#28 - Timur Kuran

0:00Today, I am happy to welcome Timur Kuran to the Network State Podcast. And Professor Koran is the author of Private Truths, Public Lies. And I've cited him many times over the years for people who want to understand the 2016 election, people who want to understand distributed consensus, anonymous consensus, people who want to understand the fall of the Soviet Union. So I've been a fan for a long time. And, you know, so that's how I came across Professor Koran's account. And then I was tickled to see he followed me on X. And I wonder how you came across my account. I don't think I've asked you this.

0:35I came across your account because people I follow forwarded your very interesting tweets. And I can't remember exactly what they were, but this was when I, it must have been around 2017 or 18, soon after I joined Twitter. And that's how I discovered you. Excellent. Well, so, you know, I want you to give a quick, uh, we can talk about the kinds of things, you know, topics you're interested in. I'm always interested in like how people are interested in each other's intellectual currents. Cause I, I do talk a lot about tech, but I talk a lot about politics and, you know, the future and history and pieces like that.

1:21So there may be strands that appeal to you or what have you. I wanted to go through your work though, because, um, I think private truths, public lies is something which has just started to break down in the U.S., not just started, like last six months or so, we've seen an enormous preference cascade. And maybe you can describe, summarize the book and talk about what motivated it and then how you apply it in other places. Certainly. Well, let me start with the definition of preference falsification. It's the misrepresentation of one's wants under perceived social pressures. it involves of course self-censorship hiding one's actual preference from the from the public but also and this is very important this second component is extremely important it involves pretending to like someone something that one actually dislikes and And one engages in preference falsification to avoid social sanctions, to avoid ostracism, to avoid being called names, to avoid being associated with disreputable groups, to win favors.

2:47Now, there are many negative consequences that preference falsification has. One is that it distorts collective decisions. Let's imagine a department, an academic department, that is trying to decide whether to continue using standardized tests in admitting PhD students. Now, thousands of departments across the United States that faced exactly this question, this issue, because there were people who argued that standardized tests discriminate against marginalized communities, so they brought the matter to department meetings. Now, imagine a department where the majority, even a very clear majority, a large majority, believes that there's a great deal of valuable information in standardized tests.

3:51Even if you don't use standardized tests exclusively, do you look at other things as well? There's a great deal of valuable information in them. But these same people are afraid of being labeled as racist, as misogynist, as sensitive, etc. And the people who are pushing hard for eliminating standardized tests are quite powerful. They may falsify their preferences and just to avoid the wrath of the promoters of eliminating standardized tests. And as a result, the department might eliminate standardized tests, even though a clear majority believes this is a bad decision. Now, preference falsification is accompanied by what I call knowledge falsification, which is that you, in the process of trying to project to the public a preference that you don't actually have, you conceal what you know and you pretend that claims made by others that you believe to be false are actually true so in the process you distort knowledge if you are pretending that if you believe that standardized tests are valuable, but you are pretending that they're not, you cannot publicly talk about all the virtues of standardized testing.

5:47You have to say things to be convincing to the people who want to eliminate standardized tests or want to retain, want to prevent their reinstatement, you have to pretend that you accept all the arguments made by the opponents of standardized testing. So one thing that I remember from your book that I thought was interesting was, in my head, I was comparing it to a method actor. because sometimes these people who are forced to mouth all of these falsehoods for a long enough time find themselves quasi-believing the falsehood or rationalizing it to themselves. And the external lie sort of shapes the internal narrative, at least partially.

6:40It does. I mean, there are some people who actually come to believe their lies, uh perfect falsification in in a wide array of contexts it's becoming a visual it's like the vaklov hovel thing about uh workers of the world unite right yes and they start they they they get used to the notion that whatever slogans they're supposed to parrot they they will just parrot them They know it's futile to question them, to inquire into their veracity, whether they mean anything. So they simply accept the worldview that is presented to them and they start believing them. That's a very shallow belief, though.

7:29It can be because it doesn't have much depth, because it doesn't involve actual evidence, because it doesn't involve a study of the mechanisms that might be driving certain outcomes. It's quite shallow. So it's actually, it can be changed quite rapidly. But you're absolutely right that some people will come to believe the slogans that they are parroting. At the same time, there are people in society, this was certainly the case in Eastern Europe before the fall of the Berlin Wall, there were people who did not believe the lies that they were being told. They did not believe that central planning was provided a better economic system than a market-based economy.

8:34But they had to conceal their knowledge, or many of them concealed their knowledge, the reasons they believed that a market system would be more efficient. And they had to, in order to advance at work, in order to retain acceptance in their communities, they had to pretend to accept the party's ideology. So that process distorts public knowledge. And it affects what affects the ability of others to see the world correctly. correctly a couple of things on that i think a lot of people not certainly not you but many people don't understand just how crazy crazily taboo capitalism was in communist countries like in maoist china um you know people were executed for uh you know having some grain held back rather than throwing it all into the common pool um the zhao gang thing which may be apocryphal about the villagers who asked, do I own the teeth in your head?

9:48And they said, no, the teeth in your head belong to the collective. And they signed a contract amongst themselves and hid it in the roof of a hut because the split of property and private property rights were so taboo that these were capitalist running dogs and reactionary. So like capitalism was punishable by death, right? And part of what justified that is that so much blood had been spilt against the evil capitalists. What do you want to let them back in? These selfish, crazy capitalists, are you a traitor? So it got mixed up with nationalism and you had this crazy level of taboo against something that we don't think of as being taboo.

10:26And then another piece of it is in the, in like the Soviet Union, especially, and, you know, I'm less familiar with exactly how it's played in China, but there are only two kinds of people who are capitalists, which were A, these sort of intellectuals who could see beyond their own society and whisper about it in psalms.literature and so on and so forth and then be the black marketeer who'd cut you for a tomato and who lived down to all of the regime's stereotypes about capitalism right and because they were selfish they were greedy they were fraudulent they they were they just cared about money they didn't care about the people and so forth so it's a little bit like the you know if guns are outlawed only outlaws have guns kind of thing, where if truths are outlawed, only outlaws have truths.

11:15And that's a secondary effect of the, you know, public lies, private truths kind of thing, right? Because it pushes the truth to the margins and the people who voice them tend to be either intellectual or disreputable. I'd love to know your thoughts on that. No, no, this is, this is, this is absolutely true, but also makes it, let's pursue your, your example, under central planning, even the heads of enterprises that were given quotas to fulfill production quotas, to fulfill their quotas, they had to use black markets. So they all knew that these markets worked. They all had personal knowledge that these free markets worked.

12:13worked. They pretended, many communist leaders pretended that they were implementing central planning when they knew very well that the only way to approximate their goals involved using... Shadow prices, quasi-capitalism. Exactly. The shadow process is so on. And that all this was going on, but they maintained a facade of commitment to central planning. Students were taught in schools and universities. They were taught that central planning works and it's so much more ethical. It's so much more efficient than capitalism. Yes. It's similar to in the U.S. system using your university diploma as a socially acceptable proxy for IQ or SAT scores or something like that, like a proxy score, like a shadow price.

13:15Because the college board and SAT tests, colleges can administer them, but private employers couldn't for a long time. Um, but one, one thought I have is, um, so I say, you know, you know, you're mathematical enough person to, you know, you know, the difference between a vector and a scalar, I assume. Yes. Okay. So one thought that I've had, which kind of builds off of your point just now about these individual administrators, they knew that the system was broken. One thing I observed, especially during the Biden administration, was, you know, obviously people were lying about Biden being senile, but they were also faking employment statistics and they were faking inflation stats and they're faking these different things in ways that they themselves.

14:09So like, for example, Larry Summers put out a post said something like if you included shelter and like the rise in costs of loans and other kinds of things, actual inflation rates like in double digits or something along those lines. Right. So he was he was somebody who was expert enough to understand how this particular agency was faking their particular thing. And then David Sachs and I looked at some like employment data and we're like, why is it beating expectations so many months in a row? this is like kind of weird and it keeps getting revised downward afterwards. And now there's like some revelation that, oh, actually it was fake after all.

14:44And of course, but so the point is though, that each of these kind of administrators was putting a thumb on the scales or faking things or whatever you want to call it in a way that it would be hard to figure out how they were doing it unless you had some domain expertise there. But then they were also seeing everybody else's fake numbers as if they were real. Right. So you have like 50 pieces of fake information out there and for 49 of them, you have no internal information on how it's being faked or the 50th one that administrator is faking it. So they think maybe the system is OK. And then like it's only their piece, which is a little questionable or something like that.

15:22So they can't build an overall system map of how it's broken. Let me know your thoughts on that. Actually, there were many people who did sense that the entire system was increasingly built on lies. And that is one of the reasons why Trump, who is a very flawed man in many respects, has received so much support in three elections in a row, winning two of them. Because many people outside of the Washington bubble, outside of the centers of major influence, of major power in the United States, they have sensed that it's not just one number that is fake, that there are many, many numbers that are fake.

16:16And they can imagine that if the press has been systematically hiding the explanation for COVID, if it refused to, for a long time, allow genuine debate on COVID, it must be happening in a lot of other areas. so i want to i want to show you something that maybe get your reaction to this so you know this is a post i wrote it's about two years ago too fake to tell you know like too big to fail right so because you know sax and i you know saxon made this off-end tweet hot jobs for today i don't know where they're finding all these jobs all i see are layoffs and it got four and a half million views so clearly like this just struck a nerve because nobody they wouldn't have gotten so mad if it hadn't actually been like fake in some way right and um so you know and trusted media as you you know as we were just talking about half of americans now believe news organizations deliberately mislead them but it's one thing to kind of have an overall suspicion and then it's like another thing to have a specific suspicion and one of the things that i was just doing is i was thinking about a bunch of issues where it's clear the information was false and there's a there's a whole spectrum between like you know maliciously faked and you know true and some of them include like bernanke talking about the great moderation how it means to attain macroeconomic volatility before the financial crisis that was just like you know incorrect right because obviously you know volatility was coming back with a vengeance cheney saying deficits don't matter Bush and Clinton blew up the mortgage bubble.

18:03Bernanke saying we might see a mild recession. The ratings agency in 2008 called subprime mortgages AAA before the collapsed global economy. S &P got retaliated against for downgrading things. So this is a really interesting one where the guys who put out true information where they're downgrading U.S. debt, they got retaliated against. Then Bernanke said QE isn't money printing. Powell saying it was. Greenspan saying the U.S. can never default because it could just print money. erroneous projections of unemployment rates. This is like this scrap that I don't know if you remember from almost 10 years ago that showed.

18:37By the way, I didn't know that unemployment rates and growth rates were systematically being tampered with. But that information reminded me of a thesis that was written under my supervision a number of years ago by a Chinese student who studied Chinese growth numbers and official growth numbers that are issued quarterly. And what she found was a systematic pattern whereby the numbers were inflated when they were announced. And before the next quarterly announcement, the previous quarter's numbers were systematically reduced. Yeah, exactly. making the new number new exaggerated number look even better yes exactly so that this was so the graph systematic process yeah so this is u.s had a few job fewer jobs past years initially reported it's 2024 admission and now 2025 just a few months ago now that the new admin is there and of course the you know other people would say that these guys are you know but the biden administration claimed to have added 400k jobs from july through september in order to win re-election new data suggests none of those jobs ever existed they overestimate the number of jobs and so a point is that there's basically a very severe dispute over of course trump in his own way has a reality tv of um many things in particular like you know a lot of the stuff around trade and tariffs and so on and so forth is is incorrect but anyway my point here's my my thinking on this, and I'd love your thoughts because this gets, I think, a deep overlap of where we think about things.

20:38So if you go through all these things, all these different economic stats, that's 20 years of economic fakery. But it's not all just lies. Some of it is error. Some of it is very material revision. Some of it is spin, like quantitative easing isn't printing. Debt ceiling doesn't matter. It was tribalism. Some of it's like pseudoscience. The AA mortgage back security is arguably outright flawed. And then transitory inflation in 2021 was all of the above. It was error. It was revision. It was spin. It was tribalism. It was pseudoscience. And eventually it was outright fraud because everybody who bought treasuries in 2021 bought them on the false pretense that rates were going to be low forever.

21:19And they weren't. They were hiked radically in the next few months. So the reason I say this is when you have the centralized authority that you have some local information and you can compare it to what the centralized authority is putting out. And you can see that based on your local information, it doesn't match up with what they're saying, right? You know, you're mugged by reality, but what are you going to believe them or your lying eyes, right? You've got local sensor data, right? You can diff it and you can say that centralized authority is false. And I know what is true in this local area, right?

21:51but you it's hard to form a system perspective because can they actually truly be lying about everything well are they lying about whether it rained in panama like the weather channel or something like that can you really go and check that for yourself can they be lying about everything no you need to have some decision role to figure out what is true and what is false and that's really expensive because it's essentially reinventing the centralized institution again and that leads to the you know you know the concept of gentleman amnesia by uh michael criden no okay so he has this concept which is uh which you'll like he basically says you know there's this great physicist that he marie gentleman famous noble laureate and basically michael criden said if when you go and you read the paper on any subject you know they reverse cause and effect they get everything wrong you know they're just like they just just get it wrong and because you have local knowledge and then you flip the page and you read about japan or you read about steel and you forget that they just completely messed up the area that you know you read those other stories and you have amnesia and and his friend called it gel man amnesia where anytime they wrote a story about physics marie gel man would find out all the errors and so on about it then he'd flip the page to read about japan or korea as if it was true right and the way i think about this is if you imagine, let's say the New York times or legacy media as like, or the U S government is like a centralized institution, a hub in the center, right.

23:24Then you're a spoke over here and you, you have, let's say local knowledge on, on Duke on academia, right. And I might have local knowledge on biotech and somebody else's local knowledge on Japan or something like that. But how is the Japanese person supposed to know about what's happening at Duke better than NYT? he would just go to NYT to figure out that information because he couldn't talk to you directly until social media. And now we can route around, right? Go ahead. Well, social media then changes things because you get influencers and you, you have influencers that you trust and you trust them because people you, you, you trust are following them, are, are uh retweeting them and when you yourself follow them you find out that generally what they say turns out to be uh turns out to be true or or at least at least it comports with one's worldview even right and then then the tricky part though is of course social media itself is not a panacea because uh you have um you have lots of people who are posting fake stuff on social media just for clout or what have you.

24:34Right. And so eventually what I came to the conclusion on is that, you know, the thing about Bitcoin and thinking about cryptocurrency and one of the reasons, like I was interested in your work beforehand, but then I matched it up with my interest in, in cryptocurrency is, have you heard that there's a, there's a great book on Bitcoin, just the title alone is actually pretty good. It's called the truth machine. Okay. And the reason is that I read it. No. Okay. Right. So you've heard about it. So the thing is, whether you're Democrat or Republican, Indian or Pakistani, Chinese, Japanese, you know, British or French, everybody agrees on who holds what amount of Bitcoin, which is this incredible thing to have consensus on not a trivial truth, a trillion dollar truth, a two trillion dollar truth, right?

25:18People will fight, you know, over a million dollars, a billion dollars, let alone trillions of dollars. And the same algorithms that allowed us to get consensus, because they made it very expensive, extremely expensive and challenging to fake the state of the Bitcoin blockchain. The same algorithms can get you consensus, not just over who holds what Bitcoin, but who holds what amount of stocks, what amount of bonds, what amount of any kind of property. And that actually then extends, and this is something you may not be aware of, it extends off-chain because, have you ever used a smart lock? You know a smart lock like for a door?

25:50Yes, I have, yeah. Like a lock that you can open with your phone, for example, right? Or think about a key card at a hotel. Yes. So the same technology that's used to secure electronic, you know, like digital stuff on chain can be used to digitally unlock and lock doors, right? So that means that all capital equipment, all property, every door to a room, every door to a train, a car, a tank, a plane, everything that makes a drone fly or a robot animate, everything that every machine in a factory, every camera, all of that can actually also be gated by cryptography, the same crypto as Bitcoin. It doesn't, to be clear, it doesn't cover the shirt on your back.

26:31It doesn't cover the food on your plate. Okay. There's still personal property that it wouldn't cover. What that means is we can get to consensus on who owns what piece of property for all capital equipment and all capital. And we can actually extend it further because these are like armored cars for information. So when people are trading on, for example, temperature data, right? There's oracles that sort of take the temperature data to the blockchain and then you have smart contracts on that basis, like you can do insurance contracts on that basis. So this to me struck me as something where it has implications for how we fix a society that's been affected by public lives for a long time.

27:12We develop shadow statistics on inflation, on temperature, on this, on that. We make them tradable one at a time on chain. We have contracts on the basis. We allow them to be posted pseudonymously. There's various cryptographic techniques that allow you to do that. and then we eventually start getting a picture of what is actually true and false in society by kind of editing around the edges right and we started with really the monetary system which may be the most valuable piece and then we can extend that to other kinds of facts so i know that's maybe a little technical but that is how i've been thinking about you know you you pose a problem which is a recurrence of this public lies kind of thing that causes a dysfunction the centralized organ society and here's a potential solution which is decentralized Let me pause there.

27:58You're talking about a method for extracting information, extracting true information in a context where people have incentives to lie, to falsify. Right. So I'm not familiar with the methods that you have suggested. Perhaps I can talk about other methods. Like Bayesian truth serum or something. Similar methods that are used to predict, to measure private opinion in an environment where private opinion, the distribution of private preferences differs from public opinion, which is the distribution of public preferences. So talk about that. I'm very interested. Go ahead. If there's a lot of preference of falsification, you know, these two are going to differ dramatically.

28:59Now, one way that scholars have used is ethnographic studies. Jim Scott, who died recently, the Yale political scientist, he was an expert at this. He wrote a book called Weapons of the Weak to collect his data. He just lived among rural Malays to develop their trust, to have them reveal their secrets. And he discovered that all sorts of things that they told the state, that they told the tax collectors and so on, were not true. And so that's one way. And many of the preferences that they projected publicly to and to authorities, to people outside of their immediate circle were not. So that's one thing you can do and also conduct surveys that offer anonymity in very repressive climates.

30:08Of course, let me go back to ethnographic studies. Of course, that's it's very expensive to to do. It's very time-intensive. Surveys that offer anonymity in repressive environments, it's difficult to be convincing. Yeah. So, I mean, the thing about it is... Go ahead. Yeah. Repressive regimes know that there's a great deal of preference falsification and knowledge falsification because of their repression. but they have an interest in knowing where the true sentiment and opposition are, what the truth is. So they all offer anonymity to certain groups. Is that true? I didn't know that. So basically, I didn't know that.

31:02So they actually do their own private polling. Yes, they do their own private polling. This was done in all communist regimes in East Europe, the Soviets. did this in... I did not know that. Okay. They did this in East... How do they conduct that? Was it like, how many thought criminals are there in your district, do you estimate, Comrade? Do they do it like an indirectness thing so that the person who said, I think there's lots of saboteurs here could be off the hook? Or do they literally ask it directly with an anonymous service? They did everything. The one I'm quite familiar with is the anonymous polling they conducted in Leipzig for over 40 years.

31:50Now, how they convinced in the 1950s, how they convinced a group of students, college students at Leipzig, to believe that the check marks they put on pieces of paper that did not include their names, how these would not be used somehow to punish them if they were giving the wrong answers, I don't know. But once the system got established and students at Leipzig learned that there were no reprisals, this information became valuable. And the information, not that there was no preference falsification at all in these surveys, but over time, changes were meaningful. So that's interesting. So the time series approach where they could at least detect trends up and down, even if they weren't sure about the absolute level.

32:58Yes, they could detect these trends and they, leading up to 1985 to 89, you see a dramatic increase in opposition to the Communist Party in the belief that communism was not an efficient system, that dramatic reforms were needed. So the communist leaders were aware that there was growing discontent in the population. So that's one thing you can do. Of course, it's very difficult to establish a system and develop the trust of a certain community. Google searches can be revealing. People often will, if there's, there might be no public enthusiasm for a particular candidate, but Google searches can reveal that there's actually a lot of enthusiasm, hidden enthusiasm for a candidate.

34:14Their social circle polling, this involves asking people not about their own preferences, but the preferences of their acquaintances or in their communities. Yes. So I'm filling with Bayesian truth serum, which is one variant on this. And this is also kind of relevant to prediction markets. but basically i think this is a patented technique or something like that but there's there's like maybe some unpatented version or it's coming on patent but the idea was very similar what you just said which is you ask people let's say for example it's a uh drug use like marijuana or something you ask people straight up do they use marijuana then you also give them an incentive of guessing what uh percentage of people actually use marijuana in the population and the closer they are, the more money they get.

35:14And it's like drops off symmetrically or whatever on either side. So if it's like - You reward them partly for how well they can predict the actual distribution. Yes, exactly. That's right. And under certain assumptions, of course, your people have to be smart enough to know that by predicting the distribution, they're off the hook themselves, right because they're just you know trying to predict all those other people are misbehaving right so um this is similar to uh prediction markets where you are you know trying to estimate whether something is going to happen or not um and sometimes the prediction market is really like a verification market where you're trying to see whether sentiment will you know hit a particular range, you know?

36:02Um, and, uh, you know, I, I think a lot of your concepts are good things for people in crypto to think about. Actually, there's another piece of it because your other set of books on Islamic finance, um, I often think about that with respect to Bitcoin maximalism. Do you know what Bitcoin maximalism is? No. Basically, you know, Islamic finance, wants no debt, right? And actually, Christianity also - No interest-based debt. Sorry, sorry. No interest-based debt. Correct. You're right. I'm more precise. Christianity also had prohibitions on usury or on high debt rates. And the Soviet Union had prohibitions on a different economic thing, which is profit.

36:49And what Bitcoin maximalism is, it basically says bitcoin is the only coin there shall be no other coins it essentially has a prohibition on let's call it equity issuance right any any any new coin which is created is sort of taboo in the same way that interest was under christianity or islam and um profit was under communism and uh issuance of new assets is under bitcoin maximalism and there's a similar sense in which it has a somewhat distorting effect on the economy, which is to say that Bitcoin maximalism is so anti-finance that it's very hard to do decentralized finance on the Bitcoin blockchain.

37:32So it's moved to other chains like Ethereum and Solana. Now, one of the pros of this is that they basically think everything outside of Bitcoin is a scam. And so they're okay with everything that is not Bitcoin being pushed off of the Bitcoin chain. They want a very simple economy. They want gold and only that. They don't want any tricks. They want it to be simple. And this is actually, I would argue, in part also related to our previous point, which is it's a reaction against, let's say, an administration with all the lies of the Soviet Union or maybe the Biden administration or all the past kind of things.