00:04:09.140The future is with rocket ships in space.
00:04:11.900I don't understand why that idea has been so culturally durable, especially here in the U.S.
00:04:19.220You know, one of the weird things about being a historian is you just see, like, everything happening over and over again.
00:04:25.300And, you know, the same words even, like, re-skilling is one of those.
00:04:29.360But also these imaginations of the future coming from that post-war period.
00:04:33.880Now, I'm kind of interested in the way that there's, like, a break between those two periods.
00:04:39.380But I think there is something about the vision of the future that's maintained. So I think there's like this original construction of like a scientific state apparatus and political economy around that, that like codifies some elements of the World War II state.
00:04:54.280And that is brought into full maturity with, like, sort of the Vannevar Bush era of, like, imagining, you know, physicists and basic research as, like, the prototypical model for the science-state relationship, much like it was in World War II.
00:05:11.560That's stable until, like, the early, maybe, like, the mid-1960s.
00:05:16.800I'm reading a book that's making me rethink the periodization even earlier of this, like, an annoying historian thing to do.
00:05:22.020Um, but, but part of that, that, that Benavar Bush vision of like this relationship between
00:05:29.680state science and industry is remade in sort of like calling upon those, those same moments
00:05:37.120in those same imaginaries, but in a very different like structure, beginning with the erosion
00:05:42.500of that original World War II research state or like the Cold War part of it.
00:05:47.580And it transforms into like kind of more of the big tech model where you have venture capital, you have these lifecycle models, and they appropriate, I think there are certain features of that conversion that maintain a sort of imaginary of the future that is very consistent.
00:06:05.460It is weird, right, that the kinds of things that Elon Musk's saying are always on the horizon and that those are consistent from, I think, the 70s and 80s.
00:06:16.360But I think what this maybe does for us, this periodization, is that it's not so much a product of the post-Cold War period, but, like, they're products of, like, the Cold War model maybe, like, falling apart and this need to, like, reauthorize, to reimagine, like, what science can do for society.
00:06:35.960Like, physicists start freaking out about this due to, like, defense cuts and all of these kinds of business efficiencies that are brought into government in a bipartisan way.
00:06:46.360in the mid to late 60s i think these things are related i think there's this great american
00:06:53.040institute of physics meeting where they're they're really worried about this it's like i think 1967
00:06:58.4201968 and they're like you know we did the bomb we did radar we did all this like amazing stuff
00:07:03.380but it's not enough to convince government and the public that they need to continue funding
00:07:08.660our institutions like they think darpa might go away they think all these things might go away
00:07:12.220um and so we have to like offer a different model of like what we do for society and
00:07:19.500a bunch of them become convinced that like physics had had this purchase on the future
00:07:26.320right like you could use physics to predict like where objects would be in space in the future and
00:07:33.680that was wonderful yeah i'm cribbing like john wheeler's archives he's like oh this gave like
00:07:39.600you know physics a special authority we're gonna talk about john wheeler yeah um yes yes absolutely
00:07:45.860so wheeler is like oh like we had this purchase on the future but actually like quantum physics
00:07:50.620really unsettles this and like we don't have this anymore but you know what does computing which
00:07:55.980again the relationship of physics and computation is something that peter gallison has written a lot
00:08:01.180about for instance but like yeah they have this really credulous relationship to computing that
00:08:05.300not shared by computing experts themselves right like um rebecca slayton has this great book about
00:08:10.280the evolution of or the emergence of software engineering as a field unto itself and like
00:08:15.720they're much less credulous about what computers can do and their purchase on the future and like
00:08:20.400truth and nature right in reality yeah yeah yeah physicists because of monte carlo simulations and
00:08:27.140their role in sort of simulating the bomb and like simulating physics this idea that you could
00:08:32.720simulate all kinds of physics quantum right obviously complicates that yes gives physicists
00:08:38.020this real belief in the power of computing and so wheeler in this moment is like well yeah so we
00:08:44.540have to like build this closer relationship with computing or like all these it's not just him but
00:08:48.620all these physicists get really interested in this period in computing because it seems to be
00:08:52.960something that would have this greater purchase on the future at the same time as they're
00:08:58.120reconstruction reconstructing the case in the institutions attendant to that for like why you
00:09:04.820should fund them their students big science and what comes out of it is very different right than
00:09:11.940the world war ii state but we can get to that okay i have so many questions and there's also
00:09:17.080a lot of things we need to unpack um yes yeah so uh vannevar bush publishes in what like 1946
00:09:24.760the endless frontier um right uh which is this um yeah yeah somewhere around there so like right
00:09:32.820after world war ii vannevar bush who is um how would you describe him scientific administrator
00:09:37.680um yeah yeah and proto entrepreneur i mean there's like really interesting things that
00:09:44.040have been written about him and like he had this like um commitment to basic science but yes maybe
00:09:49.820was a little bit more applied in practice however you want to divide those categories but yeah so
00:09:54.660So he ends up proposing the National Science Foundation, the NSF, and helping to build
00:10:01.500a bunch of the post-war science, the military-industrial complex, the triple helix, the
00:10:08.020Cold War triple helix, however you want to talk about it.
00:10:18.120And the government, again, I think it's not just defense, but it is a lot of defense.
00:10:22.460So Bush publishes this thing saying, like, science is an endless frontier, explicitly invoking this frontier metaphor. This idea of a scientific frontier and this post-World War II vision of a sort of shiny physics, rockets, and chrome future lasted until sometime into the 60s and then sort of got reinterpreted.
00:10:49.200And and parts of it persisted to the present, but parts of it sort of shifted over to being more focused on computers and computation and ultimately ended up with this like Silicon Valley model.
00:11:05.100So along the way to answering the question that I asked you, you are also answering a different question, which is, how is it that these, you know, tech billionaires became the guardians of this fantastical vision of the future, for better or for worse?
00:11:27.760Spoiler alert, for worse, not for better.
00:11:30.620I mean, I think there's so many different things happening that all kind of complement each other at the same time. So, you know, I'm probably leaving out part of the story that someone would add, but I think it is the case that like the sort of physicist center modeled of the future has to be re-articulated and the state, like this research apparatus, research industrial, scientific defense, you know, government apparatus gets remade beginning in the late 60s.
00:12:00.460There's this big crisis of funding, defense cuts, but really that like applies to like
00:12:20.960So Moore's law, for instance, which is, you know, I mean, it has come to mean many different
00:12:27.340things but it's like this a graph a log graph so log means like you can present an exponential
00:12:34.840curve as a line going up by taking the logarithm so the rate of change and more as well as a log
00:12:40.960graph that just shows a line going up and tells a story about the relationship between computing
00:12:47.760power speed and size right that like things get smaller cheaper and faster in standard increments
00:12:54.080I want to just jump in and say, like, one way of explaining Moore's Law is you can double the number of transistors that you can cram into the same space on a silicon chip every 18 or 24 months, depending on which formulation you want to look at. Right? Something like that?
00:13:09.920Yeah. And so, and again, it's like, it means very different things to different people, like how it works. But it is like the image itself is a way of telling history and a way of projecting the future at the same time. And we are maybe less attendant to the ways in which it shapes our understanding of the past and more attendant to the way it shapes our understanding of the future. But I think these things are related.
00:13:34.800And I think the tech CEO as the person delivering this in the popular imaginary means that, okay, so then they understand you have to fund these industries in a specific way to maintain Moore's Law.
00:13:47.440Moore's Law is something that can perpetually be in crisis.
00:13:50.840And the role of Moore's Law has changed over time.
00:13:53.080So I don't want to say that it's a monolith or that it hasn't shifted.
00:13:57.540And the power in the tech industry hasn't shifted, which it totally has.
00:14:00.360This is fascinating because in my book, my latest book, not the first book about quantum physics, More Everything Forever, I talk about Moore's Law and I say, you know, it's not a law of nature.
00:14:11.540it was a decision made by people running the semiconductor industry that this was something
00:14:17.680that they were going to shoot for. And you're saying, yeah, it was. And it's also a way of
00:14:26.340justifying actions and funding requests and saying like, we have to stay on this track
00:14:33.020in order to save Moore's Law. Yeah. I mean, if you're going to get really specific about it,
00:14:37.480Moore's Law starts is like a marketing tool because like it turns out Intel doesn't actually
00:14:41.940have better products than its competitors but what they're able to do is to sell not just the
00:14:47.360products but the promise that it's going to continue you know increasing at these intervals
00:14:51.660and like Moore's plot becomes a big part of advertising especially within the industry
00:14:56.080then there's a shift where it's sort of mathematized by this guy at IBM named Robert Dennard
00:15:01.260and made more into like sort of a it's scientized it and it makes people believe in the future of
00:15:07.080silicon which again one of the ways in which moore's law remakes our understanding the past
00:15:11.660is that silicon as a medium for transistors as a medium for computing is taken for granted but at
00:15:17.700the time it was not at all clear right that this would work right technically and so moore's law
00:15:23.480is a part of selling that the mathematization of it and like sort of the the kind of work that
00:15:28.720carver made the scientist at caltech who also becomes a big proselytizer for moore's law and
00:15:33.700for silicon as a medium like takes up and makes very popular so it becomes very popular and then
00:15:39.280the late 70s japan emerges as this really impressive like technological power and one0.90
00:15:45.360that's like threatening core industries in the u.s including electronics and so there's you know
00:15:51.600like the japanese model which is quite different from american capitalism does two things right it
00:15:57.060causes this crisis for american capitalism you know like there's the 70s question of like can
00:16:02.020we grow forever? But then there's also the question of, is the American model the right
00:16:05.760model if the Japanese are overtaking US industry? And then within the industry, okay, so in order0.91
00:16:11.600to beat the Japanese, we're going to have to come up with a different way of doing things1.00
00:16:15.020that includes a lot of planning. So individual companies start planning, then consortia start1.00
00:16:19.680planning. And then eventually, by the time you get to the first Bush administration, and then
00:16:24.180Clinton, you get these national planning efforts. And they use things like Moore's Law as a model
00:16:30.520originally, and then eventually just call what they're doing Moore's Law. And they build it into
00:16:35.820these roadmaps that coordinate the entire industry, suppliers, like the scope keeps building and
00:16:40.680building, using U.S. power to do so into the early 2000s when things start to fall apart for various
00:16:49.620reasons, like around 2006. Why did things start falling apart? That's a question that many people
00:16:55.560are trying to answer and there's so many um different kinds of answers the focus of my
00:17:01.860research so one of the things that happens right is the separation of like the design and the
00:17:07.620manufacturer model apple went to intel actually to manufacture chips for the iphone right but
00:17:12.640intel thought the money we make per chip is just not worth it why don't we we'll send you to tsmc
00:17:18.200they end up working with tsmc um and that's that creates this separated design you know that like
00:17:24.080or makes it clear that this model works.
00:19:32.140It's like, oh, this is how technological development happens.
00:19:35.420And, you know, like a lot of these same policy tools or, you know,
00:19:38.700like we just generalize it out in a way that like gives technologists,
00:19:43.460especially like consultants, right, about tech, this power.
00:19:46.860I mean, there's also movement within economics, right, to say, okay, it used to be just like growth is capital plus labor. And it turns out, you know, it's obviously more complicated, but that's the model until like, I guess like mid-century, when all these economists start thinking about, and actually like some other like management consultants, proto-management consultants, started thinking about technology having a role, sort of playing the, why some economies do better than others.
00:20:14.380So there are a lot of parallel things happening, but it ends up locating that in tech.
00:20:22.400And this generalized idea that Moore's Law is like a model for how technology works as a whole gets us to a whole set of ideas that I'm perpetually fascinated by.
00:20:34.820Like, you know, this is where you get someone like Ray Kurzweil.
00:20:38.580Yeah. Or you get this way of telling human history that starts with like fire. And then, I don't know, I'm going to butcher this. Fire, like the steam engine. And then, you know, you can tell these modeled stages of history, which again, there are precedents for, right? Like there's, you know, stages history in terms of like feudalism, capitalism, socialism, or communism, but they sort of gets adapted to this linear model too, right? But where technology is the agent of history.
00:21:02.480And you see this over and over again in the way that people in Silicon Valley talk about the future. They just talk about, oh, we're going to have this technology and this technology and this technology, and it's going to solve all of our current problems.
00:21:16.260And I don't want to say I'm confused by it, I'm just amazed that they believe that you can solve every single problem with technology. There's so many obvious counterexamples to it, I don't understand how they, I mean, because some of them are grifters, but some of them really believe this stuff, and it's very hard for me to understand how they can do that.
00:21:39.160intent is kind of a hard thing because it's like oh where do my interests end and where do my
00:21:43.300beliefs begin and yeah exactly i don't know i can't get in someone's head but the weird thing
00:21:48.380that it does right is it's not people who are agents of history right the right history is
00:21:53.700driven by technology and there are people who can speak for technology and what technology needs
00:21:58.780you know at first it's like physicists it's like other people but now it's like you know these
00:22:02.400entrepreneurs but but it's the technology that's acting and it's particularly for me it's like a
00:22:08.740pet peeve about the way people talk about AI. It's like AI is speeding up my work. It's like,
00:22:13.720no, your employer is speeding up your work, making you work more for less. Or like AI is like
00:22:19.380inevitably going to do X, Y, and Z. It's like, no, these are political decisions made by employers,
00:22:25.200bosses, regulators, politicians. People are doing things with technology. Technology is not acting
00:22:32.560on its own technology and especially ai can be the medium through which people experience these
00:22:39.680kinds of political and social relationships they're displaced right but it is not the cause
00:22:46.860of them and i think i think that confusion is something it's it's like worth changing the way
00:22:53.340we talk about technology so it's not like the technology acting there there are let's locate
00:22:58.620the specific people who are making these kinds of decisions or not making any decisions, right,
00:23:03.160as the case may be. Yeah, no, when you talk like that, it absolves the individuals involved of
00:23:10.100any sort of responsibility and just saying, oh, no, it's the technology. This is what the
00:23:14.160technology is doing. Well, and back to what you're saying, it's like it defers questions
00:23:19.380or it subsumes, it hides questions that are really political and social questions,
00:23:24.440where like technology is the agent, is the actor
00:23:28.200in a way that I think like makes it really hard
00:23:31.100to do the kinds of politics we need to do
00:31:31.760Quantum Foundations is basically the area of research that tries to figure out what it is that quantum physics is saying about the world.
00:31:43.400And it has turned out to be like one of the weirdest things in the history of physics is that that is a really contentious, like knockdown, drag out fight that has been going on for over a century at this point.
00:31:57.760and weirdly it's like the subtext of a lot of history of computing like wheeler is really
00:32:03.140invested in but not just wheeler like carver meat to this day when i interviewed him like
00:32:07.280a couple years back he was still very focused on questions of quantum foundations like as
00:32:12.520these are the big questions of our age now this is not something that the general public
00:32:16.200spends a lot of time thinking about but i think this is the preoccupation of this generation of
00:32:20.680scientists because it is i mean it's such a weird counterintuitive thing it breaks physics purchase
00:32:27.080on the future but also it does this really interesting thing to the scientific method
00:32:30.740which is like no longer are observer and observed separate categories and in a way like social
00:32:37.440science is dealing with that on its own terms like i think people in my field think that's an
00:32:42.100original insight it's not it's like really something like you know the quantum physicists
00:32:45.980are like wait what do we do about this yeah like this is a huge deal you can't have this neat
00:32:50.780separation you can't have these neat experiments and so what do you do do you remake the scientific
00:32:56.080method around this insight they don't but they don't know what to do with it right they don't
00:33:01.440know what it means they don't know what to do with it it's this huge intellectual crisis happening at
00:33:04.660the same time as all these like material crises are hitting the field and these political crises
00:33:09.580are hitting the u.s like it's a crisis of u.s capitalism it's like all many things happening
00:33:14.140together right like the limits to growth discourse comes out at the same time it's just like
00:33:17.820so many things are in flux i'm so glad you brought that up because i really want to talk with you
00:33:51.640okay, what is a measurement and what is going on in the quantum world? And we are not going
00:33:57.620to get into exactly what the measurement problem is on this podcast episode, because that would
00:34:01.840take another hour. And if you want to know about that, you can, I don't know, go get my book or
00:34:06.580something. But it's stake, right? Is that, is the universe deterministic or are there elements of
00:34:12.300uncertainty? That's the stakes, right? I think. That's part of the stakes. Right. Yeah. Part of
00:34:17.500the stakes is like determinism versus randomness at the fundamental level. Part of it is, you know,
00:34:23.440can physics be done without talking about the physicists? Part of it is, is physics about
00:34:32.640the world as it is? Or is physics about doing physics? What is it that we're doing when we do
00:34:40.860physics? There's a lot of really deep questions at work. Yeah, it's fascinating, right? The observer
00:34:46.100observed alone is like a huge, huge, like you can't separate the measurement from the measurement
00:34:50.460instrument. Exactly. And then Everett comes in and says, oh, one way of solving this problem
00:34:57.100is, you know, which, which weirdly is a way of championing the idea that physics is really about
00:35:04.960the world as it exists, rather than just being a kind of tool for issuing forth predictions.
00:35:12.020Everett says, yeah, one way that you could do that is you could say that quantum physics is about the universe splitting off into a bunch of, you know, separate individual universes.
00:35:23.560And Wheeler likes this for certain reasons, but is also nervous about this for other reasons, including his, Wheeler's relationship with Niels Bohr, the granddaddy of quantum physics.
00:35:38.320And long, long story short, again, this is all in what is real.
00:40:06.140But he's not, yeah, no great men of history.
00:40:08.760Because otherwise it gets stuck in the sort of like, okay, then we should empower great men to do, you know, whatever it becomes.
00:40:14.440Well, yeah, well, I do think that that's part of going way back to what we were talking about at the beginning.
00:40:18.680I think that the seductive power of the great man theory of history is something that's connected to the sort of rocket ships and chrome final frontier idea of space.
00:40:32.200you know we can we can go out into space and there the great man can you know do what it is
00:40:37.980that he's going to do and blah blah blah blah and like yeah but none of that's real yeah it's the
00:40:42.760figure you know the entrepreneur is like shumpeter's new man but then he becomes and it's
00:40:48.420often he let's be real um yeah uh no no i mean just in their imaginary right yeah yeah exactly
00:40:55.400they always imagine it as a dude yeah exactly very often right and so it but it becomes really
00:41:02.160useful right there's a reason why palantir wants to retell the entire history of u.s industrial
00:41:07.480production in terms of guys like them right they're totally historically revisionist it's
00:41:12.660something that i wrote about in that new york review of books article but they do they you
00:41:16.780know they're rewriting history and i think that they're very attentive to history we always very
00:41:22.900attentive to history the fact that as you know as a society we sort of take this as like a soft
00:41:28.200you know soft science and not stem it's a mistake right because everyone else is focused on history
00:41:33.180and there's this idea that like studying histories and futures are separate but they
00:41:37.020kind of remake each other you know what i was getting at with moore's law right the
00:41:40.680the power of like this this sort of way of thinking about the future also has remade our
00:41:48.600ability to think about the past in certain ways that are really interesting and it happens
00:41:53.240reciprocally too right like if you think about history in terms of stages of development where
00:41:57.300technology is the primary actor where great men who speak for technology are the primary actors
00:42:02.340where you get something different about like what you should do in the future or who who controls
00:42:07.180the future who can predict the future and these things have performative effects like if you
00:42:11.560believe the world to be a certain way and you act on it the world may become more like it may become
00:42:16.840less like it but anyway yeah so that's that's my investment in history why you should become a
00:42:22.100historian. So I'm going to, I'm going to take a stab at explaining what a quantum computer is.
00:42:27.840And then I want you to explain why you see quantum computing and limits to growth is connected
00:42:33.860because I, I just, normally that's a question that somebody would ask me. And I love that I
00:42:39.380get to ask you this question. So quantum computing is this idea that we can use special quantum
00:42:45.740properties, like particular quantum properties of matter to perform computing in a completely
00:42:53.640different way where instead of having bits that are either zero or one, you have quantum
00:42:59.900bits or qubits that can take any sort of state between zero and one.
00:43:07.140And you can combine them and connect their states using a phenomenon called quantum entanglement
00:43:14.420to produce, how do I want to say this?
00:43:19.760You can use all of these quantum properties
00:43:22.440to gain certain kinds of computational powers
00:44:09.340besides like screwing up internet encryption for everyone what it would do it's not an obvious case
00:44:15.640for it but because it's you know what it does really well is it factors large numbers it's very
00:44:19.380good at the factoring number which is the basis of internet encryption through this algorithm
00:44:23.080called rsa now there are post-rsa algorithms that work well enough and they're being implemented in
00:44:28.520certain places but you can see how this is a concern for you know government agencies and
00:44:34.040yeah whomever but it you know it gets conflated with like oh this will be computing 2.0 this
00:44:39.780will be and there's a moment where you could believe that right like there's a moment where
00:44:44.820like silicon isn't hegemonic in the 80s where they're like okay maybe like heat diffusion will
00:44:50.360be a problem maybe like silicon won't scale maybe we need an alternate model and you know everything's
00:44:57.280getting smaller anyways we need we want something like more is law to continue like more is law
00:45:02.500helps you overcome the idea of limits to growth right and and so quantum computing is an answer
00:45:07.240to all these intersecting problems in the 80s looks like it might be able to do all of these
00:45:12.280things at once oh i didn't know people were talking about it like that in the 80s yeah well
00:45:17.700because the 1981 conference co-sponsored by ibm and mit at endicott house like the advertisements
00:45:25.200are about basically yeah like computers are getting smaller we don't know if more's a lot
00:45:30.580like the moir's law can continue but maybe this is the vehicle for it um this is maybe the natural
00:45:37.000progression of compute computing and you know there's all these physics elites you know it's
00:45:43.300a bunch of weird people too but there's enough physics elites that buy into this and are
00:45:47.160interested at the intersection of computing and physics like Feynman like Wheeler like Freeman
00:45:51.400Dyson they're all at this early conference not to say that there is plenty of really fun weird
00:45:56.460people there too or like ed fredkin who's a fascinating character between legitimacy and
00:46:01.760not so legitimate stuff and and some of the research that comes out of that right so the
00:46:07.0401981 conference convinces fineman that reversible computing or like he has he has a talk with
00:46:14.660uh charles bennett yep that convinces him that reversible computing is real and then he starts
00:46:20.420doing this really interesting course with um john hopfield and carver mead at caltech about like
00:46:28.620the future of computing basically like different models of the future of computing and out of that
00:46:32.320com carver mead is sort of the avatar for more is on silicon transistors hopfield is beginning
00:46:37.820of like neural nets and Feynman obviously has this this i mean people i think people i don't
00:46:42.500need to gloss Feynman but but that all comes out of that first that quantum computing conference
00:46:47.500And these discussions about quantum foundations and the future of computing are all happening at the same time with the same people.
00:46:55.240I didn't know that people were talking in the 80s about quantum computing as a way of possibly getting around limits to growth and continuing Moore's law.