Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Large language models in national security applications

arxiv.org

Post reply on HN
Print view View original

41–50 of 60 posts

Re: Large language models in national security applications

#41

Post by hatthew »

recursive wrote: post #32
exe34 wrote: post #29

Earlier quoted context omitted.

charge one penny for every post. most people can afford it. bots become less cost effective and you'd be able to trace the source of funds.

Even if they could afford it, they won't. The UX friction of charging money would send engagement off a cliff, even if the nominal charge was $0.

Only half joking, I feel like the majority of comments on the internet are garbage and not worth the time to read, and increasing friction with a nominal charge isn't necessarily a terrible thing.

Re: Large language models in national security applications

#42

Post by davisp »

joe_the_user wrote: post #15
perihelions wrote: post #5

The most obvious way the US national security industry could use LLM's right now is simply to spam foreign adversaries with chatbots. That's their greatest strength right now—a use-case they have amply proven themselves for. This paper comes off as eager to avoid this topic: they (briefly) talk about detecting foreign LLM spam, which is called propaganda, but sidestep the idea of our own side using it. If we were con…

Oh, national security professionals aren't going to be talking about psyops, offensive applications and etc, because such things make a given state look bad - they're an offense against democracy and respect for facts and they make the overt media of a given nation look bad. But hey, leave it to HN commentators to root for taking the gloves off. Not to worry post, I'd bet dollars to donuts the actual classified discu…

[flagged]

Re: Large language models in national security applications

#43

Post by Jerrrrrrry »

joe_the_user wrote: post #19
Jerrrrrrry wrote: post #13

Earlier quoted context omitted.

Sure you do, it's accumulated heuristics, no different than meteorology, or other macro-sims of chaotic systems. The difference is that human lives are intentioned for different fates; so the negative cognitive dissonance is going to persist consciously, then sub-consciously.

it's accumulated heuristics, no different than meteorology Meteorology is based on physics, meteorology doesn't have a hostile agent attempt counter prediction attempts, meteorology doesn't involve a constantly changing technological landscape, meteorology has access to vast amounts data whereas data that's key to military decisions is generally scarce - you know the phrase "fog of war"? I mean, LLMs in fact, don't p…

  >Meteorology is based on physics,
Good start.

  >meteorology doesn't have a hostile agent attempt counter prediction attempts
All chaotic systems have second-order reinforcing feedback loops by "tautological" definition of what a feedback loop is and what a sufficiently-complex system is.

  >meteorology doesn't involve a constantly changing technological landscape,
it does, it just has little incentive/motive/order to change, because knowing the probability of it going to rain vs. the cost of it changing the weather is so uneconomically comically juxtapositioned that technology wasn't advanced enough make any progress.

All phones being 5G'd into the weather apparatus would give certainty, but the pressure to do that collectively is not there.

   >meteorology has access to vast amounts data whereas data that's key to military decisions is generally scarce - you know the phrase "fog of war"?
"Fog of war" is how it is perceived from outside. Inside, it is those who have a monopoly on viol/power/surv that ultimately wins.

and that is us.

Re: Large language models in national security applications

#44

Post by recursive »

hatthew wrote: post #41
recursive wrote: post #32

Earlier quoted context omitted.

Even if they could afford it, they won't. The UX friction of charging money would send engagement off a cliff, even if the nominal charge was $0.

Only half joking, I feel like the majority of comments on the internet are garbage and not worth the time to read, and increasing friction with a nominal charge isn't necessarily a terrible thing.

It's a good thing for most people, just not the "platforms" trying to show growth.

Re: Large language models in national security applications

#45

Post by Sabinus »

idle_zealot wrote: post #37
bilbo0s wrote: post #16

Earlier quoted context omitted.

Or, to rephrase this as a question: Is it ethical to spam another democracy with millions of chatbots pretending to be citizens of that country—if the effect is to manipulate those citizens to [take any action advantageous to US national interest]...? Just, Devil's Advocate, but ethical or not, that's what we should be doing and what we are doing. Every nation has its sock puppets out there, our job is to stop everyo…

I don't think you understand the Devil's Advocate rhetorical device. It's not just a thing you say before you state a morally reprehensible opinion you hold.

Imo there's a difference between propaganda promoting functioning democracy and propaganda promoting conservative-at-best authoritarianism.

Re: Large language models in national security applications

#46

Post by jingyibo123 »

perihelions wrote: post #5

The most obvious way the US national security industry could use LLM's right now is simply to spam foreign adversaries with chatbots. That's their greatest strength right now—a use-case they have amply proven themselves for. This paper comes off as eager to avoid this topic: they (briefly) talk about detecting foreign LLM spam, which is called propaganda, but sidestep the idea of our own side using it. If we were con…

I'm pretty sure that before the era of LLM, hired personnel and computer chatbots already exists on major social media platform to serve as "opinion influencers", either for political or commercial purposes. Just that chatbot algorithms are kid of naive, and have been "fished" on several occasions....

The usage of LLM is too obvious, and way cheaper than hired personnel, so it's like "prism". They don't want us to talk about it because it's been going on for so long...

Re: Large language models in national security applications

#47

Post by immibis »

jingyibo123 wrote: post #46
perihelions wrote: post #5

The most obvious way the US national security industry could use LLM's right now is simply to spam foreign adversaries with chatbots. That's their greatest strength right now—a use-case they have amply proven themselves for. This paper comes off as eager to avoid this topic: they (briefly) talk about detecting foreign LLM spam, which is called propaganda, but sidestep the idea of our own side using it. If we were con…

I'm pretty sure that before the era of LLM, hired personnel and computer chatbots already exists on major social media platform to serve as "opinion influencers", either for political or commercial purposes. Just that chatbot algorithms are kid of naive, and have been "fished" on several occasions.... The usage of LLM is too obvious, and way cheaper than hired personnel, so it's like "prism". They don't want us to ta…

They have, and it scales orders of magnitude better with LLMs.

Re: Large language models in national security applications

#48

Post by immibis »

itishappy wrote: post #40
bilbo0s wrote: post #16

Earlier quoted context omitted.

Or, to rephrase this as a question: Is it ethical to spam another democracy with millions of chatbots pretending to be citizens of that country—if the effect is to manipulate those citizens to [take any action advantageous to US national interest]...? Just, Devil's Advocate, but ethical or not, that's what we should be doing and what we are doing. Every nation has its sock puppets out there, our job is to stop everyo…

Why stop there? Send missiles! Heck, send bioweapons and dirty bombs. It might not be ethical but if others doing it we might as well join 'em. (This is sarcasm. I don't want to live in this world.)

We do.

Re: Large language models in national security applications

#49

Post by immibis »

photochemsyn wrote: post #23

It's rather conspicuous that the most well-known use of AI systems in warfare at present, the Lavender / Gospel / Where's Daddy systems used by the IDF, don't get any mention. It's true that LLMs are not the central component of these systems, which have much more in common with Google's targeted ad serving algorithms, in the broader category of machine learning, but a no-code LLM interface is a likely component. In…

We're not allowed to talk about Israel.

Re: Large language models in national security applications

#50

Post by PittleyDunkin »

perihelions wrote: post #5

The most obvious way the US national security industry could use LLM's right now is simply to spam foreign adversaries with chatbots. That's their greatest strength right now—a use-case they have amply proven themselves for. This paper comes off as eager to avoid this topic: they (briefly) talk about detecting foreign LLM spam, which is called propaganda, but sidestep the idea of our own side using it. If we were con…

I'm sure they're already used against both foreign and domestic populations.
Post reply on HN