Balaji Srinivasan and Glen Weyl on Decentralized Governance, Politics, and History
In this segment of the fireside chat hosted by Erik Torenberg, Balaji Srinivasan and Glen Weyl critique the nation‑state as an outdated governance unit and argue that many public‑goods domains—data governance, open‑source protocol stewardship, and environmental management—are better served by decentralized, border‑agnostic structures. They explore technocracy, questioning whether its failures stem from misaligned incentives of technocrats, insufficient information flow, or both, and suggest mechanisms like retroactive voting to improve accountability. The conversation turns to the challenges of large‑scale empires, asserting that scale creates coordination dis‑alignment, making nimble micro‑states (e.g., Singapore, Estonia) more effective in the 21st century. They advocate for a "network‑of‑networks" model that builds governance bottom‑up rather than imposing top‑down imperial structures. Balaji highlights the entrepreneurial pathway that channels ambitious, rule‑breaking individuals into pro‑social creators, contrasting criminal rule‑breaking with entrepreneurial innovation. The segment concludes with calls to action, directing listeners to nakamoto.org, Radical Exchange, and Balaji’s book, encouraging participation in decentralized movements.
Chapters
This week on Upstream, we're releasing a conversation between me, Balaji Srinivasan, and Glenn Weil from the Village Global Podcast in 2019. We discuss politics, nation-states, Voting, and more. Please enjoy. Thank you everybody for coming. I'm glad we're starting promptly at 5:30.
Thank you everybody for coming.
The original plan was just to do with Glenn, but sometimes you think you're going to a Drake concert and Kanye comes. And Symbology had something, and so we thought we'd bring him on stage too. So, first off, thank you to Laura, Austin, and Longevity Fund for hosting us. Glenn, do you want to give maybe a 1-minute description of RadicalxChange?
Yeah, and it evolves all the time, but I think that right now I'd say the goal of RadicalxChange is that our political economy is sort of very old compared to the way our technologies have been advancing. We've moved from the telegraph to the telephone to video conferencing, whatever. And yet, like, the way we actually transmit signals of, like, value and identity across long distances has, like, hardly advanced. And RadicalxChange is trying to envision, build, design using rigorous academic things, art, etc., ways in which we can have much more powerful political economies that enable democracies that actually make sense and property regimes that actually make sense.
Awesome. So we're going to do a fireside between us 3, maybe for around 30 minutes, and then we're going to open it up for Q&A. I was looking at a lot of Glenn's interviews and was looking for a unique take, and I realized I didn't have one. And that's when I called Balaji and said, hey, let's do something, all of us together. So I thought I'd start by taking a stab at where I think you 2 agree and where I think you 2 see things differently, and then I'll either let you respond to it or just jump into a question for Glenn. So I think that you agree in your distrust of centralized systems, whether they are big governments or big corporations. I think you agree in the power of open source and crypto, although, Glenn, you have more skepticism there. You agree in the need to redefine how we think about identity. You agree in having an ethos of being radical and experimenting. And you agree with sort of the broader ethos of win and help win. Where I think you disagree is mostly on tactics. Glenn, I see you as wanting to abolish private property as we currently think about it today, and I think Balaji is more open-minded. Glenn, I see you more skeptical of crypto while you're excited about many of the cultural and community elements of it, and Balaji is less skeptical. Of crypto. Maybe implementation details of how you think about identity. You're both excited about new cities broadly. Glenn, I see you as very against CEO-driven societies, and Balaji is open to those among other options. I see you as— while you agree on decentralization in principle, I think Balaji is more open to a more explicit meritocracy that assumes different interests and abilities. And Glenn, maybe that— rubs you the wrong way, or maybe you see that differently. Broadly, Glenn, I see you as optimizing for justice and reducing inequality, and Balaji perhaps more overall prosperity, or maybe a different way of looking at it. Are there any reactions to how I presented it?
I think you frame the differences between us in a very libertarian-ish type of a way. So I would frame them maybe a little bit differently, and I'm interested in Balaji's reaction to this. I think what I've increasingly come to think of is that one of the things that's at the core of my emerging belief system, which didn't exist at all in Radical Markets, is a much more sociological view of individuality than exists in a lot of the economics tech imaginary. It's one where There's sort of historically individual animals. There's then social humans, like a group of humans, and we only get to individual humans like in the modern era. And the way we get to individual humans is not by like removing social structure, but by adding more social structure. So that like what actually creates the potential of being an individual is the plurality of human organizations and the diversity of human organizations that really get created in the modern era. And that attempts to sort of use something like the blockchain, or the nation-state was before that, to like strip away community and like reveal the true individual are fundamentally confused. And that it's by creating a diverse range of communities that we actually enable people to be individuals.
So I would say, actually, I agree with a lot of that, but I would say, um, I also, I mean, I would quibble with some of your points or whatever, but I think the meta thing is I would hold as the highest value choice. That is to say, if you choose to be in a society that has a centralized government and that works for you, wonderful. If you choose to be in basically a commune and people could actually make that work, wonderful. So long as you have the right to exit, the right to leave. That's actually part of the UN human rights kind of thing, by the way. That is, like, the right to emigrate. It doesn't mean you have the right to immigrate. Like, any society, you know, like Harvard does not have open borders. Neither does the New York Times. You can't just go and walk in and get a job there. You can't just go and gain admission. So, not every institution has open borders, but the right to emigrate, I think, is a fundamental right. Given that, and given a choice of many jurisdictions, then I think within those jurisdictions, you can try 1,000 different experiments. People are not fundamentally harmed. They can leave and they can choose. And that's my highest value. And all those other things are subsidiary to that, where I think, here's how I'd optimize my optimal society, but I'm happy to try other experiments. The other thing I'd just say is actually a lot of Glenn's ideas, you know, I love the book. You know, I haven't finished all of it. I'm going through all the papers one by one. And I'm sure I'll have some mathematical comments, you know, and you'll tell me I'm totally off. But I love the fact that we're talking about these things. things, that we're in a point of intellectual ferment where you can do empirical political science, you can do empirical macroeconomics. That was not possible before, not even like 5 or 10 years ago. And so we're really just starting in the last 5 years. So that's an amazing moment for us. So I'm glad to be here for that.
Yeah, I think that's absolutely right. I think in terms of differences, Albert Hirschman, a favorite person of both Balaji's and mine, and a hero of mine, Talks about voice and exit. Bology puts a lot of emphasis on exit. I put a maybe more balanced and maybe more importantly intertwining emphasis on the two of them. I don't view there as being such a thing as meaningful exit without the simultaneous exercise of voice. And I don't view there as usually being a meaningful notion of voice that isn't coupled with elements of exit as well. So I don't view these things as separate or alternative options. And almost anyone who's focused on exit as the solution, I think, is missing what's crucial to exit. So let me give an example. So a lot of people would say, well, if I'm being oppressed, if they won't let me use weed, use some service, talked to these people, you know, I should be allowed to exit. But the thing is that it's almost never meaningful for you to exit completely on your own. It's only really meaningful for you to exit with a group of other people. And in fact, for almost any meaningful desire that you have, there's almost certainly going to be another group of people who are like in solidarity with you in that desire, if it's actually a meaningful desire that you have. So The opportunity to exit is really much more meaningful exercised collectively. And if you exercise it collectively, it has to be part of a democratic collective decision that all of you are making to take that decision to, in some way, exit. And conversely, in order to really effectively exercise voice, there has to be an element of forming a community that's separate and apart from the rest of the community in order to do that. Because as an individual, one drop in a vast ocean, your voice is not usually very meaningful. And that's one reason that people doubt whether voice is meaningful. But again, usually it's not just you as an individual, it's you forming a community of people who are collectively exercising that voice and threatening to potentially exit if you don't have that voice listened to. And it's in that moment that your voice actually becomes meaningful. So I don't view voice and exit as these separate things. I view them as actually only effective when combined with each other.
So actually, I totally agree with that, and I could make a few points on that. One is that I think in the talk, and I titled it, or the old talk, Silicon Valley's Exit, simply because exit was sort of an underdiscussed kind of thing. But I agree with a few points. One is I think exit amplifies voice in a democracy because it shows that this feature that you care about so much you're actually willing to leave, and that forces reform. For example, Stripe actually leaving San Francisco, everyone's like, oh, it's real, a $30 billion company's actually moving. And then, you know, like, people realize that they put an infinity multiplier on their UIs, in your terminology, was very high at least.
Yeah.
The other thing is, I've given some interviews in the past, and maybe, I've got a paper coming out on this, You know, like, you may have written up something, and if so, then like maybe parallel evolution. But I call it CrowdChoice, which is coordinated exit. And the idea being that right now we treat our ballot as just something to be aggregated. So, you know, you have a preference vector, you have this position on, let's say, abortion rights and this on gun rights and whatever. And then that is aggregated with other people in a jurisdiction. And then you have to, in the simplest form, just, you know, submit to the majority rule, right?
Yeah. Okay.
And QV is a great thing for maybe fixing that part of it. But let's say you do lose. Well, my thought is you treat your ballot not as just something to be aggregated, but as a search query across all the jurisdictions of the world, and you rank order them by which jurisdiction most satisfies your preferences. And then you group with the other folks, and you essentially bid with that other community, and you say, look, if we all move there, your tax rate is currently 10%, can you knock it down to 9.5%, and we'll all move there? That's just the simplest kind of example. And then these guys will calculate and be like, okay, I can discount it, but I'll make it up on volume. Boom, it's win-win, let's do that. And I think that concept of crowd choice is going to be one of the biggest things that's enabled by kind of the next generation of social networks. It's kind of part of what I'm working on next.
Have you read any of Ada Palmer's work?
No, I should.
She's a science fiction author, and she has a great series which is set in a world where they've invented the ability to basically have like flying cars that are very fast so that you can very quickly move across physical jurisdictions. And at that point, it starts becoming just meaningless to have geographically bounded nation-states. And so what ends up happening is people sort of subscribe to different nation-state services, which then sort of negotiate with each other the terms on which their members will interact with each other. And people are sort of members of one or more. Anyways, it bears some resemblance to what you're describing.
Absolutely. And I give actually 4 immediate jumping-offs of that.
Yeah.
One is I think mobile is making us more mobile, and law is a function of latitude and longitude. So, if it's cheaper to change your X and Y location, it's cheaper to also change the law under which you live. VR is ramping also, and when that really works, then you have like 2 really material ways to opt out of your current location, just by like jacking into this virtual world or actually physically moving. And there's like, those are kind of complementary axes. The third is I think in the long, long term, nation states eventually become commoditized providers of land and rule sets. Like, they've got coal, they've got uranium or something like that, and they've got the rule sets. But otherwise, you can kind of pick among them. And then finally, I think the true borders of a community are no longer the geographical borders, but the ideological borders in social network space. Like, borders are a social construct.
Mm-hmm.
And unless there's genuine agreement, like in the sense of you can't just do some physical test and say, OK, this is the border of you know, the US and Mexico or Spain and Portugal, there is something in people's heads. And you need to be able to give that ideological argument as to where my community ends and the other one begins. And that, I think, is where the true borders are going to be, not geographic.
So I think that's— yes. And I think that there will always, as a result of that, be a new emergent cluster of overlapping different types of communities where voice is necessary because you won't just want to split off. And so the way I think about it is flexible, emergent collective governance, not either individual governance or rigid, historically determined nation-state democracies.
Absolutely. And I think actually something that contributes to that is it used to be that one was, for example, a card-carrying member of the Communist Party. That is to say, membership was discrete— 0, 1. You were in a group or not. You know, if you hang out on a Reddit forum or, you know, there's some community you're part of once in a while, there's a whole continuum between like hourly active user and like yearly active user or what have you. Like there's a continuum of group memberships, right? And then that means you've got optionality, you've kind of got a foot in a bunch of different waters, and you're no longer just like locked into one place.
Yeah, and I view that as sort of like a duality. So coming back to this question of What is it that defines the individual? I think what modernity offers is the possibility. There's this sociologist, Georg Simmel, who's really influenced me over the last year, who says that you can think of all the social circles that you're a part of. And he says in premodern society, all of those social circles, your work circle, your family circle, your love circle, all these overlap very heavily.
Mm-hmm.
And so basically, not really an individual because there's so many other people who all their circles are basically the same as you. And then what starts to happen in modernity is these things start to separate out. So you start to get like your work circle only overlaps with your like home circle with you and 2 other people. And then, you know, all this sort of stuff. And so then you end up being like the unique intersection of all these different like social groups that give you meaning. And it's like that that individuates you.
Interesting.
And so you can then think of yourself as almost having like a memetic code, which is made up of precisely these quantitative measures of your commitment to these different communities. And what makes you unique is the fact that that memetic code is not the same as anyone else's memetic code.
So that reminds me of a few things. One is this concept of 33 bits of information, like 2 to the 33rd. Given 7-odd billion people, 2 to the 33rd is about 8 billion. So with 33 unique bits of information, you can de-anonymize somebody. So sort of like this, right?
Yeah.
The other thing is, and this sort of points in 2 directions, but, you know, over the 20th century, we had, for example, zoning of commercial, residential, and industrial, you know, because industrial stuff was polluting and had smokestacks and so on, and you didn't want that near the residential. And now we're seeing, I think, one trend and then a second one. One trend is to come together in like work lofts, like the one we're in, where kind of the work and residential are sort of blended, the commercial and residential, the industrial and residential. But the second trend, and this is something I'm sort of spotting, and maybe it'll happen, but I think it's kind of happening, is I think you're going to see a separation of people's earning names, their speaking names, and their real names. Because if you speak under a name, you can get attacked for that. And so folks, just like you don't keep all of your money in one place, you sort of separate your reputation among multiple names. And then that kind of leads to a sort of a schizophrenic fracturing in some ways of this, where this name is part of these N communities, and this name is part of these M communities.
We'll continue our interview in a moment after a word from our sponsors. If you don't already subscribe to Turpentine's industry-leading newsletters, like our new daily AI newsletter, Emergent Behavior, or Media Empires, you should. But that's not what I'm here to tell you about. The platform we use to power these newsletters is called Beehive, and it's excellent. First of all, it was started by the same early team who helped build Morning Brew into a $75 million newsletter business, and they built Beehive to offer that same powerful functionality to anyone sending emails. From essayists to business owners. The platform is beautiful. Their text editor is intuitive, and they help you scale your audience with custom growth features. Beehive has powerful tools to help you monetize your content. You can easily launch paid subscriptions or pursue an advertising model. The Beehive platform will even connect you to premium brands to sponsor your newsletter. Not only do we use them, but thousands of the top newsletters in the world also use them, like Milk Road, Blockworks, the Lindy Newsletter, and so many more. Beehive's founder hooked up Upstream listeners with a sweet deal. Get 20% off for 3 months with code Upstream. Visit beehive.com, that's B-E-E-H-I-I-V dot com, to get started. Fast forward to the end of 2024. Think about your goals. What can you do right now to give yourself the best chance of succeeding? If learning a new language is on your list, you absolutely need to check out Babbel. Babbel offers a range of learning tools: self-study app lessons, live classes, and even podcasts, which have always been my favorite way to learn. Studies from Yale, Michigan State University, and others continue to prove Babbel is better. One study found that using Babbel for 15 hours is equivalent to a full semester at college. Babbel isn't just a game to kill time and make you feel like you're learning a language. It's not overly academic or rigid either. It's all about learning language for the real world. Babbel stands out because it's designed by real people using a modern conversational teaching approach. It's not always easy. Nothing worth doing ever is, but it's straightforward and designed to help you start speaking in just 3 weeks. With Babbel, I was able to brush up on my intermediate Spanish to ramp up for travel to Argentina last year and was able to set clear goals based on how much time each week I wanted to practice. Join millions of Babbel language learners across all age groups. Here's a special limited-time offer for our listeners. Right now, get up to 60% off your Babbel subscription. But only for our listeners at babbel.com/econ102. Get up to 60% off at babbel.com/econ102, spelled B-A-B-B-E-L dot com slash econ102. Rules and restrictions may apply. Glenn, you've been writing about identity a lot recently. How are you sort of? How do you respond to that? Or how are you sort of viewing what the?
Well, I actually think what Alaji said is quite consistent with. I need to reflect on it more, but it's quite consistent with. Some of the things that I've been writing recently. The idea is that what constitutes you is, let's say, the list of all bits relevant to you, the emails that you sent, the places you were, etc. But as Balaji said, that's unbelievably redundant identification of you. Imagine that there's a billion bits pertaining to you, or at least a million, and any 33 of them is enough to uniquely identify you. What that means is that you're redundantly identified by like 5 orders of magnitude or something like that. And what that means is that in principle, there are like 5 orders of magnitude of people that you could be and still be unique, right?
Yeah.
And so like that, that's like a huge space of potential personalities that you can have. while still being completely not overlapping with anyone else. The possibilities that opens up are really quite astonishing because it allows for a flexibility of democratic organization. Democracies depend on this notion of unique people. But given that there's all these different notions of uniqueness, the number of democracies that you can participate in potentially I want to actually put my finger on that because the democracies depend on unique people.
You realize this if you're running any kind of company that has any kind of user accounts, any money at stake whatsoever, because if people can set up multiple accounts, which people will try to do, you realize, okay, I need an API to determine whether this person is real and unique. And that recurs in lots and lots of companies. Lots of crypto companies have that problem. It's related to kind of a broader concept I have, which is The libertarian founder rebuilds the state. Okay. So meaning, you know, many, many folks who start companies are small L or large L libertarian, and they leave, you know, their— the confines of their stifling bureaucratic environments, just go onto the frontier, go onto the tundra. And many of them just die, and they have to come back. Okay. But those who succeed kind of build a lean-to, they start recruiting more people, and then eventually they build a larger organization. And then there's a critical flip where the most important number goes from your burn rate to your bus number. You know, like the number of people who— you know, everybody needs to pull their weight and be exceptional at the beginning, and then suddenly everybody has to become replaceable, otherwise you're not doing your job as CEO. So you set up a bureaucracy that's bigger than any one person, and then you start realizing, oh, we need to uniquify all of our users. You basically start rebuilding aspects of the state. You want NSA-like intelligence on every single one of your users because some of them are committing fraud. some of them are harming other users, etc. And then conversely, of course, on the user side, you want to not be spied on at all. You want to give the minimum necessary kind of thing. And so essentially being both system administrator and user is sort of like being both professor and student or both VC and entrepreneur. You kind of see both sides of the table and have empathy for both.
Yeah, so this is the reason why what I think is so wrong about the standard libertarian view of the world is that it fails to realize that, like, It is precisely, I think, sort of the combination of— you could call it private property or this extreme individualist thing— with the fact that we live in a society that's fundamentally social that actually creates precisely the conditions that libertarians are objecting to. Whereas if you actually start from the foundation of recognizing these social patterns and actually try to build for them, Then that's what opens you up to the possibility of this fluid, overlapping, emergent identity that actually gives you the chance of being an individual in the first place.
It's interesting. You have a moral critique of it. I'm not disagreeing with it necessarily, but I have a utilitarian critique of classical libertarianism, which is, as insightful as libertarian theory is on incentives, it massively understates the incentive for initiating aggression. That is to say, The non-aggression principle basically says, OK, no first strike. But if somebody first strikes you and knocks off your head, then you can't even retaliate. And so as such, the ideologies that are preponderated are those that have some aspect of first strike. Communists were definitely willing to first strike, and so on. So if you're just premised on retaliation, you're not going to spread. That's kind of one critique. The second critique is basically that, as with the crowd choice thing that we were talking about, or coordinated exit, You know, libertarians are all about live and let live, and live and let live, I would say, just from a Nietzschean utilitarian standpoint, isn't a winning ideology. Win and help win is actually, I think, more in line with at least how I think about the world, which is that rather than a bunch of folks sitting in, you know, their home with guns— that's a fundamentally defensive thing. Galt's Gulch is a retreat, it's a surrender. Whereas a tactical, like, putting— placing your foot back to then drive forward with more energy, That's a little different, right? That's like, OK, we'll move to America, build up power there, and then come back into Europe like 100 years later. That's a little different, you know?
Uh-huh.
And I think that's where I'd say sort of the dogmatic form of libertarianism leaves me cold, actually, because it doesn't attend to some of these practical kinds of concerns. If the other guy is aggressive, if the other guy is cooperating, he's going to win. And so you need an ideology that gives you the maximum possible freedom that is consistent with cooperation and deterring first strikes.
Interesting. I don't know if Glenn is only making a moral argument.
No, no, no. In fact, most of my arguments have been utilitarian. It's just I tend to make moral arguments these days because I think they appeal to a broader audience than the utilitarian arguments that sort of are my job to make at some point.
Right. And an observational one.
I didn't mean it in a bad way. Yeah, yeah, yeah.
But unpack it a little bit more. Glenn, when you find that you are convincing people of, oh, I'm not just an individual, or, oh, private property isn't as clear-cut as it may seem, What do you find is most convincing, or how did you get there yourself?
What's been most compelling to me in the last year is the fact that if you just think about almost all the consumption that we actually do, almost none of it is the sort of private goods on which standard economic theory is premised. And almost any economist would actually admit that. So it's like we consume things in a family, like most of the things that like make it nice to be in San Francisco are not features of my individual home, but features of the ambient environment around there that make it desirable to live there. You can just tell that because you could go get a great home 3 hours from here and compare it to a hole in the wall here, and you look at the prices. Most of the features of what make me be able to live the comfortable life I do are a function of the fact that I live in the United States. So 2/3 of income inequality is across countries. So you just see all these things and you realize, like, we are massively socially overdetermined. There's just, like, no possible way that we can sort of, in this linear way that, like, private goods wants to do, attribute back to the individual, to a sum of individual elements, like, this emergent thing that is sociality. And it's like, once you think about that, you just realize that everything that is supposed to be about why capitalism works great and the utilitarian account of it, it just doesn't add up. It's not what actually could be driving what's succeeded about the way that we're organized. Once you move past that, then the whole edifice that lies behind utilitarian economics-based justifications of market capitalism just kind of goes away and you start realizing there's this enormous complexity of the system that we're actually dealing with and how can we start to match that in the formalisms that we have to capture it.
I'm actually more sympathetic to this, by the way, than you might initially thought, but I think from maybe a different angle, which is as a system administrator, so you develop a new social network, you are essentially the one who has created the land Yeah. And the usernames, for example, twitter.com/jack or /jim, that's valuable virtual real estate that you've created. And you determine whether it can be sold, whether it can be resold, all of these kinds of things. Moreover, if you develop any kind of like two-sided marketplace, you know, so like Earn, we had both of these things, you find, okay, these small parameters that I'm setting centrally radically affect how much people are using it, whether they're sending it back and forth. The defaults really matter. It's not, quote, truly a free market. There's not, quote, truly private property, because I'm setting these things centrally. In a sense, Bitcoin may be the first private property ever. That's a bit of an exaggeration, but in the sense of there's no system administrator who could just deprive you of it. Now, there's personal property, of course, and there's private property in the sense of that which you can defend with your own physique or whatever, your own guns.
Yeah.
But I think once one embraces the idea that, okay, the Fed has system administrator access, it can freeze your account, it can move these things around, the state can do all these things, your company can do these things with user accounts, then you realize, okay, at least for those folks, those actors, I think a lot of your things are very applicable, the auction everything kind of model.
Mm-hmm.
I think it's useful, though, to have a counterbalance against these system administrators, by which I sort of collapse both the tech CEO and the You know, the Fedwire administrator, with something like Bitcoin, which then protects the rights of the individual against some kind of arbitrary thing like this.
One thing that I found really interesting about what you said is, like, so in econ, we have, like, all of these models where there is, like, the global administrator, the state, the social planner, whatever. And then there's, like, this market process where there's— everything's a private good, there's no increasing returns, people are just these atomized market participants. And, like, that's the image. But then it's kind of weird because actually that model is used to model the state. It's used to model individual countries. It's used to model the whole world. It's used to model platforms. We use that exact type of model for platforms. It's used to model companies. And you're like, hmm, there's like 100 different entities at 100 different levels of resolution, which are all being modeled by this global, incredibly important like I'm setting all the parameters and then everyone's atomized underneath me. Maybe the actual model features tons of layers of complex social organization, but no, we never look at that. We never think about how do I optimize at one layer for the fact that there's going to be rich, complex, and diverse social organization underneath and intertwining with me. It seems to me like that is the fundamental thing that's missing. I just wrote this thing I'm about to put out where I call that ideology alone. atomized liberal and objectivist naive epistemology. It's like a view that there's like— it's a view that there's like, there's like the objective truth or the state or the center out there, and then there's the individual, and there's like nothing in between. And that's just like such a confused view of the world that like it can't even stay consistent with itself within the same field.
It's interesting because, uh, towards Like someone at— I shouldn't say at Right Angles, but related to that, something I've thought about is if you take Uber and Lyft, I think of them as really competitors in some ways to taxi regulators, and their drivers as competitors to taxi drivers.
Yeah.
And so the real competition is actually between a regulator-driver fusion, Uber, and another regulator-driver fusion, Lyft, where these two provide star ratings and And they regulate their platforms. They're actually better at this than the taxi medallions, which only have maybe an annual inspection or what have you. The GPS track every ride. They make sure that people can pay. They resolve disputes very rapidly. They've got lost and found services. They have bidirectional ratings. All those are far superior to the services that the taxi medallion of New York even provides. And yet, there's still a market process because they're competing with each other. So these sort of fusions, are kind of— have aspects of both state and aspects of both private company. And I think you're gonna see more of those. Another aspect of this that's really interesting is to change your taxi laws, you'd have to move from New York to Texas, for example, or California to Texas. To change your rideshare, you just hit a different button in the app. We take that for granted, but we basically unbundled this whole bunch of laws, your taxi laws, your hotel laws, all these other things that were previously locked in on a particular jurisdiction, Appified them so you can choose your regulator on a real-time basis.
To make that even a little bit richer, think about the fact that Uber and Lyft operate across all these different jurisdictions, and there's a lot of common infrastructure that they're supplying across all those jurisdictions because there's algorithmic stuff, there's loyalty, there's knowledge of how the system works, etc. All those things are sort of this not quite global but close to global public good that's being provided buy these things. At the same time, you've got the local regulator who is providing a public good in terms of the environment, and I'm not just talking about the taxi medallions that's competing with them, there's actually the regulations that are on top of Lyft and Uber, and these actually make a huge difference because— so for example, in San Francisco, most people are exclusive on one of the platforms. In New York, almost everyone's cross-platform. This has a huge effect on how the system operates in a way I don't quite have the time to get into, but it totally changes the nature of the competitive environment between Lyft and Uber. And there's all sorts of features that just don't work at all in New York that do work in San Francisco because of the nature of the regulatory environment that's imposed upon them here but not there. So you've actually got these public good providers in different spaces, one in cyberspace, one in physical space, intertwining with each other and creating the possibility of competition or not.
Right.
So what's actually going on is so poorly modeled by like this absurdly naive view of sociality as sort of being like, there's this one global layer and then we're all these atomized individuals. It's actually this incredibly rich intertwining of these things. And if we can actually come up with formal systems of value accounting that allow that to take place, that allow that like rich intertwining to like actually be accounted for and the like merits and points to sort of like accrue to the ways in which these things intertwine with each other, I think we can have just like vastly more powerful economic systems than we currently do.
I think blockchains in the fullness of time will at least give us some of the ability there, because just like 6 degrees of separation went from a party conceit in like the mid-'90s to a data structure where you could actually go and the social network became a data structure you could read, write, pull it, analyze it. The economy goes from this sort of vague thing with data, of course, that like Fred and so on put out there. to a data structure, namely the blockchain, where you can replay every transaction since t equals 0. And so, what you could do is classify all edges, you could group them in this sort of multi-scale analysis that you're talking about, where you have not just 2 agents, the state and the atomized individual, but multi-scale agents and whatnot. I think your simulation starts to get quite complicated there. The one thing I actually sort of want to quibble with you on, and maybe— so far we've had a love fest, maybe this will be a fight—
Yeah.
So, is the term public good, because I think it is true, certainly there's things that the current states provide that are public goods, like police services, electricity, not so much in California. But I think that a lot of them are— I'm pretty convinced by public choice theory in general that I don't think that many of these regulations are, today at least, optimized for the consumer. I think they're optimized for optics. And they're optimized for— for that person ascending to some higher regulatory position or getting headlines or something like that. I think very few of them are actually consumer protecting. So that's why I always hesitate when I hear public good, just because of that.
Yeah, no, I agree. What I actually mean by public goods— and public good is not quite the right word, I would prefer the word increasing returns— is any time when the whole is greater than the sum of the parts.
Ah.
And the thing is that that is like, thank God, the default thing that goes on.
Sure.
Because if it weren't, like, we couldn't do any better than we could do all living in isolated, you know, things out in the woods. And like, thank God that like we can get these benefits out of scale. Like, public goods almost never— of the type I'm talking about— almost never line up with the boundaries of historically defined like geographically described entities that are like supposed to be doing public good provision.
Mm-hmm.
Like you think about the drug war, it's like this is determined— our policies on the drug war are determined by like the, the white majority of the United States of America. And like the vast majority of the costs of that and the consequences fall on like Black and brown people in the United States and in Latin America. So it's sort of like an issue which should really be determined by Black and brown people throughout the Americas is instead determined by the white majority of the United States. What the hell does that have to do with anything? It's just an area where the nature of the economy of scale or the increasing returns process is completely misaligned with what happens to have been the way that we decided to arbitrarily chop up the world.
It's interesting because I think, like, I agree with the skin in the game aspect of it. Just as a sidelight, my recollection, I may be wrong on this, is with the crack epidemic in the early '90s, actually the Black community was pushing for more stringent laws at that time. And then they found that it overcorrected, and now we've got criminal justice reform. So sometimes, you know, things overcorrect and then you have to pull it back. And I think that's probably right. But I agree with the broader point on skin in the game on this. Like, that's to say, Somebody who is affected by something should have a say in that, and too much of the time, that's simply not the case.
I mean, the thing that I really find most interesting about the cocaine issue is actually not the over-enforcement in the African-American community, but the under-enforcement among wealthy whites. So like, if you actually sample the water supply of cities, you can determine where cocaine's being most used because it's very persistent in water. And most cocaine is used in the wealthy parts of cosmopolitan cities. And the incidence of ending up in jail if you are such a consumer is like 3 or 4 orders of magnitude lower than the incidence of ending up in jail if you're coming from the communities that you were describing. And the consequences that has both for the, you know, those poorer communities, but also maybe even more importantly for my heart, because my wife's a Latin American political scientist, for the people who get killed in Latin America fighting the war to stop the importation of the things that are actually being consumed by the people who are like funding the oppression of the people in the land.
Well, yeah.
It's just unbelievable.
I think there's basically 2 optima on this. There's the Netherlands model, where drugs are basically legalized. And there's a Singapore model where they don't do this catch-and-release kind of thing. They just execute drug smugglers completely. Now, you may not like that, but basically that has worked to keep drugs out of Singapore.
Yeah.
And so it's just an extremely harsh rule that is enforced very consistently. It's highly advertised. You get on a plane to Singapore, it says, warning, we will kill you if you bring in drugs. And so because of that, otherwise, Lee Kuan Yew was like, OK, this is going to become a center for smuggling unless we take an extremely harsh view on it. So I think the US does this also with healthcare. They kind of have the worst of both worlds system. Sometimes the US falls into this kind of bad region where you have the worst of the left and the right.
Glenn, you have a post on why you're not a technocrat. And why don't you unpack a little bit more your problems with technocracy? And how we should think about it.
What are the alternatives? First, to define what I mean by technocracy, it's sort of the view that there is some relatively narrow community, usually defined by some notion of formal expertise, though it's a little bit less clear if you think of the Chinese Communist Party, which I think is very technocratic, exactly what that formal expertise is. But whatever it is, it's some narrow group of people who are trained in some way who can sort of design a better system for everybody else to live in. I actually think the whole formal way that AI is set up as a problem, if you read the Russell and Norvig textbook or whatever, is the essence of technocracy. It's like, here's an objective function, we're going to design a system that just maximizes this objective function. There was someone who I won't name, but a very prominent person in Silicon Valley who said to me, look, I don't understand all these fancy words that, you know, you guys are using, social science words, but I know how to build automated systems. And let me tell you, you tell me what a good world looks like and I can build it. You know, and so that, that's sort of the technocratic attitude, right? And the thing that I think is so wrongheaded about the technocratic attitude is that any technical system, no matter how much formal input it elicits from participants, So it could have them vote, it could have them take a survey, whatever. Ultimately, the language in which the technocrats themselves design the system is going to be completely different than the formal input. It's not like when we're talking about designing some system, we're like, oh, Balaji, all I want to hear from you is a number from 1 to 100.
Sure, sure.
You know what I mean? No, we have a conversation about it, right? Unless you believe that all relevant information to the design of the system is either held by the technocratic class and expressible in their language or is expressible in the numbers that you give to the users to give their input, then you're basically leaving most of the relevant information out of the actual process.
Mm-hmm.
If you don't think that, then I think you have to do what I would call design in a democratic spirit, which is that you need to focus as much on the way you design the formal input that you give people the chance to give as you do on making sure the system is itself explicable, describable, reusable, critiquable, and redeployable by the participants in that system. who are then going to make use of that tool. And that's precisely what AI doesn't do and what other technologies do. It's like the vacuum cleaner, you know, kind of knows what it does, but that's not what an AI is supposed to do. It's supposed to do it for you. It's supposed to optimize for you, not be understandable and usable by you.
I have a few thoughts on that. One is I think it is almost inevitable that a technocracy in a broader sense arises. And the reason being because, for example, code is only a piece of it. If I install software, I know on some level a software engineer wrote this. But I think much more implicit is the idea is in everybody's head, some political engineer wrote that, in the sense of— actually, you open your book with this quote, which I also like.
Yeah.
Many men who believe themselves to be practical are often the slaves of some dead economist. The memes and the frameworks that are out there, a troll doesn't know Just like, for example, you don't know who your great-great-great-great-great-grandfather is, neither do the trolls. The memes in their heads have been installed from many generations. All these political ideas came down.
Yeah.
And so I think the folks who are really good at designing those memes are also a form of technocrat. They are programming human brains as opposed to machine brains. And I think that combination is something that's there. The other thing is that while I, It's situational as to when this is the case. You've heard that famous saying, you know, if I asked my customers what they'd want, sometimes, you know, they would say a faster horse. So simply doing bottom-up all the time will not necessarily always get you that leap where sometimes only an individual can do that.
So I completely agree with that. I don't actually believe in bottom-up, like, I mean, I wouldn't be doing what I'm doing if I was. No, so like, I obviously believe in design. I think design is incredibly important. I think design with a democratic spirit, however, is extremely important. And like, I would contrast the French and American revolutions. This is a contrast from Hannah Arendt, who's one of my favorite philosophers. So she says that, you know, the French Revolution was sort of like, we want the following thing, like food or bread or whatever, like, give that to us now. So we're just— and then, like, however you're going to do that, oh, that's fine, just leave it to you.
Mm-hmm.
Whereas the American Revolution was like, we want legitimacy, we want a legitimate system. So we're going to go off to the US, we're going to, over a long period of time, have people experiment with these local democracies. This is going to build up a system that comes to pervade how people think about things and how they relate to each other and how they would choose to organize themselves. And then that system is going to get better and more powerful and whatever, to the point where it makes the system from which it's spraying look stupid. And then it's just going to replace that system. You know what I mean?
Yeah, absolutely.
And so that is like the— but of course, those ideas were designed. I mean, like, those came out of the Enlightenment, early Enlightenment. They came out of, you know, Locke and whatever. So it's not like they didn't— they weren't ideas that were hatched somewhere, but they went through this long process of sort of socialization and coming to pervade the way in which people interact with each other before they then were like the system.
So it's funny, the way I sort of think about the French Revolution and the American Revolution is, I think— and this is one way of cutting it, and it's not the only way by any means— but insofar as the French Revolution was like a pure left revolution, but the American Revolution was a balance of left and right. And so it was both all men are created equal and no taxation without representation. And I think you had both of those themes there, where you had sort of the moral spirit, the zeal, that the left can provide, and sort of the, for lack of a better term, the practical substance that the right provides. And I think that if you overbalance on one of those two, you lead to something which is either extremely narrow-minded and self-interested, or extremely impractical and optics only. Another example of that is the startup itself, which also combines both left and right. It's got the insistence on disruption and the lack of hierarchy, Informal dress, you know, it's very international, F the man attitude, but it's also capitalism, it's technology, it's, you know, a market process and so on. It's that fusion, I think, where some of the best stuff happens as opposed to kind of overcorrecting on one or the other.
I have one more question, then I want to open it up to the audience. You mentioned your sort of the Alone framework and that sort of leading the whole sort of theory of economics to be sort of a house of cards. Where are you right now in sort of, you know, building a replacement model for it, or how are you thinking about alternatives?
Well, so I think that what we need to do is we need to replace the— so basically, like, the standard political economy we think of is built on basically 2 tokens: an identity token for individual identity, and an anonymous value token called money. And I think we need to replace that. So that's basically like an n-dimensional vector. It's like for every person, there's how much money they have. That's like a simplified representation of it. I think we need to move to something closer to an n by n-dimensional vector. N by m. N by n. So like instead, we have a representation of the relationship that every person has to every other person. In the network. Because—
Like a David Graeber-ish debt kind of model in the sense that—
Yeah, I mean, you can think of it as being partly inspired by Graeber, but the key function—
I don't like all this stuff.
Yeah, yeah, yeah, there's a lot of issues there. But anyway, but like the idea is that then you can actually represent the social relations that people have to each other, because I think the fundamental problem of economics is not like how do you reconcile the individual self-interest with the common good, The fundamental problem should be, how do we have cooperation under conditions of diversity? Because the truth is there is no such thing as selfishness. It's not even meaningful because given that so much, almost all of what we consume, it is consumed in groups in some form. What there really is, is either I'm super aligned with you in terms of the groups that we're part of and there's no conflict of interest between us at all, or we're like really socially distant from each other and then there's a ton of conflict between us. Because we're different. And mechanisms like markets that treat symmetrically the interactions between 2 people who are very close to each other and 2 people who are very far from each other are really problematic and are going to lead to all sorts of issues because they don't actually overcome the fundamental problem, which is difference and the fact that we actually are very delighted to live in societies with lots of difference, but that it's hard because that's what creates conflict.
I like that, and I frame that, or I think about that in a complementary way maybe. Like, you have your whitelist, which are those folks who are friends, family, the folks in your contact book, and so on. There's your blacklist. Let's call that, first order, the folks you block on Twitter. And then actually, the vast, vast majority of human beings in the world are on what I call your green list, where the correct way to interact with them is not with, you know, like a handshake or the sword, But a dollar bill, right? In the sense of, you know, they're neither friend nor foe. They are somebody where there's a mutually, you know, like a valuable transaction. And so that's like kind of this huge area where current social networks are really more about the whitelist. Twitter actually does expand into strangers, but I think there's like a huge amount of that green list economic activity that hasn't yet been put online.
Yeah. So I think one of, if not the fundamental problem with the internet as originally conceived, is it was a bunch of whitelists interacting with other whitelists. And therefore, it sort of thought, oh, OK, and that's just going to scale.
Right.
But of course, it didn't. It was small, and it was a property of things that were Dunbar, 150 people small. And then you get out there, and it just totally breaks down. Then, of course, you can have platforms come in and do their thing and whatever. There's been a fair bit of that. But what we actually need to get back to is how do we actually build up white to light gray to slightly darker gray to dark gray to black lists? How do we actually have a network-of-network conception that allows us to do that? That's the sort of thing that I'm working on to try to replace the standard There's just equally spaced atoms out there.
Paint a bit more of a picture of what that could look like. What are some concrete implementations or manifestations?
The idea— I can't really get into all of it here— but is basically a network. If you have a network with weighted edges, effectively that's like a representation of those social relations. And then if you can take that as an input to mechanisms, You can have mechanisms that are very different than like a traditional market would be. So like the typification of a traditional market, this is what I talk about in my book, is an auction. Auction is supposed to put everyone in the symmetric position. But imagine you have an auction for a piece of art and 2 people are at the same office down the hall from each other. If they win the piece of art, they're going to put it into their office, which they walk— each walk by every day. Then there's an art museum that they both frequent, in the city they live in, also competing. And then there's some indigenous representative from the place where the art was allegedly stolen from, trying to get it back for their country. To treat these 4 people as, oh, they're just all participating in an auction, they have some value, this is the most ridiculous thing you can possibly imagine. Because of course, the guy down the hall winning it is 80% as good as me winning it, from my perspective. the art museum is probably 40% as good, and the guy from really far away is worth zero. If the auction doesn't treat the social relations between these people in a way that is meaningful, it's missed most of what's actually going on in the design problem. I think the problem is that money and individual identity as markers just don't have the informational content that's necessary to do design for most of the problems we care about.
Let me give you another example of this, a different example. The cap table, something I think about a lot.
Yeah.
So, you know, the cap table, the table of all shareholders in a company, arguably, you know, the most important data structure in Silicon Valley. Hahaha. And the reason is it's what built Google and Facebook and so on. It aligned all these people behind it, you know, as opposed to like a skip list or whatever other data structure you might have in CS. So here's the thing that I thought mapped to something you were saying. Whenever you're doing anything complicated in a company, like if I'm helping a founder through some complicated financing or something like that. They've got— he's got rows in the cap table, and certainly he's got how many, you know, seed, you know, common A, B, C shares these people have. And he's also got the corporate charter, which says, okay, what threshold do you need? But the thing that's not in the cap table, but it's critical, is his relationship or her relationship with every one of those rows, because that is in their head. And then you need to actually encode that as like, okay, I'm on good terms. Okay, I haven't talked to this person a lot. Okay, let me warm them up.
This is my brother.
Yeah, exactly. That's right. That's right. So, you've got like all of these kinds of relationships, and then what you do is, you know, frequently you'll put like a +2, a +1, a 0, a -1, a -2 for how warm or cold they are on something. Then you map that out, and you've got, okay, I've got 72% of Series C, but I need, you know, more of Series A, and then you work on getting the -1s to pluses. And so that's something where there's actually social information on that edge along with the economic information. The social information is often less written down.
Yeah, well, and the thing is that even there, that's still sort of n by 5 or something. But the thing is, actually, to get a voting system that would actually make sense at some level, it needs to be n by n. Because think about quadratic voting. It's great, right? You sum over the people, this intensity of preference thing, whatever. But that's really not quite right, because the thing is, like, what you want to know is not really like, okay, so what's the sum of the intensity of preferences? But like, these people are really similar. They all like this. Okay, well, that gets some points. But like, this other person who's really different likes it too? That should get like way more points, because like that's independent information, right?
Yeah.
and this other stuff is all just almost perfectly correlated with each other. So that information about how these different data points are actually related to each other is so important and it's totally missed.
You might even call it, if it's valuation in this form, you might call it eigenvaluation.
Exactly. No, but it is actually. So the thing is, the way I think about it is that actually, quadratic voting, it takes the sum of the square roots of the money that people contributed. But like, the elements in the sum are individual people. But the truth is really there's like an eigenspace out there of like influences on those people, which is actually what you want to be taking the sum of the square roots in, not in the like direct space.
Can I ask you some dumb questions on quadratic voting that you may have already answered? A few thoughts, but one was I didn't see in there, and again, maybe in one of the papers, the secret ballot concept. Like, that's to say, people putting their votes out, typically, if those votes are public and they're enumerated, you know, like, how did you think about that?
So there's a nice cryptographic paper about that by—
Commit reveal?
Actually, like, from RSA, Ron Rivest. He has a nice technology for that.
Oh, got it. Okay, so because if there's a disutility to revealing your vote, then that could be an extra term.
Yeah, and you can actually stop people from revealing it. There's both physical cryptographic techniques and equivalents that are computational that you can use for that.
All right, cool.
I want to open it up to the audience, so just say your name and ask the question.
Julia. So Julia asked, what would we change about property law?
What I would change is I would allow for more different kinds of system administrators to experiment with different kinds of property law. For example, If you had a Twitter 2.0 where usernames could be bought and sold in a free market rather than just being first come, first serve, I'd like to see that experiment. If you have something where the earlier you sign up for Facebook, the more crypto you get, so the early adopters who shared it virally get more and they have maybe more votes over whether someone is deplatformed or not, I'd like to see that experiment. Basically, I'd like to see a lot of different experiments in property law because I think there's lots of things that work in practice, but not in theory, and vice versa. And I consider myself an empiricist on that.
I think I agree with a lot of what Balaji said. I think as a matter of principle, broad level, I think we have to move towards— and by the way, I think actual property law as it actually exists rather than how we imagine it is actually more similar to this than the way we imagine property law to be is. But I think we need lots more Partial overlapping, non-exclusive, non-absolute property rights where there are— where everything is sort of like belonging partially collectively to many different layers of ownership and whatever. And that's not quite what's actually described in the book. It's in the book, it's still like this very like the individual has a share and then the collect— the collective. has the rest. But I think the reality is what we actually need is multi-layered, rich, overlapping, partial control and whatever. And that needs to evolve in a way that actually makes sense to people and that they can process. So how that's actually going to work out, as Balaji said, is going to be empirical. But I think the basic principle of trying to move to this richer, overlapping sense of ownership is Glenn, the most powerful idea in your book for me was quadratic voting because it was just so practical.
I went back to my investment firm and said, hey guys, this is the way we should be thinking.
This will lead to smarter decisions for us.
In that vein, I was just curious, if you took all the insights you've had and said, hey, for us who aren't spending all time investigating these, we were going to look at what we could actually action or what's actually applicable or that we can take to our organizations or our own lives, what would you suggest?
So the question is, what's easiest to apply and biggest near-term impact of the ideas that we've been playing with? I would say probably of the things that are really ready to go now, quadratic voting, number one, quadratic finance, number two. Quadratic finance is an idea I developed with Zoe Hitzig and Vitalik Buterin. that I now actually think is just an application of quadratic voting. I think I figured out how to make it just like a special case. But in any case, it seems different, and it's a way of sort of funding organizations based on matching funds from a central part. So I think that's very applicable within companies. If you have different divisions trying to cooperate, you can use this for that. Number 3, I actually think that the, like, cost, Harberger tax, salsa idea is quite applicable just for managing assets, like managing Teslas. Rather than selling them, you could lease them out using that sort of system. I think that's quite useful. Then there's stuff that's useful at a more political big level, but I'm focusing on more the corporate side because it sounded like that's where your question was coming from.
How do you overcome collusion in fundraising?
Suna asked a Fabulous question, a critical question, which is, how do you overcome collusion in quadratic voting? The obvious point about that is that if you have the sum of the square roots, 2 people could do better rather than by one voting -10 and the other voting +10, both voting 0, because then neither of them has to pay for it. I think that My basic answer to that is, so you don't overcome collusion in quadratic voting. That's not really possible because most of social relations are not actually collusion, they're just social relations. So if we're husband and wife, or wife and wife, or whatever, or 10 people who all live together, of course we're going to be caring about each other when we make a decision, not even about collusion or even non-selfishness. It's just like, if one of us does well, the other is going to benefit from that. Most consumption is not private, so it's just not even meaningful to think about people as these completely separated entities. The whole problem of collusion results from thinking about people as these separate, equally spaced atoms who you just want to have make independent IID decisions somehow. But that's not how any human being lives. So what we actually need eventually to do is to move beyond that whole substrate to a system that actually accounts for that stuff and build mechanisms that are based in that space. That's going to look different than quadratic voting is going to look. That said, is quadratic voting better than one person, one vote? I think in almost every circumstance, the answer is yes, because one person, one vote is even more absurdly based on this notion of, like, there are individual human beings and they each have one unit and they just put their token in this box or they put that token in that box. So that doesn't mean there's no cases where quadratic voting is going to work worse than— but I think almost everywhere, one person, one vote is even more extremely crazily in that direction, and quadratic voting starts to overcome it but definitely doesn't overcome it fully.
My immediate thought on that, by the way, is it might be interesting to look at You've got different distributions of who has what votes. If you had a voting bloc that did collude and voted as one, what size does that bloc have to be to always win and have interests that are disaligned with the rest of society? I think that would be an interesting angle of analysis.
You'd probably give me the answer. Yes, we have a paper about that. It turns out the bloc has to be pretty big to really— but the thing I realized is That was my almost defensive answer, but ultimately I don't think that's that satisfying because it's sort of like, well, it's true that if that's one group that's trying to collude and dominate, they have to be reasonably big and whatever. But ultimately the problem is not that. The problem is that the whole system is missing the fact that there's going to be little pockets like that everywhere and that you're not really going to get the right result. out of a system? Because you're probably almost certainly going to get something better than something ridiculous like one person, one vote. But can you actually get the right thing by taking more of this into account? I think you can do a lot better for sure.
Is our non-individualness, as you just described it, a function of the current limitations of technology? And in the next century, is that an ideal that we will reach closer to? Over time?
Well, I think again, like, if you think that like individualness is this like state of nature in which we were born, which was awesome, and we're trying to get back to that, you might think that it would go in that direction.
Well, maybe not back to that, but it just tends to be where we're going?
Well, I think like what will happen is we'll get increasingly socially complex, and that will make each of us more and more an individual than we've ever been before, because the more types of sociality that exist, the more possibility it is for us to actually be unique. So I do think we'll get more and more individual, but I think the way in which we'll get more and more individual is by allowing diverse social organization to flourish, and therefore us to be individuated by all the different ways in which we participate in sociality. So the goal of getting more individual is realized precisely by allowing the process of creating social groups to be more fluid and democratic and rich and so forth.
And just very practically, 100 years from now, are we having hundreds of Singapores and Israels and Hong Kongs and Dubais, or how does it manifest?
Well, first of all, I don't think that physical space is necessarily going to be the main thing in the future. I think we'll have tons of different types of transportation technology. There's a wonderful paper by Joe Lamke, someone who's involved in the radical exchange movement, who's an urban planner, who describes that as you get new transportation technologies, you get new clusters emerging that are based around the optimal distance to go with that transportation technology. So you don't walk more than, I don't know, 10 blocks, and so you have neighborhoods. And you don't take a scooter more than a certain amount, so then you have super neighborhoods. And you don't take an airplane more than a certain amount, so you have country. This is the sort of thing. And so as you get more technologies, there will be more different levels of those socialities. And of course, There's non-physical versions of that which have to do with social distance and whatever. These create these multi-scale fitness landscapes in which all these different types of sociality flourish. I think that it will be the intersection of all of those things that will be the relevant thing as more of those things proliferate.
I agree with a lot of that. I would also say social media and cryptocurrency are going to cause unbundling and rebundling of nation-states. In the sense of basically folks who live within the same physical area no longer agree with each other. They agree with their other ideological members from different countries online. And eventually that's going to result in a huge amount of migration and churn as what we currently have as borders change. The closest analogy might in some ways be partition, hopefully without the violence. But that was something where ideologies sorted geographically. And I think we may see a lot of that.
You both talked about Uber as an example of unbundling governance from land.
There's other examples, perhaps Amazon and commerce.
Is there other places that you think that the laws or the ways in which we're regulated are currently done by land-based governments that should be done by perhaps some other type of governance or protocol?
Yeah, I think actually if you look at every large-scale profitable— or not every, a lot of them— A lot of profitable entities, cloud entities, turn out to be cloud regulators. For example, Amazon has star reviews and reviews of the reviewers. PayPal has anti-fraud. eBay has star ratings. Gmail has spam filtering. Apple has reviews in the App Store. And you actually start going down the list, and it's hard to find a scaled e-commerce marketplace or really any service where humans interact that doesn't have some form of reputation and review. And so, in a sense, these tech giants have already become good at managing humans and all their foibles and kind of these automated systems and setting up these rules. I think that one of the big frontiers for this, maybe an obvious one, is crypto with smart contracts and whatnot. And I think where that's going to colonize are niches where those contracts were too expensive to do, too international to do, too simple to do. Like lemonade stands or complex cross-border things or things like that. And then over time, they're going to eventually be sort of back-recognized into law in the same way that eventually the RIAA made its peace with file sharing because it couldn't stop it. So those are the kinds of things I think about.
So maybe this will make me sound even more extremely libertarian than Balaji in one example, but I would say in the medium to long term, I don't think there's almost anything that it makes sense for nation-states to do. Seriously, what is a nation-state? What the hell does that have to do with anything? Maybe nation-states would have the role of preserving the historical landmarks that are associated with that nation-state as a historical entity, because nobody other than the people in the— but even that's probably not true, because lots of tourists like to come and see American— so I'm not even sure that's right. I think there's almost nothing where the nation-state is the optimal public goods governing entity. In the near term, what I would say are some of my favorite examples are data governance. The patterns of data being used to make machine learning algorithms or to learn— this has nothing to do with geographic boundaries. That's crazy. Another example that's related is Nadia Egbal, who's actually in the audience today. Her work about open source has really made me think about the governance of all the protocols that we use. That's another area where nation-states make absolutely no sense and there's no funding. It just doesn't line up. And then my third example would be environmental stuff. You think about how much violence nation-states have done to environmental features. You draw the line on the map somewhere and it's like the river doesn't know that. The river just goes through. And then it's like the Amazon River. There's a tiny minority of the population of Peru, Brazil, etc., etc., that live right along the Amazon River. So there's this small group of people within each of 5 different nation-states that's getting completely screwed over by the tyranny of the majority and by the fragmentation of the nation-state. And it just doesn't line up in any way. So you need a nation of the Amazon, you need a nation of SSL, you need a nation of— that's what we need to start having.
Yeah, so here again, I think the discussion of technocracy is super interesting. So I'm wondering to what extent do we each think of it as a fault, uh, technocracy as a question of aligning incentives of the technocrats versus gathering information? And could you improve the incentives by having like retroactive votes after the technocrats have implemented the policy?
And Samo, why don't you ask yours too, and then we'll have them answer.
For nation-state skepticism, The notion of the nation-state is quite welcome.
It's been overemphasized for a while.
But I will point out the nation-state was already over by 1945, right?
This is the United States and China and more empires.
So what do you think of empires?
Okay, question number 1, is technocracy primarily about disalignment of incentive of the technocrats? And number 2 is, what do we think of empires? Do you want to start?
Sure, sure. Yeah. So, with respect to technocrats, is the issue their lack of economic alignment, or is that what we need to optimize versus collecting feedback better, I think was the question. I think it's all of the above, but what I always keep coming back to is, so long as you have a choice over who is to govern you— actually, really, let me be more precise. I think there's 2 different things you need to optimize in a society. The first is sort of Let's call it the path of the citizen. And the citizen, I think, vote with their ballot, vote with their wallet, vote with their feet. I think that's important that we have all 3 options there, that they can choose among their leaders. And the second path is the path of the leader. And often in conservative societies, they don't have a path for an ambitious young person to become a leader. And so what that leads to is a bubbling up, then it becomes revolutionary energy, and things are thrown off. What's great about the whole startup world is it takes these ambitious young megalomaniacs and it puts them into something where their actions are actually mostly pro-social and lets them create things that are positive things. Without naming names, for example, there is a prominent tech founder who was a huge email spammer in college and has now become a tech billionaire. And basically he built something that many of you have used. It's almost like reforming a criminal. And it's actually— the difference is a criminal breaks rules in order to destroy. An entrepreneur breaks rules in order to build. So that path for leaders is also important. And I think if you've got both of those two, where you have a pro-social path for leaders and you have a path for, quote, citizens who don't want to take that level of risk, I think you can have something good. But what I come back to is the choice for both. You need to have choices for both.
Technocrats and empires?
Yeah, and so an empire. So if I was to restate the question, is basically the nation-state itself is sort of obsolete in the sense of since 1945 or thereabouts, we've had the US Empire, the Soviet Empire, the Chinese quasi-empire in terms of its border area, and so on and so forth. And I guess what do I think about that? I think countries like Singapore, Israel, Estonia, and then China is a big exception that proves the rule, are the kinds of things that I expect to do well in this century, which, you know, those countries are not empires. They are nimble. And I think just huge quantities of people are going to be a disadvantage because scale causes disalignment. That is to say, it's much easier to get— obviously, to get one person on the same page is trivial. To get 2 people on the same page is doable. 10 is harder. 100 is harder still. You can almost envision yourself being at the front of a long procession, And if you're, you know, running a company, for example, if you radically change direction, it's like a snake which kind of goes like this. It takes time for those people to catch up to where you were. If you keep changing direction, there's folks who are still executing the plan from like 3 turns ago. Now you take that at the scale of an empire, and it's just very hard to lead large numbers of human beings, especially any kind of shift whatsoever. Whereas you can be much more nimble if you have kind of smaller groups. So we're already seeing this at the corporate level. You know, Hewlett, you know, HP broke up. It would have been hilarious if they called themselves Hewlett and Packard. They didn't exactly do that. eBay broke up into eBay and PayPal. You're getting these kind of frictional forces that are— or rather, centripetal forces are breaking these companies apart. And I think scale is actually becoming a disadvantage in this century, where it was an advantage last century. And so I think empires are— I mean, the one exception is China, because they actually might be competent enough to keep the whole thing together. But I think for the most part, scale causes disalignment.
On the question of technocracy, I don't think these days that I really see as much of a distinction as I used to between incentives and information. Because if you have this alone view of equally spaced atoms and then the whole, then there's a very clear distinction. There's the individual incentives and then there's the information that should influence the whole decision. But if you have this other view where it's like Well, everybody is not selfish, but they're partial to their perspective and causes. Then the distinction between information and incentives really starts to blur because, well, are those economists just being selfish and serving the economists, or do they just genuinely believe in that particular worldview, which is super narrow and based on this formalism that they have that thinks that they're the most important people in the world and that they're the only ones who know anything, that every sociologist and anthropologist and person on the street is a complete moron, and that's the only reason they don't understand that capitalism is the greatest thing ever. I'm not even sure that there's that clear of a difference between those perspectives. And so I don't think voting can really solve the problem because the issue is really about fact that voting itself is a super thin formalism which only works under very limited assumptions. We've been talking about all the problems with voting. Actually, you need to have the conversation and the openness to that more unstructured type of feedback that only making things clear communicatively can do. In terms of empires, I think that my fundamental issue with empires is that what we want as much as possible is like a network of networks that builds from the bottom up, that has that rich information, rather than a single system that comes in and imposes itself. Because inevitably, that leads to everyone is this uniform, equally spaced atom from the perspective of the empire. Well-constructed empires try to avoid that as much as possible. To the extent that they can be built as networks of networks, obviously, they'll be better, but the natural imaginary of empire is conquest and imposition of a system, which then wipes away a lot of the information that could have been built up from more bottom-up type processes.
Speaking of empires, Balaji, do you want the last word?
Sure. There's a website I've heard of that you might want to go check out if you want to just go and sign up there and maybe hear more later. It's called nakamoto.org. So just go put in your email there and you might hear some interesting things in the future.
Awesome.
Please give a huge round of applause for Balaji and Glenn.
And again, if anyone's interested, check out Radical Exchange. We'd love to have all sorts of people involved. It's a social movement.
Buy his book also. It's awesome if you haven't yet.
Awesome.
Thank you.
Thanks.
Fantastic. Upstream with Erik Torenberg is a show from Turpentine, the podcast network behind Moment of Zen and Cognitive Revolution. If you like the episode, please leave a review in the Apple Store.
More from Erik
Palantir CEO Alex Karp & Rep. Ritchie Torres on Innovation & The West | Hill and Valley Forum 2024
