DP

As an undergraduate at Reed you studied Mathematics and Anthropology. I’m curious, why both of those disciplines? And how did they inform one another? And how do they inform your present work as a historian?

SG

In high school (this is going to sound obnoxious) I was very into philosophy. I was a difficult teen in a difficult situation and so, for whatever reason, I turned toward Being and Time and Martin Heidegger. I don’t know why. I got into Hubert Dreyfus and David Harvey as well. I thus came into college thinking I wanted to be a philosopher. All that was on offer at Reed was analytic philosophy, which I found very gendered. Compounding that feeling were these very gendered scandals that wracked the field at the time. I don’t know if you remember this.

DP

No, I don’t.

SG

Around 2013 this major analytic philosopher at the University of Miami, Colin McGinn, got into a lot of trouble after he was found exchanging emails with a female graduate student about hand jobs. He went on to get into all this sophistry about what he really meant by the term “hand job.” All this is to say, I was turned off to the idea of studying analytic philosophy any further.

DP

Right.

SG

I thought that anthropology might be a way to ask the same kinds of questions but on what I felt was a more empirical basis. You could ask, for instance, what it means to be human by spending time with different kinds of humans from different cultures and traditions. This was very appealing to me, so I decided to major in anthropology.

The math, though. I was reading something by Manuel Delanda about field theory. I thought, “Oh, math sounds cool.” I took a couple classes to try it out. A professor was like, “You should be a math major,” which was a thought that never occurred to me. I was like, “Okay, sure.” So I became a math major. At Reed, you do a year long thesis in each of your majors. I double majored, so I did two.

With the math thesis, I had to shop around for what I might do. I was torn between two things. One of which was about optimizing factories—complicated logistics—which I still find super interesting. The other option was to work with this guy who was applying abstract algebra, a field which I was super interested in. His work was in the sub-field of quantum algorithms, which I knew nothing about. I ended up writing my senior thesis on optimizing queries for quantum algorithms, which basically sent me down the path that I ended up on for my PhD.

As an anthropologist, I was like, “Why are we writing algorithms for computers that don’t exist, might never exist?” That question was hard for me to wrap my head around. My supervisor, this very generous, sweet professor named Jamie Pommersheim, had all these interesting answers. Like, when Charles Babbage was developing a theory of computation, he didn’t anticipate the advent of digital computers as they exist today. Anyways, none of those kinds of explanations were totally satisfying, though. I was like, yeah, but Babbage wasn’t actually anticipating the existence of these machines, right? In this specific form. Then I was like, well how was this getting funded? Why is this getting funded? Where is that infrastructure coming from to support this speculative technology?

I applied to Columbia with that series of questions in mind. And even with my background in anthropology, I got into an American history program, which was fantastic. It was 2016 when I came in. Eric Foner’s history department was a great place to metabolize what was happening with Trump, to make sense of this huge shift in my beliefs regarding what the future was going to be like, what all this meant, and where it all came from.

DP

What were you researching at the time?

SG

I started doing oral history interviews and ethnographic work with early practitioners in quantum computing. The quantum computing community was very, very generous with me. Charlie Bennett and Scott Aronson connected me with all of these figures and their networks. I got in touch with folks in other places, like in the national security state at the National Science Foundation (NSF).

I was pursuing the question: How was it possible to sustain all this infrastructure for a field that didn’t seem like it was going to deliver anything monetizable for a very long time. How, in an age where everything is about utility and applications, did this field work? This, not to say I think that speculative funding is bad. Fundamental physics research is good, which is basically what some quantum computing research facilitates.

I began to see the state and academic infrastructures that were in place for speculative technological projects. In understanding where those infrastructures came from, I realized I needed to understand the semiconductor industry, which constructed an entire political economy around future technology vis-a-vis Moore's Law.

DP

Can you briefly explain what Moore’s Law is for our readers, many of whom, I assume, are unfamiliar.

SG

Moore’s Law is the claim that chips naturally get smaller, cheaper, and/or faster over standard increments of time.

DP

How is this related to the kind of speculative funding you just mentioned?

SG

Chris Lecuyer writes well about how Moore’s Law allowed Intel to sell futures along with their “products.” It’s not that Intel’s chips were significantly better than those of their competitors, but they had this very powerful marketing device—again, Moore’s Law—which became, in effect, a self-fulfilling prophecy.

I came to understand that I was really writing about the military industrial complex, or the defense industrial base—whatever moniker you want to give it. I wanted to know how the semiconductor industry received state funding even during periods of neoliberal austerity. I got really interested in  figures like Norm Augustine to track and make sense of a bunch of these trajectories.

DP

Prior to our conversation I was doing some research on Moore’s Law and a paradox emerged, that is: Moore’s law is not a natural law, but seems to be given as such. That is to say, Moore’s law relies, like you said, on speculative research to be fulfilled and yet, it entirely guides research in the field and there is no agreement on when, if ever, it will cease.

SG

Christophe Lecuyer has a great piece on this. Coincident with the efforts to make chips smaller vis-a-vis Moore’s law was an internal debate in computing and physics about the limits of computing. Is silicon the right medium for computing? When will heat dissipation become a problem? The prominent engineer Carver Mead adopted the Moore’s Law position, which was that silicon can scale forever. Mead became a big proselytizer for the Moore’s law line and was close with Intel executives while also engaging with the quantum theory and computing people. In any case, Robert Dennard, an engineer at IBM, went to one of Mead’s talks and decided that he bought what Mead was saying and rephrased Moore’s Law into a mathematical law, the famous log plot.

Moore’s Law through this and later industrial policy like the various roadmaps becomes an expectation. It gets built into all this infrastructure, including when the government gets involved in helping semiconductor industries mid crisis plan for the future. Moore's Law becomes so powerful, in fact, that people think it’s self-perpetuating. In a way, it is: it’s something that’s almost always in crisis, right? There’s always something to act on. You don’t need an external threat. You can just say, “Moore’s Law is ending,” which becomes justification enough for further research, funding, etc.

There are some scholars of technology who consider all of this and come to the conclusion that Moore’s Law is a natural law. I think that’s somewhat misguided. The term social fact may be more accurate. In either case, by the end of the ‘90s, Moore’s Law becomes so popular that people think it’s just going to continue without investment. It became an ideology—this is how certain technologies naturally develop.

The other thread that’s worth following is the tremendous fear and anxiety about the status of American capitalism in the 1970s and 1980s. At the time, the Japanese seemed to be doing capitalism better. There was all this talk about the limits to growth (which is something that still freaks out the tech billionaire types like Elon Musk), which Moore’s Law seemed to enjoin. The Law provides an imagination of continuous capitalist growth, undergirding the post-war American ideal: There aren’t going to be limits to growth, and there won’t be hard political choices or problems, instead, there will be technological fixes produced by a techno-scientific elite.

DP

The real force of Moore’s Law seems to be the way it creates a container into which the financial and phantasmatic of US capitalists can pour.

SG

Yes, Moore’s Law is a beautiful visual. The ‘line go up’ log graph that you just see everywhere, the one that tech guys love invoking, is the basis of the belief that all tech companies will have exponential returns. Moore’s Law has been able to deliver on that promise in certain moments. However, it may be the case that the system that undergirded the political economy of big tech is falling apart with the destruction of so much of America’s scientific and industrial policy institutions. It may be that belief in tech’s ability to produce exponential returns and solve hard political problems weakens with that; I don’t know.

The visual of the graph also rhymes with the stories of civilizational development that tech guys often tell in the present. First ‘man’ invented fire, then steam power, then industry, then software…ad infinitum… this way of presenting history as stages of development toward modernity.

DP

I’m interested in the ways that the aesthetic, psychological, and effective experience of defense is altered by the understanding that chips will become increasingly smaller, efficient, and lethal vis-a-vis Moore’s Law.

SG

The whole electronics industry is a product of World War II industrial policy. Indeed, that industry remained military guided through the 1970s until the Cold War State fell apart and the civilian market expanded. There has since been some decoupling, but the American military’s imagination and strategic doctrine has been very tied to Moore’s Law since the 1970s.

In the 1970s, US military planners saw that the Soviets had many more potential soldiers than the US. The US imagined that they would offset their deficit with high tech weaponry and surveillance. As such, there’s a way in which the present focus on high tech weaponry, precision, and computing is intimately bound up with the creation of the World War II state. We often think about the New Deal State and then the Cold War State, but maybe we'd be better off thinking about a World War II State that still hasn’t died.

I’m cribbing Gary Wells’ great book Bomb Power here, but this WWII moment sets in motion the concentration of executive power, the US’s reliance on high tech weaponry, and the use of data stats to run the war machine. ex-Defense Secretary Robert McNamara’s interest in metrics, data, and optimization is, in effect, a way for him to concentrate power in his office. He’s famously a megalomaniac in that regard. This trend continues up to the present. During Vietnam, we saw General Westmoreland’s notion of the electronic battlefield. Today we have a focus on sensors, on precision strikes, on remote control of battlefields.

To get to the heart of your question, though: the emphasis on efficiency, lethality, and accuracy, in my view, paradoxically leaves more room for brutality. Obama’s drone war opened up the space for brutality. Teju Cole made this point in a 2013 New Yorker article, I believe. Drones do not precisely assassinate just one person on the first try, there are often multiple strikes; Cole mentions an attempt to kill a Taliban chief that only succeeded on the 17th attempt and after killing 280-410 others.

DP

In the present, so much money, energy, and discourse is invested into AI and related software. Can our moment be historicized by looking at how Moore’s Law steered growth in the 70s, 80s, and 90s?

SG

To be sure, Moore’s Law hasn’t gone anywhere. It undergirds the techno-optimism of the 2010s, the 2020s, and the present. For instance, the notion that every iteration of the iPhone will be much better relies on Moore’s Law. It’s also the unspoken assumption of venture capital firms that bring tech companies to IPO and cash out. All these tech CEOs replicate the Moore’s Law log curve when discussing their own technology. So, in that capacity, absolutely.

AI is a great example of that. It’s imagined that AI is just going to keep getting better along the same, logarithmic pattern. In truth, that kind of growth very rarely occurs.

DP

In “Runaway Short-Termism,” you note that you studied in Eric Forner’s history department at Columbia, and how that program attuned you to America’s regionalism. How can thinking about the regional conventions of a period like, for instance, reconstruction, contextualize the present?

SG

It was a great education. I loved the Civil War class, which I TA’d for under Steph McCurry, who’s one of the great Civil War historians. One of the things I got from her is an attunement to gender, which, now that I’m doing national security, is everywhere. I mean, just recently Peter Navarro alleged that if we bring back manufacturing, we’ll solve domestic violence, declining fertility rates, abortion, and drug trafficking, because women would be back in the home. I also learned a lot from studying American social history and activist traditions. If I were simply trained as a historian of technology or science, I might have missed those. My on the ground experience organizing with the UAW first for our [Columbia student workers’ union] contract and then participating in the autoworker strike also really shaped the way that I think about politics. In that sense, Columbia was a double education.

But, to get to your question, the US Civil War historians really did encourage me to think about things regionally, even up to the present day. For instance, people often credit the sort of Atari Democrat, high tech vision for taking the Democratic Party away from its original base in labor and the social welfare state. You can see threads of that specific vision in McNamara, obviously, but I think it’s precisely the very intense regional recessions at the end of the Vietnam War that really affected Northern and Southern California and Massachusetts—defense workers organize in response and—this is skipping over a lot but—eventually come to think of themselves as high-tech entrepreneurs. Those tech entrepreneurs are part of a coalition that built a campaign for tech-led development, venture capital, startups, small businesses and new tech companies became increasingly influential within the Democratic Party and, more recently, the Republican Party.

A question I often ask myself is: Why isn’t the Democratic Party treating the Silicon Valley takeover of, for example, the Pentagon with more alarm? Many of those Silicon Valley companies are tied into the Democratic Party or at least important figures in the party—this is in many ways still the party of Atari Dems. Democratic politicians in, say, the Midwest seem to view things with more alarm—they don’t represent the same constituents. That’s not true in Northern California. In all of these capacities, regional politics may re-emerge in a more obvious way in the future, though I think it’s been an ignored feature of the last party system as well.

DP

I found your comments on the gendered aspects of defense tech very resonant. As you note, there are a myriad of gendered vectors that cross that industry. The present administration, you argue, seems adamant on transitioning away from a “supposedly feminized service-based economy” and towards “a masculine-coded, producerist one organized around defense manufacturing.” The way that gender ideology functions as a part of this picture does not seem very straightforward to me. I was wondering if you could help untie these threads.

SG

Oh, it’s not straightforward. There’s a way in which this marries the Christian and this new re-industrialization efforts that’s enrolling some of these Silicon Valley guys into imagining ‘fixing’ the gender and racial order can come hand in hand with reindustrialization and revitalization of American power.

I think people like Mark Zuckerberg were perturbed by the accusation that they weren’t patriotic enough because they were siphoning the best minds of their generations into feminized ad tech. They’ve responded by becoming much more masculinist themselves. Have you noticed that he changed the way he dresses? (Ironically, that same ad-tech is now what ICE is hoping to adapt to make their ethnic-cleansing efforts more efficient).

DP

He’s gotten built too.

SG

Right. They’re also all being like, “Let’s go, I’m working in defense now.” They send their executives to work in the Pentagon. The prohibitions against working for the military industrial complex, prohibitions descended from the Vietnam War protests, have vanished, at least for now.

Some of the concerns and sense of urgency about revitalizing American power and defense in particular to confront China, seem to be a veiled discomfort and backlash to #MeToo and BLM as well. That might be why there’s such a resurgence of gendered politics among those kinds of elites. Alex Karp, the CEO of Palantir, talks about masculinity a lot. There’s a fixation on male loneliness, on men being left behind. Probably everyone is more lonely, this is just a way to focus discourse around a gendered grievance politics that appeals to these elites and mirrors their own feelings.

DP

If anything, the deliberate masculinization of those in the defense tech industry just seems to reveal the absolute vacancy of the category of “man” itself.

SG

Yes. Totally. I also think there’s something to look out for on the left. Sometimes I see this nostalgia for the mid-twentieth-century that I find pretty disturbing given the kinds of factory jobs people were working. Of course it would be great if we had powerful labor unions, but, in truth, by mid-century, they were already on the decline. There’s a great podcast on the history of the AFL-CIO that makes that case.

Not to mention that factory jobs are not particularly good jobs. Marx says this Capital Vol. One. He writes beautiful metaphors about monsters and vampires, about how the machinery of a factory literally consumes the sinews and the muscles of workers. I would see that when I did national work with UAW autoworkers who that actually occurred to. You realize, “Oh, this is extremely literal. These are not often good jobs.”

The gender and racial order that came along with that is not something the left should have nostalgia for. When we talk about the 1950s being a period where there was much less economic inequality according to certain measures, I think maybe we miss the gendered and racial components of that period, the kinds of inequalities of which are not as easily quantified.

DP

You mentioned that your next book will evolve around interviews you’re conducting with people in the defense industry. Why do you remain invested in that specific methodology, the interview?

SG

I think oral histories are a good compliment to archival work. I don’t ever replace archival work with oral interviews, but, especially because I study the national security state, there’s so much that’s not public. And, even if I wasn’t studying something that had redacted components, I feel oral histories often point me in interesting and different directions. I also think oral histories can provide additional ethnographic richness. When you’re writing about something it’s important to glean as much of a sense of context of the world of your object. Obviously medieval historians can’t do this, or at least in the same way, but I can. I have such luxury.

DP

I sent you this article published by the New York Times’ editorial board titled “Why the US Military Must Reinvent Itself.” That article, on my reading, appears like an advertisement for the defense tech industry. The overtly nationalistic tone of that article got me wondering: How you view yourself in relation to the material that you write about and study. Polemical? Instrumental? Critical?

SG

I think those articles are incredibly credulous, and I think the authors got a bunch of stuff that's pretty easy to verify, wrong, which is too bad. But it’s not surprising. I think tech defense firms and their funders have spent a lot on fancy PR. I think the subtext of that article, if we’re being honest, is about hedging the AI bubble; drones offer another place to send capital and potentially get exponential or at least stable returns. People are clearly getting nervous about it. There’s this whole sect of that industry who are already asking for the government to bail them out of the debt they incurred during these really insane data center construction projects. Some of these financial entities are hedging their bets. Small to medium sized drones are already a massive bubble. I think like $200 billion in last year’s NBAA were allotted to drones. There’s going to be even more money for drones in this next NDAA, the 2026 one.

In that sense, a lot of these new things coming out about the drone war, its inevitability, who’s credible, and who’s not, are a fight over that pot of money, a way of insuring investments. Who’s going to be the Nvidia of the drone bubble? When I was doing research, I met this Ukrainian drone manufacturer who suggested certain pieces about certain drone companies are placed by folks with material interests. I’m sure that’s happening all over the place. I have to imagine that this Times series is a product of the money sloshing around drones. Trump is not exactly lowering the defense budget, so there’s lots of money to fight over. I just wish they were less credulous and a little more diligent; we shouldn’t be led around by the nose by these financial manipulations.

But you were asking me a much more thoughtful question, which is not about this, but rather about how I position myself, and why it’s worth writing the stuff that I write. I think there are a lot of big changes that will be wrought on the national security state in the next few years. I’m in a good position to write about some of those changes in a critical way. And I think a lot of the loudest voices are full of shit. I don’t know if anyone will take what I say seriously. I’m not omniscient. I’m always learning, always talking to new people, and always trying to understand what’s happening. I hope my work will give someone a better understanding of how we got here so that we might not make the same mistakes again.

DP

My colleague Henry and I just published a piece in November about the overwhelming pace of history since 2020, the too muchness of the last six years. This is a feeling you circled around near the end of your conversation with Nic Johnson published in The New York Review of Books:

I wonder whether we’ve neglected just how rapid the pace of change has been, how many different sorts of things are happening—that we know about—and how quickly the administration seems to be pivoting from one point of view to another…[The] rate of flux is overwhelming, and it makes it hard to think as far as even the next month. When I was writing my piece, every day the administration would do something else, and I kept thinking, Well, now this is reshaping my picture. I have to add this. This article is never gonna end.

I’m wondering how you deal with the overwhelming sense of history and the present as a historian?

SG

Part of what’s useful about being a historian is that history isn’t changing very fast. You can get purchase on how transitions have happened in the past. You can trace things happening now in a calmer, more stable way.

I teach this class on the history of time, the history of history writing, technologies of time, and the invention of simultaneity. I’ve been reading Thomas Mann’s The Magic Mountain, which I think I’m going to teach in that class because it’s all about the porousness of time: how seven years can pass in a couple chapters, how people grapple with disruptions in their sense of time. Mann writes at the beginning about the periodic character of time. To provide a contemporary analog, the Biden years, as a period, felt slow in a certain way.

DP

Henry and I wrote in our essay that the Biden years didn’t even seem to happen. They slipped through memory.

SG

Right. Likewise, pre- and post-2016 feel like such different worlds. Thomas Mann has this interesting passage about interwar period:

But let us not intentionally obscure a clear state of affairs: the extraordinary pastness of our story results from its having taken place before a certain turning point, on the far side of a rift that has cut deeply through our lives and consciousness. It takes place, or, to avoid any present tense whatever, it took place back then, long ago, in the old days of the world before the Great War, with whose beginning so many things began whose beginnings, it seems, have not yet ceased. It took place before the war, then, though not long before. But is not the pastness of a story that much more profound, more complete, more like a fairy tale, the tighter it fits up against the “before”?

When I read this passage, I was like, wow, that feels like our present moment in such a real way. There is something about certain breaks in history.

Recently, it feels as if something has been set into motion. There is a novel openness. Anything could fucking happen. We live in a moment where I could imagine so many different outcomes. It’s been clear to me since 2016, at least since 2018, that we were at the beginning of a very weird, long journey, the outcome of which would be determined by the setting of a new global order, or a new politics. That openness hasn’t yet closed, even this many years later. And I don’t know when those beginnings will stop beginning.

There is both a good and a bad piece to that: on the one hand, things could get a lot worse and, on the other hand, things could get better. We should be organizing and building things. I think people are. We’re not stuck in the world we were born in, but we’re not yet in the world we’ll find ourselves in.

I guess I have this sense that things might stabilize. With that said, the Trump administration does stuff with such volume. There are so many stories that under Biden would be massive scandals, for instance: the fact that the Navy and the army are proposing selling off pieces of shipyards, or, outsourcing the funding of modernization in exchange for considerable private capital control over the government. That’s a huge fucking story. People wrote about this in Trump 1 and it’s no longer novel, but it still overwhelms. And it does mean that whenever I’m trying to go on a podcast or do an interview, I’m like, shit, there must have been fifty things that I missed because I’m on research leave.