AI Governance and the National Security State | Dean W. Ball

And the framework through which he approaches AI, what intelligence actually is, the nature of large language models, how they learn, what they understand, and their advantages and limitations. And we examine the broader historical thesis that animates Dean's worldview, that We are not witnessing the birth of something entirely new, but rather living through a computing revolution that began with the transistor.

It is now reaching its natural culmination in the era of machine intelligence. In the second hour, we discuss what's at stake in this transition: the model of AI governance that Dean believes will be most effective given the nature of this technology and the incentives. Driving its adoption, and what's at stake for society, the nation state, and the individual if we get it wrong. We begin with Dean's own conception of superintelligence, not as a singular all knowing entity.

But as something will derive much of its power from being embedded in the infrastructure of human civilization. We discussed where he falls along the continuum from Doomer to Accelerationist, and what a sensible approach to AI governance actually looks like when that avoids both laissez-faire abdication and heavy-handed regulation.

We get into the real-world test case of Anthropics dispute with the Department of War, what it reveals about the tensions between Frontier AI labs and the national security state, and how Dean thinks about the challenge of forging public-private governance structures adequately.

adapted for the age of AI. We close by examining labor market disruption, the overproduction of elites, and the question of which nations, societies, and individuals are best positioned to navigate the transition ahead. If you want access to all of this conversation, go to hidden forces.io slash.

And join our premium feed, which you can listen to on your mobile device using your favorite podcast app, just like you're listening right now. If you want to join in on the conversation and become a member of the Hidden Forces Genius community, which includes QA calls with guests. Discounted access to third-party research and analysis, and in-person events like our intimate dinners and weekend retreats. You can also do that.

To get access to the kinds of information and insights that are only available there and to sort of occupy that perspective. And that that was actually just gonna be an essential part of my future in this field. That intuition kept occurring to me, but I did I was, you know, doing a bunch of stuff running around and and didn't really have the opportunity to act on that intuition. And then happily OpenAI approached me and said,

You know, what do you think about sort of coming in and helping to shape our frontier policy? I think from OpenAI's perspective, where this all comes from. is really just this idea that like, you know, they have like a public policy team, a government affairs team that does exactly what you would think a large company's government affairs team does.

And they're great. They're extremely competent, but they are in this posture of reacting to policies being promulgated by fifty states, by the federal government, and Congress and the executive branch, court cases, you know, the European Union. countries all over the world, there's just a lot of incoming that that team has to deal with. And so what OpenAI kind of wants to create is like a a small organization that doesn't have to react to any of that stuff.

And instead can look out on the horizon six to twelve months. And kind of think, okay, where are things going? What are people going to be talking about? And how can we sort of proactively develop ideas? to inform that conversation. So it's a somewhat more proactive version of a sort of like policy operation.

and uses the latest AI terminology. So I don't know if she's the right person to use as the kind of baseline normie, but if we were looking for the baseline normie mom equivalent, how would you explain to her what you do?

And I you know, I wanna write stuff that's palpable for her. I think what I would basically say is that I'm fundamentally I would characterize myself as a writer and a communicator who tries to communicate ideas about about artificial intelligence, emerging technologies, and how our society can internalize those things in such a way that we We benefit from the technology and transform our world in a healthy way. I think that's basically what I would.

But it's also well done and thank God not AI slop informed. So What would you say or what do you feel like qualifies you or has given you the intellectual tools that others may lack in order to be a constructive voice in this debate?

And also to describe differences between things that a lot of people would call similar or call the same.

And like I've just like always like had this inclination to sort of look at the world with extremely curious eyes and sort of like find everything interesting. That's kind of why the substack is called hyperdimensional in a way. And so like I think what gives me an ability to do is I can often write like I can bring like metaphors and analogies and examples to mind that are surprising, but that are also illuminating in ways that I think other people

aren't able to do as much. And that's basically like the one thing that I would say like actually care. I don't even think I'm that good of a writer. You know, I think I'm like a perfectly workmanlike writer. But yeah, I think that's like the one thing I can do.

cyber offensive risks from AI models or biorisk or things like this. These are like problems that can be made sort of shaped into tractable and analyzable problems. But then there's this other category, which is like I try to remind myself almost every day that like this is 1750, right? We're in the equivalent of 1750 or 1775.

And we have an entire industrial revolution to get through. Right. And so, like, from the vantage point of 1775, it's just like, you know, trying to occupy that mindset and then imagine like You know, how are we going to govern the world of nineteen twenty-five? It's like, well, the world it's so it's so the transformations of the human experience were so dramatic that it's Like the person of 1775 could scarcely comprehend even what the problems would be, much less the solutions.

And so, I mean, imagine trying to explain what AI psychosis is. to someone from 1850 or something, right? And so I feel like a sense of awe about that because I am conscious that I am living through that we are all living through a transformation, you know, of that magnitude.

And I try to like convey that sense of awe in various ways. And it's awe and it's also humility. It's also like we really don't know. And I think in some ways, like I and many people in the AI, you know, AI is filled with a lot of people who have reflected very seriously about the industrial revolution and about technological change more broadly. And who are therefore very open to like, yeah, man, like stuff might get quite wild, you know, like things might be quite crazy because

the world of today is so wild compared to the world of a hundred and fifty years ago or something. And so there's like this kind of openness to like dramatic historical and technological change. that I think actually is quite off-putting to many people and like creeps some people out, if I'm being honest.

And I I occasionally find myself, you know, saying things and then getting clipped on the internet for saying like some crazy seeming thing. And it's like, Yeah, well, I was just riffing about the, you know, like how crazy historical change could sometimes be. So yeah.

Maybe even some will go back to information theory itself. You know, they'll go back.

Obviously, and that's what AGI is. And they'll kind of see that. That's the fundamental nut of what happened, of what is going on right now. So in other words, like we're seventy five years into an industrial revolution, in my view, a hundred years into one. And a lot of the societal transformations that we see and like this sort of instability, the the sort of sinning out, the hollowing out of institutional legitimacy, in particular in Western societies.

I think a lot of that can be characterized in some ways as downstream effectively of computation and of the internet, which is just computation. And so yeah, like that's sort of how I think about it. And viewed in that light, yes, I would Put my confidence I would feel I feel that in my book.

Look, we're building a thing that can do computations and eventually like what we'll wanna do is we'll wanna turn those computations into something that resembles or exceeds human cognitive patterns. And then, you know, that's kind of the end state, right? That's sort of the end state.

Is that when people look back on this era, it won't just be computation. It will also be like, you know, biotechnology will be a major area that people reflect on. I think probably space. Will be one. Maybe like robotics is a slightly separate thing. Physical autonomy. And the thing is that. Like AI is like this glue. It both accelerates and stitches together.

All of these other areas of technology that are advancing too. And so, much like the Industrial Revolution, you know, what you'll see looking back on it is like, There were these very, very powerful technologies that all got developed around the same time. And the development of one accelerated the development of the other for various reasons. And so yeah, that's kind of how I think it all.

age of AI, for lack of a better phrase, while simultaneously maintaining the spirit of the Constitution and the Bill of Rights. In other words, is it important when thinking about what you're trying to accomplish to consider not just the technology, but also the people and the society in which it's being introduced.

And like I think generally, sometimes that works, but usually the oak tree doesn't take, you know? It won't like the new soil and the roots it will not take root. I think you have to grow a tree over time in the soil that's native to it. And I think that that's like kind of how like everything

I think that's true that we have to do that, but I'm also extremely skeptical of the notion that you can do something like that rationally from the top down. I think that like the new thing that we do. the way the world will look in a hundred years or whatever, like that itself has to be grown. And I can't just dream that up and imagine it like, you know, from scratch, out of whole cloth.

And so yeah, like whenever I say things like that, I say things like that as a shorthand to describe what I think will be a monumental process of historical change. But the thing is is like there's a part of me that's like yelling at me when I say that. Like there's a part of my brain that's like, but be careful, don't be arrogant, don't fall into the conceit. That you with your tiny brain and limited perspective can like do this yourself, you know? Because we can't. No individual can.

helpful, but even that approach of yours is much more decentralized. Again, not a great word to use, but there is a kind of hands off approach to how you think about this that I think is really actually refreshing. And we're gonna have a chance to talk about that. So let I think we did a good enough job of giving people a sense of who you are and where you're coming from. Let's just go really high level now. What is your definition of AI?

human beings can. And it is the set of approaches to statistics, essentially, that have been invented over time to Try to make machines. think in the flexible way that humans and animals can. And that those approaches, to be clear,

are broader than just like, you know, there's obviously like large language models, which are the closest thing we have to like, yeah, that seems like it's kind of thinking in ways that are similar to humans. But there are also applications of the exact same statistical approaches. to completely different sets of data that are like totally alien to the human brain. So there are models, you know, that can predict the next nucleotide in a DNA sequence.

in the same exact way that there are models that can predict the next word in a paragraph of text. And like the next word in the paragraph of text thing feels somewhat analogous to human cognition, but obviously no human can think natively in nucleotides. But the same approaches work. So it is both. The particular thing of we want to make machines that think

like humans and that can perform human cognitive labor. And we wanna take those same statistical approaches that we use to invent the machines that can think like humans and apply them to different modalities that have nothing to do with human cognition.

patterns from the observation of data. And the reason that that's true, when you think about what that means, if I'm observing some Just the world, right? I'm looking at the world and what I eventually I find a pattern. I infer a pattern in the world. When I when I drop the ball, it falls to the ground every time. And I can predict.

Using that pattern, I can predict what will happen in the future, right? Because that pattern has been very reliable. I can also name that pattern. So like that pattern is a concept and it's a compression. It's like what I'm doing is I'm finding a faster way to go.

to understand what's happening around me because it's like, oh, I have a name for that thing. That thing is not totally alien to me. That thing is part of a pattern that I've observed before. And therefore, all intelligence is a form of compression. Language models are a compression of the internet. Well, they're a compression of their training data. Literally, that's what they are.

my brain is forming a compression of the real world. A very, very lossy compression. It's worth noting, right? Like, I mean, it's fantastic that, you know, I perceive the world and 3D space and it's quite impressive rendering of reality that my brain is able to deliver. At the same time, like We are aware scientifically that my brain perceives an extremely narrow range of light, right? There's like this whole universe out there of things that are happening that we have no sensory input to.

And so, you know, my brain is forming a really contingent. an ultimately quite lossy compression of reality. But intelligence is that act of compression. That's what it is. And so intelligence, yes, to put it simply, it is the inference of patterns from the observation of reality.

from the patterns, which is essential. It isn't just just to be clear, you're not saying it's just the observation of patterns or repeatable sequences. It's also using those patterns to infer something deeper about the structure of the data. or the structure of the world that allows you to make viable predictions.

because of just how much data they can process and have access to.

And the L L M responds with, It's a short distance, so you should just walk. What do answers like that reveal to us about the way in which these systems model the world that they can at once be both incredibly intelligent while also occasionally forming opinions or providing answers that just seem utterly idiotic. You know, granted, we have seen a major, and this is important to note, we've seen a major reduction in instances like these with each subsequent model improvement.

Like there are all sorts of things like humans fall for there are so many tricks we fall for. A, we can be rhetorically swayed in ways that large language models are harder to, you know. Like we fall for all sorts of fallacies that a large language model would catch immediately, right? Large language models would be like, that's an ad hominem. Like you're begging the question, right? And ways that like you adapt.

Or are important based on some kind of statistical correlation. I mean, it's kind of I spend enough time using these models in

sort of thinking through podcast episodes. And I again I'm in no position at this moment to sort of arrive at a conclusion about what's going on, but I can feel something meaningfully different about a how I think. versus how a large language model thinks. And Yes. I'd love for you to speak to that.

And it is 2017. There was a researcher named Alex Radford at OpenAI. And, you know, most of the AI companies at that time were trying to do, you know, they were teaching computers to play video games. They were doing robotics, board games, things like this.

And Radford had this idea of like, there's a data set of 80 million Amazon product reviews. And let me just train a model to predict the next character of those things, of those reviews. And In doing that, what they found is that they had accidentally, accidentally invented. A state-of-the-art sentiment analysis system. Sentiment analysis is like the ability of a machine to tell the underlying emotional sentiment of a chunk of text.

And it turned out that as an instrumentality of predicting the next of its fundamental goal, which was to predict the next letter in this database of product reviews, the model. Learned on its own to analyze the sentiment because understanding the sentiment of the text is useful for predicting the next token of that text. So there is like some higher level understanding going on there, but I think it's also extremely different. And I think that large language models are subject to all kinds of

You know, again, like we fall for optical illusions. Humans, there are all sorts of optical illusions. That humans fall for. And there's all these weird ticks of our cognitions that are ultimately downstream of contingent realities about the way our brains evolved. And I think the exact same thing is true of LLMs.

But I hesitate to say that that means that they're not really understanding. I think the more complicated reality is that sometimes they're really understanding and sometimes they're not. Sometimes they really are just, you know, sort of mindlessly predicting. And other times they're doing genuine understanding. And I think we've gotten much more genuine understanding over time. There are some people who would say,

you know, an LLM can never genuinely understand. And my view of that is that that is ultimately like a metaphysical claim about the nature of human cognition that you're more than welcome to believe, but like I guess my view is like at this point, OpenAI has a model that is unreleased but maybe six weeks ago made like a really important, like actually important discovery in geometry.

through connecting different ideas together. And like It kind of seems to me like that if you have an entity that is making like legitimate breakthroughs in mathematics. On actually important problems that humans have paid attention to, that have befuddled them for decades, and a large language model has like arrived at a new insight about that problem.

Seems to me that like if your definition of thinking doesn't include the entity that can do that, I feel like it's your definition of thinking that is the thing that is flawed.

and derives the important higher order insights without memorizing the entire thing. And in that sense, I am more sample efficient, which is also another way of saying, I have formed a better compression of the book. I have found the important patterns in that book, named them, and made them useful to me. I am more intelligent.

than the LLM is. And I think probably humans are orders of magnitude more sample efficient than the LLM. But the LLM, by virtue of being a computer, is like still quite Useful and it would imply that there are like many orders of magnitude we can go to make the LLM like closer to human level sample efficiency.

Then in other words, the pattern doesn't require meaning. And so when you're talking about sample efficiency and yes, human beings need a smaller sample size to derive meaning from a corpus of data. I wonder to what degree, Dean, and I don't want to take us too far afield here, because this is more of a philosophical discussion that you and I need to have at three AM in our dorm room.

But I wonder to what degree Something very different is going on here for large language models and that in order for any of this to make sense, they need human beings as the sort of person in the mirror, that the meaning is something that we derive. It's not something that they are able to derive from that information. Yeah. Feel free to respond to that. And then I have a series of, I think, more answerable questions.

You know, humans had to identify, like we find structure in the world and we make predictions and infer patterns based upon that structure. And there are common structures. that the LLMs are fighting and that the humans are fighting, but the world is also very large and rich.

And there are all kinds of structures that humans find that LLMs don't find, and all kinds of structures that LLMs find that humans don't find. And I think, by the way, we've barely scratched the surface of what the LLMs understand about the world.

that we don't understand. Because I think there actually probably is quite a bit that they understand that we don't understand. But I guess what I would say is like, I think what you're getting at is this notion that like agency and goals and like meaning, purpose are things that are

much closer to being unique to human cognition. Ultimately, we don't know the answer. Ultimately, like, I think one thing that really scares people in the AI safety world is like, you know, we call AI systems of today, we call them agents. But are they agents in the legal sense? Which is to say like, you know, my lawyer is my agent and that means he does exactly he's like, I'm delegating some aspects of my life.

to that person and but that person is expected to have an absolute duty to me and is to be like a complete subordinate essentially to me or are they a philosophical agent? Which is to say something that can act rather than behave, act on its own, have goals of its own that it autonomously acts upon.

And I think that is like a scary thing to think of like, you know, super intelligent machines with goals like that. I think that is like a really big unanswered question. But I think there is some reason to believe that like the human agency and finding of purpose. It's not purely an artifact of like our intellect or our reason. It is coming from some deeper aspect of our biology and our survival instinct and things like that that like language models are not trained to have.

So I just wanted to put that out there for anyone that's interested in obviously reading Claude Shannon's work is I think a important complement to this conversation. In your article twenty twenty three or why I am not a doomer, you wrote that in that year you went from being a skeptic to a believer and you would walk around, was it the campus? Where were you at this time?

Does AI break the existing constitutional order of the United States or does it merely challenge it, which I think is a very important question that we'll have a chance to dive into when we talk about regulation? I'm just curious to understand where would you say that you fall on this continuum of Doomer versus accelerationist boomer? And by boomer I don't mean baby boomer, I mean like boom, you know, this is the AI boom.

And like the way the world works, broadly conceived, that causes them to have this like very simplistic notion of what could bring about human extinction or doom, as it were. And I guess like my view is like Again, put yourself in the perspective of 1775 and then try to imagine like World War One happening, right? And like imagine explaining to that person like what a person who only had the knowledge of 1775, the specific

war fighting dynamics of World War One. And it's like, man, that would have been so terrifying and unpredictable. And there are, there are things

not like cool. It is cool. It is cool to be clear. But it is also like there is a certain terror to it. And so I guess when I think about like technological change in macro historical terms. There's a part of me where like sometimes I feel like I'm actually more of a doomer than the doomers because I just realize like how alien what we're going to build will be.

and like what sort of like unimaginable horrors likely lurk around the corner. But I also think unbelievable, you know, joy, like surprising joy to the upside, delight. is also just as likely. And in the end, I think that there will be more good than bad. But yeah, I I do think that like there's tremendous humility that you need to experience if you are doing the enterprise of commenting on AI.

That's what happens when you try to do that. It's very I mean, in the early days when if you were reading like, you know, Bostrom's work or some of the stuff that was coming out of the existential risk community, it was very easy to adopt the if anyone builds it, everyone dies view that Yudkowski puts forward. But I think as these models get built out and they get better and better, you begin to, I think, at least if you're being intellectually honest.

understand that the picture is much more complicated. And AI alignment's a great example. I mean, in the way that I used to think about the alignment problem, it was like, how do we bridge the explanatory first of all, how do we even sort of come to a clear idea internally about what the values are that we want to impart on these systems.

And then two, how do we bridge the explanatory or epistemological gap and convey them in a manner that they will understand? And then so that we, you know, don't all end up dead because AI decides that we'd be happier if they lobotomized humans or put, you know, sensory neurons in our brains and jacked us up with dopamine or put us in a vat or something. But actually what you realize is that I feel at least that the alignment problem is less of an issue.

Because so much of what these models are trained to do is to understand, I think in some ways better than us what it is that we want. Now I might regret saying that, and I should also say that it doesn't make me less of a doomer. I'm not a doomer, but it doesn't sort of I that's less sort of where my doomishness comes from. But I just it's just to say that it's way more complicated. And I think it's important to have a nuanced view of this. So in the spirit of that

What would you say are some of the things that both the doomers and the boomers, again, this is my term, but it's not the baby boomers. I could see how that could be confusing, get wrong in your view.

incredible amounts of agency and ability to like affect change in the world. And I think that's like not quite accurate. It's in some ways it's accurate, but I think it's a they're painting with too broad a brush. And there's a there's a lot of hand waviness in that view.

What I would say about the I mean, when it comes to the boomer side of things, the sort of techno optimist side, there's a lot of different worldviews that are captured there. So, like for example, there's one category of boomer. who really their optimism about AI and their opposition to government involvement in AI

basically comes from a place of actually ironically enough kind of pessimism about the technology. They have kind of a deflationary view of the technology. So there were a lot of people, you know, there was a leak draft of an executive order that the Trump administration, that the White House put out. It leaked from the White House, but it was real. It leaked in like November, December of last year.

reference the notion of catastrophic risks like, you know, autonomous cyber attacks or or bio and stuff like this. And it dismissed those things as purely speculative and hypothetical. And it's like Looking at the state of the technology as of late 2025, if you were willing to even put like write down the words. That those kinds of risks are totally speculative and hypothetical. You had a profoundly deflationary view of the technology.

And one thing I've always suspected about the boomers, and we actually see this a little bit starting to emerge now, one of my theories of mind of that community, which again, like I've largely been allied with for most of my career writing about this. But one of my theories of mind about them has been like I bet you they're actually like Once they fully realize what is actually going on with this technology.

They will actually be quite terrified of it and that they might actually be closeted AI pause ban people. You know, like that the notion of super intelligence. actually being real. Like very few people helped internalize that notion in their bones. and then or at least made an effort to and then also been like, and we should build it. I feel quite lonely in that camp. There's a few of us, but there's not many of us.

And that a lot of the people who are boomers actually just really their optimism is rooted in fundamentally a quite deflationary view of the technology. That means that like, Oh, well, these problems that the AI safety world points out, they're just not problems that we're gonna have to deal with really. They're just dismissing the problems. They're kind of kicking the can down the road. I would say like that's the fundamental mistake of most boomers.

But it is worth noting that there is a spectrum there and not all of them have that deflationary view.

Which is your approach to regulation. Again, not a really good word. Maybe your approach to how to oversee and manage this transition into the era of AI so that we can maintain and protect the things that we value, whether those things be our humanity, our Constitution, our Bill of Rights, whatever. And so adjust our socioeconomic and cultural systems and socioeconomic systems.

to this new reality. And I'm also curious how that fits into the larger geostrategic competition which between the United States and China, where this has also shown up in the ongoing dispute between Anthropic and the Department of War. what this means for other countries, their economies, societies, and cultures, whether some are better adapted or more maladapted than others. And also what does this do to the labor market and the economy? This is something that people are very concerned about.

And I would love to talk about all of these things with you in the second hour. For anyone new to the program, Hidden Forces is listener supportive. We don't accept advertisers. or commercial sponsors. The entire show is funded from top to bottom by listeners like you. If you want access to the second hour of today's conversation with Dean, head over to hiddenforces.io slash subscribe and sign up to one of our three content tiers.

All subscribers gain access to our premium feed, which you can use to listen to the rest of today's conversation on your mobile device using your favorite podcast app, just like you're listening to this episode right now. Dean, stick around. We're gonna move the second hour of our conversation onto the premium feed.

If you want to listen in on the rest of today's conversation, head over to hiddenforces.io/subscribe and join our premium feed. If you want to join in on the conversation and become a member of the Hidden Forces genius community, you can also Do that through our subscriber page. Today's episode was produced by me and edited by Stilianos Nicolau. For more episodes, you can check out our website at hiddenforces.io. You can follow me on Twitter at And you can email me.

Info at hiddenforces.io. As always, thanks for listening. We'll see you next time.