tracked, and ballistic weapons can accurately place a cloud of shrapnel on a per drone basis.

With the same kind of track and target platforms that would direct a laser ..

> the same kind of track and target platforms

Speed of light is the game changer

Pfft. - Not really your field, is it?

How about total cost, weapon size and footprint, time to charge ..

Speaking of time to charge, you dodged the real question .. how many drones can be targeted and destroyed in (say) 30 seconds?

> Pfft. - Not really your field, is it?

You got me. I know nothing about this and you would definitely be better off asking someone else.

I wonder how easily highly reflective coatings will mitigate Dragonfire.

Very well indeed I think. AFAIK Dragonfire is fully within the optical spectrum. However, there's a similar project we've been working on over here that's a beamforming MASER - that travels right through the housing and fries the electronics.

A fifth generation fighter jet can bomb the factory that makes them, the truck that carries them, jam their coms, and land in a parking lot to rearm, refuel and do it all again.

I'm sorry, but playing king of the hill is a great way to get people to try to kill you. Absolutely moronic foreign policy.

It's appalling how many people on these threads take pride on being an SME on these topics, without recognizing the real point. It is not a question for a drones SME. It's question for defining the meaning and purpose of human progress. We are creating a form of natural calamities and calling it progress. Drone swarms are calamities similar to earth quakes, tsunami or wild fire. If you have a swarm your enemy has it too. If you have a nuke, the enemy has it too. You might reason, well, even if we don't push for it, others would. Alright, then the only way to stop it is by natural balance. The balance would be achieved by devaluing the resources humans are after. Say there are too few humans remaining on earth, so that they don't care about owning stuff, country borders or money or security ...

ukraines fpv drones supply a mountain of training data..

Imagine a planetwide human party in resistance to self extinction , that would be the worst terrorist organisation ever.

I don't think it's a conspiracy theory to say: The Stargate Project will be the largest surveillance tool ever, it will be pointed squarely at the US population, and it will be abused in historically unprecedented ways.

It is not. Genuinely the "best" use of AI, and why it was made by ad companies (surveillance companies)

Reply on news.ycombinator.com