Governing democracy, the internet, and boardrooms

0:00:04 2024 is an important year for democracy.
0:00:07 This year, over 60 countries have and will head to the polls,
0:00:11 representing just about half of the global population.
0:00:15 But governance extends past the leaders of nation states.
0:00:17 In fact, as the revenue of some companies
0:00:20 dwarf the economies of some nations,
0:00:22 the way these entities govern themselves
0:00:26 is both more interesting and more important than ever.
0:00:27 Listen into this episode,
0:00:29 originally published on our sister podcast,
0:00:31 “Webthrough with A16Z,”
0:00:34 as they explore topics ranging from content moderation
0:00:36 to incentivizing participation
0:00:37 to new governance experiments
0:00:40 from dows to oversight boards.
0:00:41 I hope you enjoy.
0:00:50 Hello and welcome to “Web3 with A16Z.”
0:00:51 I’m Robert Hackett,
0:00:53 and today we have a special episode
0:00:55 about governance in many forms,
0:00:57 from nation states to corporate boards
0:01:00 to internet services and beyond.
0:01:02 Our special guests are Noah Feldman,
0:01:04 constitutional law professor at Harvard,
0:01:07 who also architected the meta oversight board,
0:01:09 among many other things.
0:01:11 He is also the author of several books.
0:01:14 And our other special guest is Andy Hall,
0:01:16 professor of political science at Stanford,
0:01:20 who is an advisor to A16Z crypto research,
0:01:23 and who also coauthored several papers and posts
0:01:25 about “Web3” as a laboratory
0:01:29 for designing and testing new political systems,
0:01:32 including new work we’ll link to in the show notes.
0:01:35 Our hallway style conversation covers technologies
0:01:36 and approaches to governance
0:01:40 from constitutions to crypto blockchains and dows.
0:01:43 As such, we also discuss content moderation
0:01:45 and community standards,
0:01:47 best practices for citizens’ assemblies,
0:01:49 courts versus legislatures,
0:01:52 and much more where governance comes up.
0:01:56 Throughout, we reference the history and evolution of democracy,
0:01:58 from ancient Greece to the present day,
0:02:00 as well as examples of governance
0:02:04 from big companies like meta to startups like Anthropic.
0:02:07 As a reminder, none of the following should be taken
0:02:11 as tax, business, legal, or investment advice.
0:02:14 See A16Zcrypto.com/disclosures
0:02:16 for more important information,
0:02:18 including a link to a list of our investments.
0:02:24 So I want to start with a really broad question.
0:02:25 Maybe it’s too broad,
0:02:30 but what is the right model for internet governance?
0:02:33 We have all of these companies that host platforms
0:02:36 that people participate in,
0:02:37 that they build on top of,
0:02:41 that they connect and communicate on.
0:02:44 And so a question is, what is the right way
0:02:46 to govern those platforms?
0:02:48 Right now, the most popular ones at least
0:02:50 are owned by corporations,
0:02:53 and they sort of get to call the shots there.
0:02:55 Is that the way that it should be?
0:02:57 Idealistically, in like a world
0:03:00 that we all live in and work in and play in and…
0:03:04 – May I start with a little bit on your formulation?
0:03:07 Corporations don’t exist in the ether,
0:03:09 much as we sometimes like to fantasize
0:03:12 either positively or negatively that they do.
0:03:15 They’re products of law, usually Delaware law,
0:03:16 because of the weirdness
0:03:19 of how the American federal system works,
0:03:22 and they’re governed by a plethora of obligations
0:03:24 and regulations that are not only
0:03:26 in their place of incorporation,
0:03:28 but are also state law everywhere and federal law.
0:03:30 And a lot of the duties that they have,
0:03:31 they can’t get rid of,
0:03:33 even if they tried to get rid of them.
0:03:35 So if you make and sell a product,
0:03:36 you’re liable for the bad stuff
0:03:39 that happens to people as a result of that product,
0:03:41 even if you pretend you’re not,
0:03:43 and even if you don’t want to be.
0:03:44 That’s just a given example of a duty
0:03:45 you can’t get away from,
0:03:48 that means that all corporations
0:03:50 are intensely regulated all the time already,
0:03:53 even before we get to the specific regulations
0:03:54 that are relevant, say,
0:03:57 to a user-generated content social media platform.
0:03:59 So the first thing is,
0:04:03 you’re gonna be governed by state and national laws,
0:04:04 and in a more tenuous way,
0:04:07 you’re gonna be governed by those international laws
0:04:10 that your government bothers to apply to you.
0:04:11 – And which, by the way,
0:04:12 there are plenty of governments
0:04:13 that are democratic formulated,
0:04:16 and some of these laws are presumably
0:04:19 coming from the will of the people.
0:04:23 – Yeah, and some of them are undemocratic and lousy,
0:04:23 but they still affect you
0:04:26 if you want to do business in a country.
0:04:29 So all of these things are in place.
0:04:30 That’s the first thing I would say.
0:04:34 So there already is a lot of governance
0:04:35 that comes from governments,
0:04:37 like the good old-fashioned kind of governance.
0:04:41 Then there’s a further and really important question
0:04:44 of whether there should be more governance
0:04:45 coming from governments,
0:04:46 and if so of what kind,
0:04:49 and that’s very hot topic right now at the US Supreme Court,
0:04:52 this year is deciding three different sets of issues
0:04:55 all related to what are the rights of platforms,
0:04:56 what are the rights of users,
0:04:57 what is the role of government
0:04:59 with respect to social media.
0:05:01 And until now, there’s been exactly zero
0:05:04 Supreme Court doctrine on social media companies,
0:05:05 and now we’re gonna have a whole bunch of it.
0:05:08 So this is a major transformational era.
0:05:09 Unfortunately, we don’t know the answers
0:05:10 to any of those things yet definitively.
0:05:13 We have sort of tea leaves from the oral arguments.
0:05:14 So that’s another really huge issue,
0:05:16 like how far should it go?
0:05:19 How much regulation should the states of Florida and Texas
0:05:21 be able to weigh in on content moderation?
0:05:23 – Yeah, so you’re referring to the net choice case
0:05:25 that the Supreme Court is reviewing,
0:05:28 and what’s at issue there is these states
0:05:30 would like these social media platforms
0:05:34 not to censor or moderate certain types of speech.
0:05:36 – I mean, that’s how the states would put it.
0:05:41 The companies would say that the states want to force them
0:05:44 to carry content that they may not wanna carry,
0:05:47 or wanna limit the ways that they can moderate the content
0:05:49 on what they call their platforms.
0:05:51 And so that’s, like any Supreme Court case,
0:05:52 you’ve got two sides,
0:05:55 and those are the two main perspectives.
0:05:57 And then beyond that,
0:05:59 there’s to me a super fascinating question.
0:06:02 I know Andy’s thought very deeply about this too,
0:06:06 of modes of regulation that platforms can choose,
0:06:09 first of all, for themselves and use internally,
0:06:12 then there’s potential collaborations
0:06:14 as in between different platforms
0:06:17 on collective self-regulation,
0:06:19 which is a form of governance.
0:06:22 And then there’s the weird kind of hybrid structure
0:06:25 that the Facebook oversight board represents,
0:06:26 and then in a different way,
0:06:29 the anthropic long-term benefit trust represents.
0:06:32 And those are methods whereby a company
0:06:34 creates a kind of hybrid structure
0:06:36 where there’s some independent actors
0:06:39 whom the company agrees to be governed by
0:06:41 for some limited set of purposes.
0:06:43 And that’s another really interesting
0:06:47 and to my mind kind of innovative and cutting edge
0:06:48 way of doing this,
0:06:50 which like any innovative cutting edge thing
0:06:51 has its pros and its cons
0:06:54 and is also very much in an experimental phase rate.
0:06:56 – I think that we’re already getting into something
0:06:57 very deep and interesting,
0:06:59 which is like, where are the boundaries?
0:07:02 Where do we think the real world governance ends?
0:07:05 And when or why would any type of organization
0:07:07 want to go beyond it?
0:07:08 Seems like there’s at least a few reasons,
0:07:09 but no, I’d be curious what you think.
0:07:13 I mean, there’s obviously just a sense
0:07:14 as a company or as an organization
0:07:18 that you want to do more than government has told you to do.
0:07:20 That’s like one possible motivation.
0:07:21 Seems like there’s also something
0:07:24 about the global reach of some of these platforms.
0:07:26 We know that global coordination
0:07:28 around regulation is very challenging.
0:07:30 And so in the absence of that coordination,
0:07:32 you might feel the obligation
0:07:35 or the strategic need to do something
0:07:37 to coordinate that yourself.
0:07:38 And then there’s also something,
0:07:39 this hasn’t come up yet,
0:07:42 but I think a third explanation
0:07:44 is something to do with competition.
0:07:49 And I think going back to your original question, Robert,
0:07:52 in a world where we had perfect competition
0:07:53 in a very classical sense
0:07:56 between all different types of internet services,
0:07:58 I think a lot of the governance
0:07:59 beyond traditional regulation
0:08:03 would be done by users voting with their feet.
0:08:04 And the challenge we get into
0:08:07 is the scale of many internet services
0:08:09 and the network effects,
0:08:11 put them in this kind of interesting position
0:08:13 where they’re at a very, very large scale,
0:08:15 which makes it in some ways hard
0:08:16 for users to vote with their feet
0:08:19 because they like being where their friends are.
0:08:20 But at the same time,
0:08:22 there’s still a lot of economic competition across services.
0:08:23 It’s not obvious that these companies
0:08:25 aren’t really monopolies
0:08:27 in the traditional antitrust sense.
0:08:29 And it’s put them in this weird zone
0:08:33 where there is a sense that users
0:08:37 are unhappy being locked in to any particular platform.
0:08:38 But at the same time,
0:08:42 traditional antitrust tools don’t really seem to apply.
0:08:44 And in that weird gray area,
0:08:47 it might make sense for large platforms
0:08:49 to play with additional modes
0:08:51 of giving their users the ability
0:08:54 to decide together how the platform’s gonna work
0:08:57 in a world where they can’t freely individually move
0:08:59 between platforms for the same service.
0:09:01 – And Andy, I don’t think you would disagree with this.
0:09:02 There’s also the advertisers
0:09:04 who in a perfect economic picture
0:09:06 would just go where the people were,
0:09:07 where the users were.
0:09:09 So in that sense, they seem less important.
0:09:11 But as we’ve seen in the,
0:09:13 what you might call the X files,
0:09:15 the advertisers have turned out to be major players
0:09:18 because they don’t wait to see what the users will do.
0:09:20 I mean, they also do that,
0:09:22 but they take preemptive steps
0:09:24 based on their perceptions of what might happen
0:09:26 and the reputational costs
0:09:27 and all of those sorts of things.
0:09:29 And so they’re big players too.
0:09:32 And they’re another reason why a company,
0:09:34 a platform might want to have
0:09:36 its own self-regulatory mechanisms.
0:09:37 And here I just wanna add,
0:09:39 in this super polarized world,
0:09:43 there are no neutral decisions that you could say,
0:09:44 oh, well, everybody should leave me alone
0:09:46 ’cause I made a neutral decision.
0:09:49 Whether you take content down or leave it up,
0:09:50 you’re non-neutral.
0:09:53 Whether you amplify it or don’t amplify it,
0:09:54 you’re non-neutral.
0:09:56 Depending on how much you amplify it,
0:09:56 you’re non-neutral.
0:09:57 Depending on your algorithm,
0:09:58 you’re non-neutral.
0:10:00 Everyone is sort of realized by this point,
0:10:01 certainly in this ecosystem,
0:10:03 that there’s nothing neutral.
0:10:06 And where a lot of people are mad at you
0:10:08 and there is no neutral position,
0:10:11 you might have an incentive to offload some of the decisions
0:10:14 just so people will be mad at someone else.
0:10:17 And also because users might not trust you as the platform
0:10:20 and then you might think it’s better off
0:10:22 that people will trust somebody else
0:10:23 more than they trust me,
0:10:25 even though they may not trust them fully either.
0:10:26 – I totally agree.
0:10:29 I think it’s an exercise in trying to rebuild trust.
0:10:31 I think something notice that’s really important,
0:10:35 which is like trust in tech companies is falling
0:10:36 in a lot of parts of the world,
0:10:37 not in all parts of the world,
0:10:38 but in many parts of the world.
0:10:40 But at the same time, faith in government
0:10:43 is falling in a lot of those same places.
0:10:46 And so there is no obvious actor
0:10:49 to make some of these very hard calls
0:10:53 in a neutral, procedurally fair way.
0:10:55 Where the key is that I think you,
0:10:57 as a user, you’d like to be in a position
0:10:59 where a platform decides on a piece of content
0:11:03 or what app is allowed or what transactions are allowed,
0:11:05 in a way such that even if you disagree
0:11:07 with the particular decision,
0:11:09 you still buy into the process
0:11:12 by which the decision was taken.
0:11:15 And today, I think there’s a lot of people
0:11:18 who question both a company’s ability
0:11:19 to build such a process,
0:11:23 but also a government’s ability to build such a process.
0:11:27 And in the absence of trust in either of those processes,
0:11:29 it makes sense from a business perspective
0:11:34 to try to improve your trust in society as a company
0:11:36 by finding a third option
0:11:39 for a fair way to make that decision.
0:11:40 – This sounds like a really good segue
0:11:42 into the oversight board,
0:11:46 because that is an attempt to satisfy these conditions
0:11:47 that you’re talking about,
0:11:50 to kind of offload some of this responsibility
0:11:54 onto a third party to enhance transparency and trust,
0:11:57 hopefully, and create a process
0:11:59 that is reasonable and rule-bound
0:12:02 for moderation in all sorts of decision-making
0:12:03 behind the scenes.
0:12:05 Well, let’s talk about the formation of that
0:12:08 and the needs that led to its creation.
0:12:11 – I think the core background situation
0:12:15 that then Facebook was facing, now Metta,
0:12:20 when Mark Zuckerberg decided to create the oversight board,
0:12:23 was the recognition
0:12:25 that there were some very, very hard
0:12:27 content moderation questions.
0:12:29 They were still pretty simple at the time.
0:12:30 They were sort of, “Do you leave up content?
0:12:31 “Do you take it down?”
0:12:35 They were not as fine-grained as they now have become,
0:12:38 but in which he might have a strong intuition
0:12:40 about what the right answer was,
0:12:42 but reasonable people could differ.
0:12:46 And I think the insight that he was able to see
0:12:51 is that actually there might be a right and wrong decision.
0:12:52 I’m not a relativist.
0:12:54 I don’t think that there are always two decisions.
0:12:57 No, there might be a right and wrong decision on this,
0:12:59 but reasonable people could differ
0:13:01 about what the right decision is.
0:13:03 And it probably didn’t matter that much
0:13:06 from the standpoint of platform governance
0:13:09 which of those two decisions were taken.
0:13:12 But if he made it, it was gonna take him a lot of time.
0:13:13 It was gonna take him a lot of effort.
0:13:15 He had no special expertise
0:13:17 in any of the underlying questions.
0:13:20 And people were gonna be really angry at him
0:13:23 and distrust him so that no matter what he decided,
0:13:26 it would decrease the legitimacy of Facebook
0:13:27 as it then was.
0:13:30 And so the idea that I proposed to him
0:13:33 and that I think resonated for him was,
0:13:36 “Maybe this decision doesn’t have to be made by you.”
0:13:37 In the first instance,
0:13:39 the company is going to make a decision.
0:13:39 They’re gonna have a policy
0:13:42 and they’re gonna try to file that policy.
0:13:43 But if it’s controversial,
0:13:46 why not create a group of independent experts
0:13:50 and bring those questions, the hardest questions, to them?
0:13:52 Have them answer the question
0:13:54 in light of Facebook’s principles,
0:13:58 international principles, commonsense good judgment.
0:14:00 Have them explain their decision
0:14:04 and see if it worked.
0:14:06 And I would say that the secret sauce there
0:14:09 is the explanation of their decision.
0:14:11 The secret sauce is that if you’re answering
0:14:12 a really hard question where you admit
0:14:14 there are different possible answers,
0:14:17 you should explain why you’re doing it.
0:14:20 And that explanation, that practice of reason giving
0:14:24 has the capacity to confer a lot of legitimacy
0:14:25 when it comes to hard decisions.
0:14:27 ‘Cause I think it reassures people
0:14:30 that thoughtful people gave thought to this,
0:14:31 that they had some degree of expertise,
0:14:33 that they deployed that expertise,
0:14:35 and that they tried their best.
0:14:36 It’s not a magic solvent.
0:14:39 Some people will still be really mad about the decision,
0:14:41 but everyone will be able to say, I hope,
0:14:45 this is the closest thing to a fair decision-making process
0:14:46 that we could have come up with.
0:14:48 And my last point on this is,
0:14:51 obviously this idea of legitimacy and fairness
0:14:53 did not come out of the blue.
0:14:56 It’s borrowed from real-world institutions,
0:15:00 especially courts, that often are not elected.
0:15:02 Often their decisions aren’t appealable.
0:15:04 If it’s like a Supreme Court or a High Court,
0:15:05 they’re the last word.
0:15:06 They’re not always right.
0:15:08 They’re often controversial.
0:15:09 People get mad about them.
0:15:13 But the institutions on the whole
0:15:14 have a fair amount of legitimacy.
0:15:16 I mean, it’s a funny moment to be saying it now
0:15:18 when the Supreme Court has, you know,
0:15:21 done a lot of self-harm on the legitimacy front,
0:15:23 but the Supreme Court still has much greater legitimacy
0:15:25 in the eyes of Americans.
0:15:27 And that’s even after a range of other things
0:15:29 that the Supreme Court has done that they don’t care,
0:15:31 that the majority doesn’t care
0:15:32 that it’s undermined their legitimacy.
0:15:35 So the idea that you can get legitimacy
0:15:39 from reason decision-making was borrowed from that realm
0:15:42 and applied experimentally to the Opposite Board.
0:15:45 And I think on the whole it’s working,
0:15:47 not perfectly, but that it is working.
0:15:49 But that’s a topic for further discussion.
0:15:51 – You mentioned that this is a funny moment
0:15:54 to be talking about the legitimacy of the courts.
0:15:55 Now, of course, it’s an institution
0:15:57 that’s hundreds of years old,
0:16:01 but based on the way things have shaken out,
0:16:03 is there anything you would have done differently
0:16:08 in your conception of this sort of analog body
0:16:10 to a court system for–
0:16:12 – Well, actually, if I can answer in the reverse way,
0:16:16 there is something that I’m actually kind of proud of,
0:16:18 which differentiates us from the Supreme Court.
0:16:22 We have, the Oversight Board is a pretty big body of people
0:16:27 who decide cases, sometimes all of them,
0:16:32 but sometimes in subsets, and it isn’t partisan.
0:16:35 So it’s not set up so that when you change the composition
0:16:38 by a vote or two, everything flips on you,
0:16:40 which is what happens at the US Supreme Court.
0:16:43 And so, and then retirements and deaths.
0:16:47 And so that kind of, it can all change in a second.
0:16:49 It’s highly partisan and it’s very responsive
0:16:52 some of the time to political changes
0:16:54 are all features that we intentionally
0:16:56 did not put in the Oversight Board
0:16:59 to lower the temperature around its decision-making process.
0:17:03 – Is this something that the government should take
0:17:04 into account?
0:17:06 Like if you were setting up a new government
0:17:10 and a new court system, would you make changes of this sort?
0:17:12 – Yes, absolutely.
0:17:14 I mean, it’s crazy that the Supreme Court,
0:17:17 I mean, you have to go back to 1787.
0:17:19 People died a lot younger.
0:17:21 Being on the Supreme Court was not that good a job.
0:17:23 A bunch of people who had that job would quit
0:17:25 and go back to private law practice
0:17:27 ’cause they couldn’t stand the job and it didn’t pay enough
0:17:29 and you had to get on a horse and ride around
0:17:32 in your cases all, it was just wasn’t that pleasant.
0:17:36 Now people live longer, they get into that job
0:17:39 and they never leave, they leave the court
0:17:40 when they die a lot of them, which is very,
0:17:43 I mean, it’s unhealthy, it’s not good.
0:17:46 And we also have a much more polarized politics
0:17:48 than they certainly did at the beginning,
0:17:50 although there have been periods of intense polarization
0:17:51 in US history.
0:17:53 So yeah, if you were designing it from scratch now,
0:17:56 you would have staggered terms with time bound.
0:18:00 It wouldn’t be the luck of Justice Scalia died,
0:18:02 Barack Obama was president, but Mitch McConnell
0:18:04 was able to block the confirmation hearings.
0:18:06 I mean, that kind of crazy town,
0:18:09 hardball politics should not affect the composition
0:18:10 of the Supreme Court.
0:18:14 And since one of my jobs is that I watch the court closely,
0:18:16 part of my job shouldn’t be having to have
0:18:20 like a not quite expert, but pretty good amateurs knowledge
0:18:22 of like the health of the individual justices,
0:18:24 if they’re diagnosed with a certain cancer,
0:18:25 what are the probabilities that they’ll live
0:18:30 and how long, like that’s distasteful and also like absurd
0:18:32 that that should be part of our way we decide things.
0:18:34 We shouldn’t be thinking in those terms.
0:18:36 – I wanna raise one thing that I wanna make sure
0:18:38 we spend more time on how the board is doing
0:18:40 ’cause I think there’s a lot of interesting
0:18:42 different aspects of that to unpack.
0:18:44 One thing I just wanna raise that no one I have talked
0:18:47 a lot about in the past and lots of other people
0:18:50 thought about that I think is important is one,
0:18:53 one important difference between say the Supreme Court
0:18:57 or US courts or courts in general and the oversight board
0:19:01 is sort of whose laws they’re interpreting.
0:19:06 So Noah referenced this earlier that they’re basically,
0:19:08 they’re considering cases on the basis
0:19:11 of Facebook’s principles or Meta’s principles.
0:19:14 And I think that’s the logical way to start.
0:19:16 But one thing people obviously then point out is,
0:19:21 well, the Supreme Court isn’t referencing
0:19:23 single corporation’s rules.
0:19:27 It’s referencing democratically written rules
0:19:29 that the legislature writes.
0:19:33 Nevertheless, it’s an interesting I think thing
0:19:34 to consider in the long run.
0:19:37 Can you get a sufficient amount of legitimacy
0:19:41 as Noah called it from a court online
0:19:44 that’s interpreting the rules of the company
0:19:46 when part of the concern is whether the company
0:19:49 is setting the rules in the right place or not.
0:19:52 And so one thing people have talked quite a bit about
0:19:54 but I don’t think we’ve cracked
0:19:58 is how would you democratize the legislative process
0:20:00 as well so that something like the oversight board
0:20:03 would be making decisions with respect
0:20:06 to democratically written policies.
0:20:07 – Yeah, this is super interesting.
0:20:10 And let me add one more twist
0:20:12 to the way things have been playing out.
0:20:14 ‘Cause this is something that has not been broadly discussed
0:20:17 or broadly covered except in a very narrow group
0:20:21 of specialists, mostly academics and NGOs
0:20:22 who follow the oversight board.
0:20:24 The oversight board has been really worried
0:20:27 about the thought that they’re following Meta’s rules.
0:20:29 They’re more ambitious than that.
0:20:31 They wanna follow something more democratic
0:20:33 but they can’t use the laws of any particular country
0:20:36 ’cause there isn’t one particular country that’s binding.
0:20:37 So what they’ve been doing is they’re claiming
0:20:40 to apply international law,
0:20:41 the international law of free speech
0:20:42 which is not anywhere near as democratic
0:20:44 as the laws of a democratic country
0:20:47 but it is sort of second order democratic
0:20:51 because these are treaties that were enacted by countries
0:20:54 some of which are many of which were democratic
0:20:58 and which all countries recognize as legally binding.
0:21:02 Now, the upside of that is that it sounds
0:21:03 a lot more legitimate than we’re deciding
0:21:06 these important questions based on Meta’s rules
0:21:08 and it’s also a little more democratic.
0:21:10 The downside of it is international law actually
0:21:12 isn’t that well set up to answer these questions
0:21:15 ’cause it’s not designed to govern what a platform should do.
0:21:19 It’s designed to govern what countries should do.
0:21:20 That’s why it’s international.
0:21:22 It’s the law of stuff between countries.
0:21:25 So the oversight board has had to be really creative.
0:21:28 They’ve basically relying on some work by scholars.
0:21:31 They’ve basically created a whole new body
0:21:34 of what they call international law
0:21:37 that they say binds Meta and binds should in principle
0:21:39 bind other platforms.
0:21:41 And they’re out there making this law
0:21:43 and there’s no one to tell them no.
0:21:46 So I just supervised a student who wrote a dissertation
0:21:48 saying this is unjustified.
0:21:51 But my response to that was always that might be true
0:21:53 but who’s gonna stop them?
0:21:55 And they’re trying to do that.
0:21:57 So that’s a bid that is out there
0:21:58 and it’s interesting to watch.
0:22:00 And as I say, it hasn’t been very well covered.
0:22:02 But I also, I do wanna get back to the point
0:22:04 that Andy’s talking about which is
0:22:06 what are the experimental creative things
0:22:08 that a big platform like Meta could use
0:22:11 to bring democratic input to bear
0:22:13 on some of the hardest decisions that it makes?
0:22:16 – Well, it might be good to start with the history
0:22:19 which is that this is something apparently
0:22:21 Mark Zuckerberg has thought about for a long time
0:22:23 because back in the 2000s
0:22:26 when the platform was still pretty young,
0:22:30 they had a sort of a disagreement among users
0:22:33 about what the new community standards should be.
0:22:38 And Mark held a global referendum on the platform
0:22:40 and I mean, it’s super interesting.
0:22:41 You can actually, on the internet archive,
0:22:44 you can go and see the video he posted
0:22:47 encouraging all the users on Facebook
0:22:49 there were about 300 million at the time to vote.
0:22:53 And what they learned through that process was actually
0:22:55 it’s really hard to get people to vote.
0:22:58 So out of the 300 million users,
0:23:01 I think something like 60,000 of them voted.
0:23:04 And the reason for that is probably because
0:23:08 most people are not logging on to any part of the internet
0:23:10 to do the hard work of governance.
0:23:12 They’re logging on to have fun
0:23:14 or see their friends or whatever.
0:23:17 And in addition, the decisions being considered
0:23:18 where they’re pretty abstruse.
0:23:20 Like you had to read these like hundred page documents
0:23:21 and stuff.
0:23:24 So I think it was a really innovative idea
0:23:27 that suggested early on this idea
0:23:31 that there’s a way to potentially make decisions
0:23:33 about a platform democratically,
0:23:34 but it turned out to be challenging
0:23:36 to make work in practice.
0:23:38 And so now if we fast forward to today,
0:23:43 yeah, I think lots of people are interested in ways to set
0:23:46 at least some kind of broad, simple policies
0:23:48 where the values that users hold are relevant
0:23:53 to the decision through some kind of democratic process.
0:23:54 No one has really figured out
0:23:57 how to solve this core problem of participation
0:24:00 as well as informed participation,
0:24:04 as well as if it’s a platform that’s global in scope,
0:24:06 how do you actually get people in different countries
0:24:09 who speak different languages to even work together
0:24:11 to figure it out?
0:24:15 What if they have radically contrasting values or goals?
0:24:17 There’s a lot of really hard questions,
0:24:18 but I think there’s, yeah,
0:24:21 there’s at least two veins of experiments
0:24:23 worth going deep on.
0:24:25 One, which I think no-
0:24:27 – By the way, voter participation rates are a problem,
0:24:30 not just in online voting, but in, you know,
0:24:31 IRL voting too.
0:24:32 – That’s true, that’s true.
0:24:33 Although the participation-
0:24:35 – It’s a giant thing and you could make a mandatory, you know,
0:24:38 you can make it, you open Instagram
0:24:40 and you can’t use Instagram unless you vote on these issues,
0:24:43 but then you’re not gonna get highly informed voting, right?
0:24:46 I mean, these ones that he’s describing,
0:24:47 just to make them concrete,
0:24:49 ’cause I think they’re two kind of good examples.
0:24:50 One is the problem that,
0:24:52 I don’t know if people who are listening
0:24:53 all have heard of this or not,
0:24:54 because it may be a little dated,
0:24:57 but they call the Boatemic Boatface problem.
0:25:00 – Oh yeah, this was Reddit, right?
0:25:01 The, didn’t Reddit-
0:25:02 – It was the British government.
0:25:03 The British government decided,
0:25:06 they were gonna use, you know,
0:25:08 this new amazing thing, the internet,
0:25:12 to gather votes on what they should name a new battleship.
0:25:17 And, you know, the winning vote getter was Boatemic Boatface.
0:25:18 – Which I think we can all agree
0:25:20 is a fantastic name for battleship.
0:25:21 – I don’t see the problem.
0:25:23 – Yeah, Her Majesty’s Ship,
0:25:24 I guess it was then Her Majesty’s Ship,
0:25:25 Boatemic Boatface,
0:25:28 somehow it didn’t land for the Royal Navy.
0:25:30 And then the other example that I like to use,
0:25:31 and Andy mentioned this,
0:25:34 is let’s imagine you’re setting a nudity policy.
0:25:36 Well, there’s one policy
0:25:39 that works in the South of France on a beach,
0:25:42 which is more permissive than what we have in the US.
0:25:44 And then there’s Saudi Arabia,
0:25:47 which is a lot less permissive than we have in the US.
0:25:49 And it’s not obvious that either of those
0:25:52 should control the entirety of the platform.
0:25:54 It’s not obvious that our model
0:25:55 should follow the platform either.
0:25:59 It’s mostly a default that, you know,
0:26:01 a lot of these platforms are US-based companies,
0:26:04 and so US-based standards end up controlling.
0:26:04 Americans think, oh,
0:26:06 those are the reasonable standards.
0:26:08 But of course, there’s no inherent reasonableness answer
0:26:10 to the question of like how much skin you should show.
0:26:12 That’s a classic example of something
0:26:15 that different cultures are equally confident
0:26:16 that their way of doing it
0:26:18 is the only right way to do it.
0:26:20 So that’s a genuine challenge.
0:26:22 And then that leads to thinking,
0:26:23 well, maybe we should have regions
0:26:26 with different rules and different regions.
0:26:28 And it turns out it’s harder to do that
0:26:31 on a global platform than you’d think.
0:26:33 And most of the global platforms
0:26:35 have tried as much as they can to avoid that
0:26:37 because they don’t want this kind of
0:26:40 extremely balkanized platform
0:26:42 where it all depends on where you logged in from.
0:26:44 And, you know, if you have a VPN,
0:26:45 you can get different standards.
0:26:48 And you can imagine what a mess it can quickly become.
0:26:50 So those are part of the challenges.
0:26:54 And then the last part of the challenge is
0:26:56 all of our examples of democracy.
0:26:58 Andy, you can correct me if this is wrong
0:26:59 because you’re the political scientist,
0:27:01 but at least the ones that come to my mind,
0:27:04 all of our examples of democratic governance,
0:27:09 assume a denominator that is the relevant group of people.
0:27:10 It’s the people who live in your town
0:27:14 are gonna vote on town policy or your state, your country.
0:27:17 We don’t really have true democracy at the global level.
0:27:21 And so that raises the question of who’s the denominator?
0:27:25 Who’s the relevant community to vote on these issues?
0:27:26 And maybe it’s the users,
0:27:29 but a lot of issues affect non-users.
0:27:33 And so, you know, is it everybody, including non-users?
0:27:34 Is it a subset of the users?
0:27:37 That problem is, I think, the very first problem
0:27:38 to grapple with.
0:27:41 And in a lot of ways, it’s the most challenging.
0:27:43 – Let’s unpack the Voting McBoat Face thing for a second
0:27:45 ’cause I think it’s really important.
0:27:49 It’s a great example of two or really three
0:27:52 deeply related problems in the use of voting
0:27:54 to make collective decisions.
0:27:57 The first is just the lack of participation.
0:28:00 So very few people actually voted
0:28:02 on the Voting McBoat Face decision
0:28:04 because why would most people pay attention
0:28:06 or even know about it?
0:28:09 Second, which is closely related to the lack of participation
0:28:13 is that who then are the weirdos who choose to vote
0:28:15 when not very many other people are voting?
0:28:18 It’s people who have very extreme weird preferences
0:28:21 like the kind of people who think it would be hilarious
0:28:23 if there was a vote named Voting McBoat Face.
0:28:26 And so those two things go together quite tightly.
0:28:29 Low participation and then selection
0:28:33 for unusual, unrepresentative, extreme views
0:28:35 or preferences or trolling.
0:28:38 So I think that’s like an important thing to understand.
0:28:42 And then the third, which is also deeply related,
0:28:46 is just the general problem
0:28:49 that you’re asking people to vote on decisions
0:28:52 and they have no skin in the game.
0:28:55 And that, again, lowers the incentive to participate
0:28:57 and lowers the incentive for people
0:28:59 with representative views to participate.
0:28:59 And I think–
0:29:01 – What if you had to serve on that ship?
0:29:03 You might care a little bit more.
0:29:06 – Well, that kind of gets to Noah’s denominator point.
0:29:08 I think there’s a very broad challenge.
0:29:11 It’s much broader than just democracy,
0:29:13 which is how do we govern things
0:29:18 that are not economic in nature in ways that are smart
0:29:20 when the people were asking to do the governance
0:29:22 that don’t have skin in the game.
0:29:26 And that was the problem with open AIs governance.
0:29:28 It’s been a challenge with university governance.
0:29:31 The trustees in some sense don’t have skin in the game.
0:29:33 – You measure the skin by how much money they put in.
0:29:35 They would say, I mean,
0:29:36 not that I’m always sympathetic to them,
0:29:37 but the donors– – Good point.
0:29:39 – We’re the only ones with skin in the game.
0:29:40 We’re donating the money.
0:29:42 All you guys are doing is taking the salary.
0:29:44 – That’s a good point, that’s a good point.
0:29:45 – That’s what they say.
0:29:46 I mean, I’m not saying I’m on board with it.
0:29:47 Anyway, go on, yeah.
0:29:50 And when we look at all these voting problems,
0:29:53 if you’re asking people to vote on things
0:29:55 that don’t clearly affect them,
0:29:59 that’s a reason you might not get very informed participation.
0:30:02 Furthermore, when you ask large groups of people to vote,
0:30:05 you have this problem we call the paradox of voting,
0:30:08 which is that even if you do care a lot,
0:30:11 even if you do feel like you have skin in the game,
0:30:13 if lots of other people are voting,
0:30:15 then you know your vote doesn’t matter.
0:30:18 It’s not gonna swing the outcome.
0:30:22 And so you could care infinitely much about the decision,
0:30:23 and yet your incentive to actually pay attention
0:30:25 about it was very, very small.
0:30:28 So we put those things together and we ask,
0:30:31 okay, how are you gonna do something democratic
0:30:35 to make a big decision and a big online platform?
0:30:37 I think you have to start with this fundamental set
0:30:40 of problems that are, you wanna get people to participate,
0:30:42 you want them to pay attention,
0:30:44 you want them to feel like it matters,
0:30:46 and how on earth do you do that?
0:30:49 And I think there’s been several directions
0:30:51 of experiments that are pretty fascinating.
0:30:53 One, which Noah alluded to earlier,
0:30:56 is what Anthropic as well as Meta and others
0:30:59 have been playing with where you don’t use voting,
0:31:02 you use these things called citizens assemblies.
0:31:04 And then the other is what’s going on in crypto,
0:31:08 where you’re experimenting with different types of voting
0:31:10 that try to draw in only the people
0:31:13 who have some sense of skin in the game
0:31:14 through the tokens that they hold.
0:31:17 So those are I think two sets of experiments
0:31:18 we can learn a lot from.
0:31:21 – And I would just add one more framing thing here.
0:31:23 Most forms of democratic government
0:31:25 in the history of the world,
0:31:27 I tried to address these questions by saying,
0:31:30 we’re not gonna have direct governance
0:31:33 where every single person has to vote on every issue.
0:31:35 ‘Cause that’s also like a time consuming thing.
0:31:37 And so one way to solve both the,
0:31:40 how much time do you put in and the,
0:31:42 how much do you care and the skin in the game
0:31:45 is to have representative decision-making.
0:31:46 This is the move from like ancient Athens
0:31:49 where everyone can show up and vote every,
0:31:52 to be clear, every free man can show up as a citizen
0:31:56 and vote in the assembly to a system where you elect people
0:31:58 and your representatives are professionals
0:32:01 or quasi-professionals and they do the decision-making
0:32:04 and the voting, they have a skin in the game.
0:32:07 They have the job of caring about getting reelected
0:32:09 based on whatever incentives they have.
0:32:11 And that gives them the incentive to try to guess
0:32:13 or figure out what you’re thinking.
0:32:16 And one of the really interesting things about the internet,
0:32:18 and I think it’s important background to when we dive
0:32:20 into both crypto and citizen assemblies,
0:32:22 is that from its early days,
0:32:24 the internet has been broadly speaking,
0:32:26 the ideology of the internet
0:32:27 or the ideologies of the internet
0:32:30 have been really skeptical
0:32:32 of the idea of elected representatives.
0:32:34 There’s this impulse.
0:32:36 And it’s really interesting to explore why.
0:32:38 Is it because of hacker culture?
0:32:39 Is it because of coder culture?
0:32:41 Is it the personality type of the people
0:32:43 who first got really good at using the internet
0:32:44 before everyone used it?
0:32:46 I mean, there’s a lot of possible explanations.
0:32:48 But there’s been a deep impulse
0:32:50 to get around the idea of elected representatives.
0:32:54 And both of these experiments, the citizen assemblies
0:32:57 and the crypto or DAO or organizations
0:33:01 are attempts to like say, we are gonna reinvent the wheel.
0:33:03 Like we’re not gonna do the thing
0:33:05 that all democracies have done,
0:33:07 namely rely on elected representatives.
0:33:09 We’re gonna do something different and better.
0:33:11 And that’s kind of appealing
0:33:13 because especially if you have some skepticism
0:33:14 in our elected representatives,
0:33:17 it’s a good time to try to reinvent what they’re doing.
0:33:19 But there’s also a reason to be modest
0:33:22 when all of the countries in the world
0:33:25 that use a certain technology, namely democracy,
0:33:27 have converged to a first approximation
0:33:29 on one solution to this problem,
0:33:31 namely elected representatives.
0:33:32 And you’re like, nah,
0:33:33 they don’t know what they’re talking about.
0:33:34 We can do better.
0:33:35 Maybe you can.
0:33:38 It’s definitely worth trying.
0:33:41 – So you wanna have like some degree of modesty
0:33:43 about the probabilities of coming up with something
0:33:46 that in 2,500 years of thinking about this,
0:33:49 no one has yet really managed to solve in any other way.
0:33:51 – I wanna build on that quickly
0:33:52 ’cause I think it’s really important.
0:33:55 And I completely agree with Noah 1,000%.
0:33:59 There’s something about technologists and people online
0:34:02 that makes them really crave direct democracy.
0:34:04 And one of the key claims they make
0:34:08 is that technology is going to allow for direct democracy,
0:34:11 that it didn’t work in the physical world
0:34:12 because it was too burdensome.
0:34:13 – This is gonna be my question.
0:34:16 Like does the internet, is it a step change?
0:34:17 Like have we entered a new era
0:34:21 where based on the tech that’s available now,
0:34:22 this is conceivable.
0:34:25 – There are no certainties in the social sciences,
0:34:28 but this is as close to certainty as I’m willing to go
0:34:31 is the technology is gonna make no difference
0:34:32 to this problem.
0:34:35 And this is just the way it is.
0:34:36 – It goes back to the paradox of voting
0:34:38 I was talking about.
0:34:42 It’s very burdensome to be asked to understand
0:34:45 and become informed on every possible decision
0:34:47 that a group is going to make.
0:34:50 And you might think, and this is I think the fallacy
0:34:53 that a lot of people in tech have committed,
0:34:58 you might think that having more quote unquote democracy
0:35:01 by putting more votes to the people
0:35:05 is gonna get you a more empowered user base,
0:35:07 but the opposite is in some sense true.
0:35:09 The more things you ask them to do,
0:35:11 the more burdensome it is for them to do it,
0:35:14 the less that they’ll choose to do it.
0:35:17 And that then creates a really important
0:35:19 and dangerous vacuum because now you have
0:35:23 very few people voting and then interest groups
0:35:27 can come in and capture the decision making process
0:35:28 at a relatively low cost
0:35:31 because there’s not that many other votes out there
0:35:32 for them to compete with.
0:35:36 And this is I think fundamentally why yeah,
0:35:38 no successful society on earth has stuck
0:35:41 with direct democracy for very long.
0:35:43 And it’s why technology is not gonna solve the problem
0:35:45 because it’s actually not a problem
0:35:48 about voting in person being too difficult
0:35:49 or anything like that.
0:35:53 It’s a problem of information acquisition and analysis.
0:35:55 And I’ll just note on that point,
0:35:58 people have also claimed that we’ll be able
0:36:01 to provide voters with all the information
0:36:04 they need to make every decision through some kind of app.
0:36:06 I’ve had a lot of undergrads in my office pitching me
0:36:09 on the next app that’s gonna inform us all about politics
0:36:12 so we can vote on everything ourselves.
0:36:13 And the problem is that information
0:36:15 or that data is not enough.
0:36:16 You actually need to analyze it
0:36:18 and then decide what to do.
0:36:20 And that’s actually quite difficult.
0:36:25 And so being buried in data on what your community
0:36:29 is voting on is just not enough to help you do this.
0:36:32 And so I think it’s a completely fundamental problem
0:36:35 that goes way beyond technology that has to be solved
0:36:38 through other models besides direct democracy.
0:36:43 And I’ll just note almost all of the big voting DAOs
0:36:47 and crypto now have some form of representative democracy
0:36:48 as a result of this.
0:36:51 If you’re listening to this and you think,
0:36:53 no, that, you know, no one and you’re wrong,
0:36:54 you know, like technology–
0:36:55 – Yeah, you guys are just elitists.
0:36:56 Come on, let the mob rule.
0:36:57 – Yeah, exactly.
0:37:00 What I would say is if you’re an aspiring technologist,
0:37:01 ask yourself, what’s your vision
0:37:03 for the company you’re gonna found?
0:37:06 And in 99.9% of cases, people are like,
0:37:07 oh, I wanna found the company
0:37:08 and then I wanna have founder shares
0:37:10 and I wanna control the decision-making.
0:37:12 And then you say, well, why?
0:37:13 And they say, well, because otherwise,
0:37:16 like the VCs will first get involved,
0:37:17 then they’ll be shareholders.
0:37:18 They’re gonna dilute my votes
0:37:19 and they’re gonna take the company
0:37:20 and they’re gonna do all kinds of stuff.
0:37:22 I don’t wanna do with it.
0:37:24 And I’m the one who knows, ’cause it’s my company,
0:37:26 I spend the most time, I don’t, I care about it.
0:37:28 That’s exactly the problem that Andy’s describing.
0:37:31 So I always find it amazing when, you know,
0:37:33 those students come in and say that ’cause I say to them,
0:37:34 oh, are you imagining a company
0:37:36 where you will have no special input
0:37:37 into how the company runs?
0:37:39 You know, and every decision will be made collectively
0:37:40 and they’re like, are you crazy?
0:37:41 Of course not.
0:37:42 And I’m like, well, then there’s a contradiction
0:37:43 in what you’re proposing.
0:37:46 So it’s the, you know, the intuition
0:37:47 that technology will interfere.
0:37:49 You can easily fix that intuition
0:37:50 by just asking yourself,
0:37:52 when you get rich with your own company,
0:37:54 do you want to be in charge of it?
0:37:55 And in my experience,
0:37:58 almost everybody thinks the answer that is yes.
0:38:00 Although there are some tweaks, right?
0:38:03 I mean, so another thing that Anthropic did,
0:38:04 and I know about this
0:38:06 ’cause I was involved in advising them on it,
0:38:09 is they’ve created a long-term benefit trust,
0:38:13 which is a trust that appoints at first just a few,
0:38:15 but eventually will appoint a majority of members
0:38:18 of the board of directors of Anthropic.
0:38:20 And the people on the trust
0:38:22 who are going to appoint those board members
0:38:25 are not themselves shareholders.
0:38:27 So what Anthropic is doing there is,
0:38:28 it’s not the full open AI thing
0:38:32 where the parent board was completely made up of people
0:38:33 who had no stake in the company,
0:38:35 which blew up as we know.
0:38:38 This is a more modest in-between version
0:38:39 where there will still be shareholders
0:38:41 on the board of directors.
0:38:44 But the idea is to have some external check.
0:38:45 So it’s a little bit better.
0:38:46 But that’s very rare.
0:38:48 – So let’s talk about the example of Anthropic.
0:38:50 You’re saying that in this case,
0:38:52 there are actual shareholders
0:38:53 that are able to serve on the board
0:38:58 and that this is different from Meta’s oversight board.
0:39:01 And it’s also different from open AI
0:39:03 where people did not have a stake in the company.
0:39:07 And so maybe there will be a more representative
0:39:09 collection of views in there
0:39:11 for keeping the concern going
0:39:15 and also mitigating risks that could arise.
0:39:17 – Yeah, it’s designed to avoid problems on both sides.
0:39:19 So take the open AI problem.
0:39:21 You wanna have some, you know, break glass measure.
0:39:23 And so the break glass measure
0:39:26 was they created this nonprofit entity.
0:39:30 And that was it’s the overarching body
0:39:33 that the board of directors belong to.
0:39:34 And then underneath that board of directors
0:39:36 was the for profit entity
0:39:39 that was that we think of as open AI,
0:39:41 the actual company that does stuff.
0:39:44 If the people on the board of directors
0:39:47 of the nonprofit who had no financial stake at all
0:39:49 believe the company was going the wrong way,
0:39:50 they could break the glass
0:39:53 and fire the management of the company.
0:39:54 And they did that.
0:39:57 They believe that things were going terribly awry
0:39:58 that this was bad for the world.
0:39:59 And they broke the glass
0:40:01 and they fired the directors of the company,
0:40:02 the management of the company.
0:40:06 – Until it backfired and the glass went back in their face.
0:40:06 – Exactly.
0:40:08 They didn’t realize that what would then happen
0:40:11 would be that Sam Altman would then say,
0:40:12 well, I’ll just go to Microsoft
0:40:15 and I’ll take every single employee in the company with me.
0:40:17 And then the employees all went online
0:40:19 and signed up that they would go with him,
0:40:20 which is maybe not so shocking
0:40:22 because of course you’re gonna agree to jump ship.
0:40:24 And so then it backfired.
0:40:27 And then the members of that board directors had to resign.
0:40:30 So what you need in the real world
0:40:32 is something that doesn’t go that far
0:40:37 where there are people who are not financially incentivized
0:40:38 who serve on the board.
0:40:40 But there are also people on the board
0:40:42 who understand the real world
0:40:44 and who do have financial incentives
0:40:46 and they develop a relationship with each other
0:40:48 so that it is possible for the people
0:40:50 who are worried about a problem
0:40:51 and want to break the glass
0:40:54 to do so in a thoughtful and responsible way
0:40:57 where it won’t blow up in their faces.
0:41:01 So that’s what the long-term benefit trust is trying to do.
0:41:06 And it does give some real reasonable responsibility
0:41:08 to people who don’t work for the company.
0:41:12 I mean, in that sense, it’s a kind of cousin of the oversight board.
0:41:13 But it’s different from the oversight board
0:41:15 ’cause those folks don’t have day-to-day
0:41:17 decision-making responsibility.
0:41:18 They’re there at the level of oversight
0:41:21 and break glass at an emergency.
0:41:24 Whereas the oversight board is meant to be an everyday,
0:41:26 there are hard problems that Metta is dealing with
0:41:28 and the oversight board weighs in on that.
0:41:30 So I would call them cousins.
0:41:32 They have some conceptual similarities,
0:41:36 but they’re not siblings and they’re definitely not twins.
0:41:38 Hey, it’s deaf.
0:41:42 Look, people will endlessly debate and record podcasts
0:41:45 on the most important ingredients to success.
0:41:47 Is it technical talent?
0:41:48 Is it timing?
0:41:49 How about hard work?
0:41:50 Luck?
0:41:52 Well, in my opinion,
0:41:54 one of the most critical ingredients
0:41:56 is the power of effective communication.
0:42:00 And that’s why I recommend the Think Fast Talksmart podcast
0:42:03 from the Stanford Graduate School of Business.
0:42:05 Host and Stanford lecturer, Matt Abrahams,
0:42:07 sits down with the experts.
0:42:09 To discuss the best tips to help listeners
0:42:11 unlock their verbal intelligence,
0:42:13 whether it’s to excel in negotiations,
0:42:16 make a wedding toast or work presentation,
0:42:20 or even communicate better with AI chatbots.
0:42:22 And if you want to sneak peek,
0:42:27 Matt’s also got a great TED talk with over 1.5 million views.
0:42:29 So what are you waiting for?
0:42:32 Check out Think Fast Talksmart every Tuesday
0:42:33 or every Get Your Podcast,
0:42:36 whether that’s Apple Podcasts, Spotify, iHeart,
0:42:37 or even on YouTube.
0:42:44 – It seems a little bit interesting
0:42:46 that you have a certain set of people
0:42:48 who have this opportunity working side-by-side
0:42:50 with people who are, I assume,
0:42:51 just getting a regular salary.
0:42:54 And then how you get those people to really care about,
0:42:56 you know, the decisions that are being made.
0:42:58 – They have to care based on their reputation.
0:43:00 And this was relevant at the Oversight Board too.
0:43:02 So how do you get the Oversight Board members
0:43:05 actually to care that they do a good job?
0:43:06 – It’s not so simple,
0:43:09 but the short answer is it’s nobody’s full-time job.
0:43:13 And the jobs that they have are as activists
0:43:17 or as scholars or as people who care about free speech.
0:43:19 And so the idea was that they will care a lot
0:43:21 about their own reputations.
0:43:24 And that will give them the incentive to do a good job.
0:43:25 It’s not perfect,
0:43:27 but it was the best we could come up with.
0:43:29 If it was their full-time job,
0:43:32 their incentives might be too closely allied,
0:43:35 you know, with the power of their institution.
0:43:37 And they still probably want to do a good job,
0:43:39 but it wouldn’t be that they had a reputation
0:43:41 to preserve elsewhere.
0:43:44 So that was a kind of complicated compromise decision.
0:43:46 ‘Cause it’s not, it turns out it’s a hard problem.
0:43:49 You know, how do you get people to care about their jobs?
0:43:51 Usually it’s they get paid more if they do it well
0:43:53 and they get fired if they do it badly.
0:43:55 And on an independent body,
0:43:57 you can’t really do either of those things.
0:43:58 You know, you can’t like reward them
0:44:00 for making good decisions.
0:44:01 And if you can punish them by firing them
0:44:03 for doing, for making the wrong decisions,
0:44:05 they’re not independent.
0:44:07 So then you need something somewhere in between.
0:44:11 And reputation, I think, is the most powerful motivator.
0:44:12 And there I would just say, you know,
0:44:15 you might think, oh, no one’s motivated by reputation.
0:44:16 But there’s, well, it turns out
0:44:18 there’s a whole bunch of people in the world
0:44:19 who take the jobs that they,
0:44:20 a lot of them are academics,
0:44:22 who take the jobs that they have
0:44:25 rather than some job that would pay a lot better
0:44:26 ’cause they care about the thing they’re doing.
0:44:28 They like doing it.
0:44:31 And then so you ask, like, okay, now you have tenure.
0:44:32 When I remember, you know, getting tenure and think,
0:44:34 oh, this is so great.
0:44:34 And then someone said to me,
0:44:37 this is so great, you never have to do a day’s work again.
0:44:39 And I thought to myself, like,
0:44:40 I never thought of it that way.
0:44:41 Like I’m so much of a neurotic,
0:44:42 I’m gonna keep on working hard.
0:44:44 But rationally, if I were a real rational actor,
0:44:46 I maybe I would stop.
0:44:47 And the main reason not to is,
0:44:49 then my reputation would be in tatters.
0:44:50 And people would say,
0:44:52 as they do say about some academics,
0:44:53 oh, there goes Feldman.
0:44:54 He’s one of those people who,
0:44:57 the day you got tenure, he never did anything ever again.
0:44:58 And, you know, that hurts.
0:44:59 So, and I care about it, not hurting.
0:45:01 – It’s impossible to picture, you know,
0:45:04 stopping working after tenure.
0:45:04 – Such a layabout.
0:45:08 – So, anthropic is also one of the examples
0:45:10 of these citizens’ assemblies.
0:45:11 So it might be–
0:45:12 – Yeah, describe the citizens’ assemblies anyway.
0:45:14 Let’s dive into that.
0:45:16 – I think the general idea here
0:45:20 is you have cases where you’re either worried
0:45:24 that too many people won’t participate if you hold a vote
0:45:27 or that they won’t actually know enough about the issue
0:45:30 prior to having some kind of deliberation together
0:45:35 and/or briefings on the issue to make an informed decision.
0:45:37 And so instead of holding a vote,
0:45:42 you try to randomly sample a representative group
0:45:45 of your citizens, in the case of a citizens’ assembly,
0:45:48 or users, in the case of one of these online,
0:45:49 I’ll call them user assemblies.
0:45:52 Other people call them different things.
0:45:53 And then you bring those people together.
0:45:56 The idea is because you’ve randomly sampled,
0:46:00 they’ll be representative of the user base of your platform.
0:46:03 You get them together, you have them debate the issue,
0:46:05 you have them hear briefings from experts
0:46:07 so that they’re informed about the issue,
0:46:09 and then you have them make a decision
0:46:12 through voting or some other collective process
0:46:16 on whatever the difficult issue is.
0:46:18 And this is something,
0:46:22 it’s had some history in the real world going way back.
0:46:24 And it’s closely related to this idea
0:46:27 for a major grease of sortition.
0:46:29 – What is sortition?
0:46:34 – Sortition was this idea that you would randomly choose
0:46:36 people to be in charge of various issues.
0:46:38 – It’s election by lottery.
0:46:41 – Yeah, some people call them lotocracies also.
0:46:42 – The idea is basically,
0:46:45 we need someone to have the full-time job
0:46:47 of running some part of the government.
0:46:48 We don’t wanna vote for them
0:46:50 ’cause that creates all kinds of weird incentives.
0:46:54 So every year, we have a lottery, or twice a year,
0:46:57 and we pick one or two people who are gonna have this job.
0:46:59 And so, and they only do it for one term,
0:47:02 they don’t have to worry about getting reelected,
0:47:05 but for reputational reasons, they try to do a good job
0:47:08 because they’ll look bad if they don’t do a good job
0:47:09 and people will think well of them if they do a good job.
0:47:11 So you pick them randomly,
0:47:13 and then you do it again the next year.
0:47:14 – It’s sort of like jury duty,
0:47:17 but like heightened to the absolute utmost level.
0:47:20 – Well, the Athenians had very aggressive jury duty.
0:47:22 They had much larger juries
0:47:25 and they paid them quite well, which is interesting too.
0:47:28 But I think the main justification for “Sortician”
0:47:31 was this idea, I’m gonna paraphrase this,
0:47:34 but there’s this great quote from Douglas Adams,
0:47:35 the author of “Pitchhiker’s Guide”
0:47:37 and several other great books
0:47:41 that’s basically like any person who wants to be in charge
0:47:45 is ipso facto unqualified to be in charge.
0:47:47 And I think that was the idea of “Sortician”
0:47:52 is we don’t want these untrustworthy, overly ambitious,
0:47:58 potentially corrupt or sociopathic people
0:48:02 who desire power to be the ones in charge.
0:48:07 We want the average well-meaning patriotic citizen
0:48:09 to make decisions instead.
0:48:10 And so the way we’re gonna make sure
0:48:15 we don’t allow that sociopath to insert himself into power
0:48:18 is we’re gonna randomly choose who’s in charge.
0:48:19 And I just wanna note,
0:48:24 it’s an interesting solution to that problem.
0:48:26 It also raises a bunch of its own challenges.
0:48:30 If you think that a big part of what makes democracy work
0:48:32 is what we call accountability,
0:48:35 the idea that the person whose job it is
0:48:37 to pay attention, learn the issues,
0:48:39 figure out what maps to your preferences
0:48:42 and make decisions on that basis.
0:48:44 If it’s the fact that they’re worried about
0:48:46 whether they’re gonna get to keep their job
0:48:49 is what makes them do all those things well,
0:48:52 then “Sortician” is a terrible, terrible method
0:48:54 because I have no incentive to do anything.
0:48:57 I’ve been randomly chosen to do this job
0:49:00 and now I just have to do it no matter what.
0:49:02 I don’t have those strong incentives
0:49:05 that are coupled with my desire to win re-election.
0:49:07 So there’s no accountability in some sense
0:49:09 unless, like Noah was saying,
0:49:12 considerations like reputation, patriotism,
0:49:15 genuinely just caring about the issue are dominant.
0:49:18 So those are kind of, that’s one of the key trade-offs.
0:49:19 – But if you do a bad job,
0:49:22 maybe you’ll get ostracized and exiled.
0:49:24 – Also, they used to do that.
0:49:25 They would actually do that too.
0:49:26 But there’s another key trade-off,
0:49:28 which is the background assumption of “Sortician”
0:49:30 is that you don’t need very much
0:49:32 that all the expertise you need to do the job,
0:49:34 you already have or can pick up really fast.
0:49:37 So imagine you’re running your startup
0:49:42 and you’re gonna choose someone to be your CTO
0:49:43 and you’re like, oh, I’ll do it by “Sortician.”
0:49:46 Like bad luck for you if your lots fall on me
0:49:51 because I don’t code well enough to be able to do a good job
0:49:52 of reviewing the other people’s code
0:49:55 that’s necessary in the startup for the CTO.
0:49:56 And so, and there’s no chance
0:49:58 that in the six month or one year period,
0:50:00 I could get up to speed fast enough.
0:50:01 Like maybe if you gave me a year,
0:50:03 I’d make some progress, maybe,
0:50:06 but I don’t have the relevant expertise.
0:50:09 So it’s also probably based on a great amateur,
0:50:11 in a positive sense, amateur idea of government,
0:50:13 kind of, you know, you can imagine
0:50:15 19th century British aristocrats
0:50:18 who think that like anyone in their club,
0:50:20 if called upon, could sit in government
0:50:21 and figure it all out, you know?
0:50:24 But life is very complicated today.
0:50:26 And the things that government does today
0:50:28 are infinitely more complex
0:50:32 than they were in a, you know, an antique society,
0:50:34 you know, in Athens.
0:50:36 And so even if you assume
0:50:39 that people could have done a good job then,
0:50:41 it’s not so clear they could do it now.
0:50:44 And even the Athenians didn’t do it for certain jobs.
0:50:46 Like they didn’t pick their generals by sortition.
0:50:48 They weren’t dumb enough to think
0:50:50 that they could go out and win wars
0:50:52 by picking a random person to be the general.
0:50:54 That required some expertise.
0:50:56 So that’s a second problem.
0:50:58 And then a third problem is,
0:51:01 is this job the kind of job you get better at over time?
0:51:03 ‘Cause if you get better at it over time,
0:51:04 you’re not gonna get a chance to,
0:51:05 ’cause you’re in for six months or a year,
0:51:07 and then you’re out.
0:51:09 And so you’re not, if it’s the kind of thing
0:51:11 where there’s upsides to doing it for longer,
0:51:12 you’re gonna lose the upsides.
0:51:13 Obviously there are also downsides
0:51:16 to having someone in the job for too long.
0:51:17 But, you know, this sort of like,
0:51:19 it’s sort of like turn limits on steroids.
0:51:22 – So I think where I completely agree with all that.
0:51:23 And so where this has gone,
0:51:27 so I was one of a number of people
0:51:30 who helped design this user assembly for Metta.
0:51:32 It’s called the Community Forum.
0:51:34 And based on these concerns,
0:51:38 but also knowing that having straight up user voting
0:51:39 would be very challenging.
0:51:45 We really wanted to focus on really value-laden decisions
0:51:49 where it’s not really a matter so much of expertise,
0:51:53 but really it’s about trying to capture in this community,
0:51:57 what are your core values that affect this decision?
0:51:59 And that might be a place
0:52:02 where this is a more workable model,
0:52:04 but I do think it’s quite limited.
0:52:08 And going forward, I mean, my personal view,
0:52:08 I don’t know what, no, thanks.
0:52:10 My personal view is,
0:52:14 I think these things have to be expanded in some way
0:52:17 to bring in more expertise and to allow for delegation,
0:52:19 as we call it in crypto,
0:52:21 because of the things Noah was just saying.
0:52:24 I think it’s particularly hard
0:52:26 that someone serves on one of these things,
0:52:28 learns a bunch and then disappears.
0:52:31 And so developing ongoing expertise
0:52:35 and the types of decisions that are socially fraught
0:52:37 going back to the beginning of this whole conversation,
0:52:40 the kinds of decisions a platform might want to give over
0:52:43 to its users, we’re gonna want people to develop
0:52:46 ongoing expertise and a sense of accountability
0:52:47 for that, I think.
0:52:48 – And as Andy says, there’s also,
0:52:49 there couldn’t be a lot of upside.
0:52:53 So I mean, take a hard kind of values-based,
0:52:56 ethical or moral question.
0:52:59 Let’s go back to what should the nudity policy be
0:53:01 on a social media platform, right?
0:53:02 It’s a hard question.
0:53:07 And you could imagine that if you take 1,000 people
0:53:10 and they’re genuinely representative of the users,
0:53:11 which is a tricky thing as we’ve talked about,
0:53:13 but imagine that they are.
0:53:16 And maybe they all have an impulse
0:53:17 one way or another way when they start,
0:53:19 but then you give them enough information
0:53:21 for them to have some thoughtful conversations
0:53:23 and then see where they land after, say,
0:53:27 three longish conversations over three days.
0:53:30 They may all land in a place
0:53:31 that’s a lot more thoughtful.
0:53:32 In fact, they will land in a place
0:53:36 that’s a lot more thoughtful than where they started.
0:53:37 And they may have a different perspective
0:53:40 than say the people who work inside the company
0:53:43 or the outsiders who start off very committed
0:53:44 to one point of view or another,
0:53:46 either because they’re like, I don’t know,
0:53:48 they’re pro-sex and they wanna have
0:53:51 as few clothes as possible or they’re deeply religious
0:53:54 and they wanna have as many clothes as possible.
0:53:57 And so you could imagine that you get,
0:53:58 and I think it’s plausible.
0:54:00 And I think in some of the experiments
0:54:02 that Andy’s run with Metta,
0:54:05 you get a more thoughtful, nuanced, balanced answer.
0:54:08 And I think for those purposes, it’s great.
0:54:10 And it’s better than a focus group
0:54:12 ’cause there’s maybe a little less control.
0:54:13 You know, in focus grouping,
0:54:15 the problem is if you’re good at running a focus group,
0:54:17 you can make the focus groups say almost anything.
0:54:20 And here, there are more people, there’s more space,
0:54:22 there are more protocols to stop
0:54:23 the people who are presenting the question
0:54:26 from driving the answer in a particular way.
0:54:28 And it just makes sense that we would get a,
0:54:31 in general, we get a better result, not every time.
0:54:33 I mean, because there’s some scenarios
0:54:37 where deliberation produces perverse results,
0:54:38 where people get into a cascade
0:54:41 where this is like how people burn witches, you know.
0:54:43 – Group think and herd mentality.
0:54:44 – Yeah, you get into that
0:54:47 and then everyone goes towards some crazy extreme thing.
0:54:49 That can happen, but there are a lot of techniques
0:54:52 that Andy’s design builds in
0:54:55 that are designed to stop that from happening.
0:54:57 – Now, part of this community forum
0:54:58 is you’re trying to provide information
0:55:01 so people can make an educated decision about something
0:55:04 and maybe change their mind about some things.
0:55:07 They arrive at a, you know, workable solution.
0:55:11 But my question there is how do you ensure
0:55:12 that the information and education
0:55:17 that people are receiving is non-biased or objective?
0:55:20 And, you know, ’cause whoever’s presenting you
0:55:22 the menu of options, they have a great amount of control
0:55:25 on steering you toward a certain outcome.
0:55:26 – Good question for Andy.
0:55:28 And then the other tricky problem is
0:55:31 how do you make decisions about what informational decisions
0:55:33 about what’s the right information?
0:55:34 Right, imagine what we’re discussing
0:55:37 is precisely the misinformation problem.
0:55:38 What’s our baseline?
0:55:42 You know, where do we even start on what is reliable?
0:55:43 – It’s a super challenging problem.
0:55:46 I think there’s a couple different answers,
0:55:47 none of which are perfect.
0:55:50 And this is a broad problem in all democracy,
0:55:54 not just the online world and not just these assemblies,
0:55:56 though it’s perhaps sharpest in these assemblies
0:55:58 since the entity organizing the assembly
0:56:01 makes decisions about what information to present.
0:56:02 I would say a couple things.
0:56:07 One is you try not to have a monopoly on information.
0:56:10 And so, you know, a big part of these assemblies
0:56:13 is that people get to discuss and deliberate
0:56:15 and they’re completely free to bring in any information
0:56:18 they want gathered from whatever source they want.
0:56:21 So that’s, I mean, that’s the first most important thing,
0:56:21 probably.
0:56:24 And of course, you try to do as good a job as you can
0:56:27 of bringing in experts that are credible
0:56:30 to represent the different sides of the debate.
0:56:33 I think it’s become in vogue a little bit
0:56:37 on a certain part of the Twitter community
0:56:40 or ex-community, I guess it would be these days,
0:56:43 to say that there’s lots of fraught debates in society
0:56:47 where it’s obvious that one side is factually correct.
0:56:49 And so we shouldn’t show both sides
0:56:53 that both ciderism is itself a problem.
0:56:56 Of course, there are instances and places
0:56:58 where that might well be true.
0:56:59 But I don’t think you can get people
0:57:02 to make an informed decision about any issue
0:57:05 if you don’t let them hear both sides
0:57:07 or more than two sides of the issue,
0:57:10 even if you think the preponderance of evidence
0:57:12 rests on one side than the other.
0:57:14 You really, I think, can’t get to an informed decision
0:57:16 unless you’ve heard all the arguments.
0:57:19 And so I think that’s a big effort in this
0:57:24 is get a diverse set of experts to present all sides.
0:57:30 But long run, I think you want to be in a position,
0:57:32 if you really can build out a democratic system
0:57:34 that goes beyond these assemblies,
0:57:39 you’d want to foster a competitive ecosystem
0:57:42 in which different information providers crop up
0:57:44 and provide their different analyses and views.
0:57:46 And that’s something I’ve talked quite a bit about
0:57:49 with DAOs ’cause I think that’s an issue in crypto too.
0:57:51 – On that point, there’s also a body of literature
0:57:53 about people in brainstorming settings
0:57:56 and when there are bad ideas proposed,
0:57:58 it actually enables the group to arrive
0:58:00 at better decisions over time.
0:58:02 – Yeah, I mean, that idea, which I mean,
0:58:06 the classic modern formulation of that is John Stuart Mill.
0:58:08 His argument for free speech was,
0:58:10 we actually need the bad arguments
0:58:13 because working through why they’re bad and wrong
0:58:16 and false will help us get the right arguments.
0:58:18 And that’s, it sounds kind of like a trite argument,
0:58:21 but if you really think about it, it’s pretty deep
0:58:25 ’cause it’s not obvious and it’s interesting in that way.
0:58:28 I think another really interesting kind of aspect
0:58:33 of this whole challenge to me is jumping out
0:58:36 to the bigger political world of democracy.
0:58:39 So you were asking before Robert about,
0:58:40 it’s a funny time to be talking
0:58:41 about the US Supreme Court as a model.
0:58:44 And I think the same in a way is kind of true
0:58:47 about these ideas of democratizing,
0:58:50 but one of the aspirations of the community assemblies
0:58:52 or forums that citizens assemblies
0:58:54 or community forums that Andy’s talking about
0:58:57 is to get real disagreement that is thoughtful
0:59:01 that doesn’t turn into polarized yelling at each other.
0:59:05 And one of the ways to do that is by narrowing down the issue.
0:59:08 So let me point to one weird feature of our polarization.
0:59:11 One weird feature of polarization is that people turn out
0:59:12 to have really strong views about things,
0:59:14 they don’t know anything about,
0:59:17 and they have them instantaneously.
0:59:18 So you’re like, how?
0:59:19 Like, how do you, this happens to me sometimes,
0:59:21 I’m like talking to somebody and they’re like,
0:59:22 oh, I’m sure I think X.
0:59:23 And I’m like, I’ve known you for 30 years,
0:59:24 you don’t know anything about that.
0:59:26 Why do you have this strong position?
0:59:29 And the answer is as a time saving device
0:59:32 in a world where there’s so many issues
0:59:33 and they’re so complicated.
0:59:36 Once we’ve picked a team,
0:59:38 we’re blue or we’re red or we’re libertarian
0:59:41 or we’re anarchist or whatever we happen to be.
0:59:43 Once we’ve picked a team,
0:59:45 there’s now a list of positions
0:59:47 that’s associated with that team.
0:59:52 And so we just default as it’s defensible time saving heuristic
0:59:54 to siding with our team on those issues.
0:59:56 ‘Cause we kind of think that someone else
0:59:57 has thought about it or a whole bunch of people
0:59:58 have thought about it.
1:00:01 And probably this is where I’m gonna end up
1:00:03 because I generally agree with this group of people
1:00:04 on various things.
1:00:08 And one of the things about doing a citizen’s assembly
1:00:10 or a community discussion about a narrower topic
1:00:15 is you can sometimes avoid this habit that we have
1:00:18 of defaulting to a polarized position
1:00:19 as a time saving device
1:00:21 because you’re gonna be told about it.
1:00:23 Like you’re gonna get the information in front of you.
1:00:25 And so it’s not only that we’re closed-minded
1:00:28 and that we don’t want to listen to the other side,
1:00:29 although that does happen.
1:00:30 We’re also partly closed-minded
1:00:33 because we’re barraged with so much information
1:00:35 that we don’t have time
1:00:37 to consider every perspective on every issue.
1:00:40 We just don’t have the cognitive bandwidth to do it.
1:00:42 I actually think that’s one of the problems
1:00:43 with our current polarization.
1:00:45 It’s not the only cause of it.
1:00:47 There have been polarizations in many societies
1:00:50 that do not have anywhere near as much information
1:00:50 as we’re getting.
1:00:52 So I’m not making some like,
1:00:54 this is the main causal factor,
1:00:56 but it is one reason that we do better
1:00:59 in a well-designed citizen’s assembly
1:01:02 often than we do out in the wild with politics.
1:01:04 – I think that’s been one of the most interesting
1:01:06 to me surprising learnings from them,
1:01:09 both in the physical world and online,
1:01:11 is they’re generally organized around
1:01:15 the most fraught kind of culture wars type issues.
1:01:17 And when they’ve been run,
1:01:20 one of the big learnings I think has been that,
1:01:22 if you get a relatively small group of people together
1:01:25 and you structure the conversation well,
1:01:28 people generally are pretty reasonable.
1:01:32 It turns out they don’t want to be so partisan
1:01:36 or so hostile when they’re put into the right environment.
1:01:39 And that connects to other things
1:01:40 about our polarization, right?
1:01:42 Which is that it’s much larger than it seems
1:01:44 because we consume so much of it
1:01:47 through these online platforms
1:01:49 where we hear the loudest voices,
1:01:52 but we don’t see the in-person interactions
1:01:54 that turn out to be usually less polarized.
1:01:56 – I had a funny example of that for my own life.
1:02:00 I was moderating a panel at Harvard
1:02:03 and some students decided they were gonna leaflet against me
1:02:05 and they were handing out leaflets to people as they came in.
1:02:08 And the leaflet was like, it was like a four-page leaflet,
1:02:14 but its text was entirely a download from a Twitter,
1:02:17 like a Twitter thread that someone had created
1:02:19 attacking something that I had written.
1:02:23 And I was like, this is so, it was just so weird
1:02:26 because it didn’t translate well to the format of the,
1:02:29 you know, to the format of the leaflet.
1:02:31 But also I just thought to myself
1:02:33 the extremity of the formulations
1:02:37 that this person was using, you know, against me online,
1:02:39 you know, they didn’t really translate to the forum
1:02:41 where like we were having a thoughtful conversation.
1:02:43 And I sort of felt bad for the people who were leafleting
1:02:46 because it’s one thing to, you know, call what I said,
1:02:47 I think VAPID horse shit,
1:02:49 which is what they were describing me as saying,
1:02:52 like they seemed in context weird and rude to be saying that
1:02:53 ’cause everyone in the room
1:02:55 was having a rational reasonable conversation,
1:02:57 whereas on Twitter, I’m sure it seemed great.
1:02:59 And then the punchline of it was,
1:03:01 there was like a substantive issue where I had,
1:03:03 in the thing I’d written that they made them angry,
1:03:05 I’d made an argument, possibly wrong,
1:03:07 but I made an argument and I gave evidence for it
1:03:08 and I supported it.
1:03:10 And the response to this argument was,
1:03:12 it’s obvious that this is wrong,
1:03:14 so we’re not even going to entertain it.
1:03:17 And I was like, that just sounded so absurd
1:03:18 in the context of a university.
1:03:20 On Twitter, it sounds great, right?
1:03:22 Like Feldman is full of VAPID horse shit,
1:03:24 so why should we even bother to refute him?
1:03:27 All you need to know is that we think he’s stupid.
1:03:29 And then that’s enough to reach the conclusion.
1:03:31 But in personal conversation,
1:03:34 very hard to carry that off without looking silly, I think.
1:03:36 – Since you brought up Harvard,
1:03:38 are you feeling this intense tension
1:03:41 between the administration, the student body?
1:03:44 It just seems like such a powder keg over there.
1:03:47 – It was a very, very, very intense fall.
1:03:49 And I mean, very intense.
1:03:52 As intense as I have experienced on the university campus
1:03:54 and I’ve been on and off of this campus since 1988.
1:03:56 So it was, it was extreme.
1:03:59 Now things are sort of slowly getting back
1:04:01 to maybe normal is too strong a word,
1:04:03 but towards a calmer way of being.
1:04:05 And I will say during all of this,
1:04:11 our classes were normal, like school went on.
1:04:13 People learn stuff, they studied for exams.
1:04:16 We did have a normal campus life,
1:04:19 but all of the intensity that was being felt was,
1:04:20 it was felt.
1:04:23 So it was an experience and not for the most part,
1:04:25 I think it’s fair to say for most people,
1:04:27 they don’t think it was a good experience.
1:04:29 – Have things subsided a little bit?
1:04:33 – Yeah, I think so, you know, I mean, there’s now,
1:04:35 so what are universities bad at and what are they good at?
1:04:38 Universities are really bad at making fast decisions
1:04:39 about anything.
1:04:40 They don’t like to do it.
1:04:42 And if they do it, they make bad decisions.
1:04:44 ‘Cause if your personality were that you could really
1:04:46 react in real time, you know, you’d be like a trader.
1:04:48 You’d have, you’d have some job
1:04:50 where you could really take advantage of that.
1:04:53 And of course, there are some people in university
1:04:56 who are really intellectually quick,
1:04:59 but the best of them use that intellectual quickness
1:05:02 to fill in their debt.
1:05:04 And then you make the decision over time.
1:05:06 So universities are bad at the fast stuff
1:05:08 and the fall was all fast stuff,
1:05:10 responding to a news cycle,
1:05:13 making issuing statements and declarations.
1:05:16 Now we’re entering a phase where certainly here,
1:05:19 you know, the university has two task forces already set up
1:05:20 to study these issues.
1:05:23 There’s two more at least coming.
1:05:24 And this is going to be the part of the universities
1:05:26 do pretty well, like take a deep breath,
1:05:28 figure out what we’ve done wrong,
1:05:29 figure out what we can do better,
1:05:32 give reason explanations of how we should do better
1:05:33 in the future.
1:05:36 And so our kind of, the nervous system of the university
1:05:40 will return closer to what it’s best at.
1:05:41 You know, and again, I want to be clear.
1:05:42 There are some people whose, you know,
1:05:44 nervous system is always like, go, go, go.
1:05:45 And that’s really valuable
1:05:47 in a whole bunch of dimensions of life,
1:05:50 but the university is not really one of them.
1:05:54 And so we do our best when we read a little bit.
1:05:58 And that’s now what’s happening, and that’s better.
1:06:00 So in that sense, we’re heading directionally
1:06:03 in the right way, though by no means are we there yet.
1:06:06 – I think it’s a really good and important example
1:06:08 of just the general topic we’ve been discussing,
1:06:11 ’cause a lot of it has to do with governance
1:06:15 and the governance of these super important very,
1:06:18 in some cases, very old, long running institutions
1:06:22 that are not straightforward businesses
1:06:25 and therefore face quite complex governance issues.
1:06:27 And certainly some of the ones, you know,
1:06:31 I’ve observed over the last, let’s say 10 years,
1:06:33 that make it so challenging is,
1:06:37 you have the exact same uneven participation problem.
1:06:40 So not every faculty member leans in and participates
1:06:42 in the governance of their university evenly.
1:06:45 The ones who choose to do that may have different views
1:06:47 than the ones who don’t.
1:06:48 And then on top of that, you have,
1:06:50 I don’t necessarily mean this in a negative way,
1:06:53 but like the mission creep of the university,
1:06:57 which is a lot less focused on protecting academic freedom
1:07:00 and developing the very best research in the world
1:07:03 and a lot more focused on a lot of other issues
1:07:06 that extend way beyond research.
1:07:09 And I think it’s hard, it’s hard to figure out
1:07:11 how to fix those governance problems
1:07:13 because the institutions are so big,
1:07:17 entrenched in so many different areas
1:07:18 and trying to spread their attention
1:07:20 across so many different things.
1:07:21 And I worry–
1:07:22 – One thing to do is to step back from,
1:07:26 at least in my view, to step back from thinking that
1:07:28 by making declarations,
1:07:31 the institutional part of the university contributes
1:07:33 to our understanding of the truth and of knowledge.
1:07:34 I don’t think that’s true.
1:07:37 So if the deans get together and hold a meeting
1:07:40 and announce that the second law of thermodynamics is true,
1:07:42 I don’t think that gives us much special knowledge
1:07:44 into whether the second law of thermodynamics
1:07:45 is in fact true.
1:07:49 However, when the people who are at the cutting edge
1:07:51 of some area of inquiry publish their research
1:07:53 in a peer review journal,
1:07:54 there’s reason to take that seriously.
1:07:55 It’s not always correct.
1:07:57 It’s always subject to revision,
1:08:01 but I would stop and listen closely to what they had to say
1:08:05 and give it some epistemological benefit of the doubt
1:08:07 because they are expert
1:08:09 and they’re speaking on their area of expertise
1:08:10 in a thoughtful way.
1:08:12 And I think one of the things that’s happened
1:08:15 is that our universities, and not only our universities,
1:08:16 but it’s happened in universities
1:08:18 have kind of fallen into this part of the mission creep
1:08:19 that Andy’s describing,
1:08:22 of thinking that they have to express a public statement
1:08:25 on every matter of public importance.
1:08:28 And I understand the moral impulse to say the right thing,
1:08:33 but that has to be balanced against, are you good at that?
1:08:35 And are you contributing to the university’s mission,
1:08:39 which is to pursue the truth in the broadest sense,
1:08:41 you know, through reading, writing and teaching.
1:08:43 But that’s different from, you know,
1:08:47 being in the announcement and declaration game.
1:08:48 And I don’t think universities are very good at that.
1:08:50 And I think stepping back from that is,
1:08:53 it’s not by no means an overall solution,
1:08:54 but it’s the first step.
1:08:57 – This is happening all across corporate America and beyond,
1:08:58 you know, these calls for activism
1:09:02 and for institutions to take a position on various matters.
1:09:03 – I have clients coming to me,
1:09:05 corporate clients every day saying,
1:09:07 “We’re in so much trouble about this.”
1:09:10 You know, and I do try to tell them,
1:09:12 there’s no such thing as being neutral.
1:09:13 That’s their first point.
1:09:16 Genuine neutrality is not possible in the world.
1:09:19 But within the framework of realizing
1:09:20 that you can’t be neutral,
1:09:23 you can sometimes step back and say,
1:09:26 “Look, you know, we’re not gonna take a view
1:09:28 on this or that important thing.”
1:09:30 And the companies find themselves in these positions
1:09:32 because they’re lobbied.
1:09:34 And they’re lobbied by people
1:09:37 who are trying to affect consumer behavior often.
1:09:39 And sometimes they’re able to pull that off.
1:09:41 I mean, Met has been subject to a boycott.
1:09:43 You know, I mean, other big companies
1:09:46 are also subject to various boycotts.
1:09:48 So you are living in a real world.
1:09:49 If you have customers,
1:09:51 you have to worry about your customers.
1:09:52 But you also have to be aware
1:09:55 that your customers could be all over the place.
1:09:59 And that’s one of the reasons to have step back policies,
1:10:01 including policies of referring something
1:10:03 to some other group of people
1:10:05 and saying, “This is a really hard one.
1:10:07 We’re not qualified to decide this.
1:10:08 And we’re handing it off to somebody
1:10:10 who is qualified to weigh in on it.”
1:10:11 – I agree.
1:10:14 And I think the key value is the tying your hands.
1:10:16 You really, you have to make what we would call
1:10:19 in the social sciences a credible commitment.
1:10:20 You have to get to the point
1:10:23 where the people who would try to force you
1:10:24 to take a position on this issue
1:10:27 that’s irrelevant to your core mission,
1:10:29 yet important to them,
1:10:31 believe that there’s nothing they can do
1:10:33 to make you take the position.
1:10:34 And I think the critical mistake
1:10:38 that a lot of the most important universities made,
1:10:40 as well as a lot of corporations made,
1:10:42 was to give in to those demands
1:10:44 and then create common knowledge
1:10:47 that they can be forced to make those statements
1:10:49 and that they aren’t committed
1:10:50 to not making those statements.
1:10:52 And people often point to it,
1:10:54 but one of the few universities
1:10:56 that had less of a problem with this stuff
1:10:58 has been the University of Chicago
1:11:00 because they had made a preexisting written commitment
1:11:02 to not take such positions.
1:11:06 And that has turned out to be a huge luxury for them
1:11:08 that other universities haven’t afforded themselves
1:11:10 because they hadn’t made that commitment
1:11:12 or the commitment wasn’t credible.
1:11:14 And I think that’s, I think, you know,
1:11:17 the central problem is that over a 10 plus year period,
1:11:20 a lot of the top universities demonstrated
1:11:23 that they did not have a credible commitment
1:11:24 to not taking those statements.
1:11:26 And now they’re trying to walk that back,
1:11:28 but it’s hard to create that commitment out of nothing.
1:11:30 – I’d be remiss not to bring the conversation
1:11:32 back to some of the subjects we were discussing
1:11:36 at the very outset around internet governance.
1:11:40 And you’re mentioning this concept of credible commitments
1:11:42 in the context of the University of Chicago,
1:11:45 but when it comes to making commitments
1:11:49 with internet services and binding people to rules,
1:11:52 I wonder what the path forward is there.
1:11:54 When you look at something like the Meta Oversight Board,
1:11:56 they have made some decisions
1:12:00 that Meta doesn’t necessarily have to adhere to.
1:12:03 Now, it adds a lot of transparency to the process,
1:12:05 but I wonder if there are other systems
1:12:10 that would, you know, bind services and corporations
1:12:13 and applications to the way that they,
1:12:19 to some set of, you know, agreed upon circumstances.
1:12:23 Of course, the thing that comes to mind is DAOs,
1:12:25 distributed to autonomous organizations
1:12:29 and how you can sort of create these commitments
1:12:33 in blockchain code that bind all participants.
1:12:35 – I’ll just say, I mean, my view on this,
1:12:36 I think there’s two things that are interesting
1:12:40 about blockchain and DAOs with respect to this topic.
1:12:42 The first is this trust problem
1:12:46 and being able to write and commit to a process.
1:12:48 So one of the first things, you know, projects do
1:12:50 when they start in crypto is essentially
1:12:55 write a constitution, i.e. write down and code
1:12:58 how they’re gonna make decisions over different issues.
1:13:02 And, you know, exactly how binding those are
1:13:04 is up for debate, it’s a little bit complicated,
1:13:06 but in the long run with blockchain,
1:13:09 I think it really is true that the constitution
1:13:12 you commit to through that process in code
1:13:15 is a lot harder to change without resort
1:13:17 to some kind of democratic process
1:13:19 than if it’s not committed in that way,
1:13:22 especially if it goes way beyond what can be protected
1:13:25 through normal real world legal processes.
1:13:28 And so I think there is something really interesting
1:13:30 about being able to build an online community
1:13:33 where you can make a sort of promise into the future
1:13:36 that anytime this type of decision comes up,
1:13:39 this is the process by which we’re gonna make that decision.
1:13:42 I think there’s something quite interesting about that.
1:13:44 The second is more economic in nature.
1:13:46 It has to do with, you know,
1:13:48 one of the most important types of trust
1:13:50 for online platforms is the trust
1:13:53 between the different sides of the platform’s market.
1:13:54 And so, you know,
1:13:57 almost all online platforms have this characteristic
1:14:00 that they’re trying to bring together two sides of a market,
1:14:02 drivers and riders or developers and users
1:14:05 for the app store and so forth.
1:14:07 And you really need the producer side
1:14:12 of that two-sided platform to believe that into the future
1:14:15 they’re gonna have an economically beneficial relationship
1:14:16 with the platform.
1:14:17 And I think one of the biggest challenges
1:14:19 we’re seeing in the space right now
1:14:22 looking like the huge dispute between Epic Games and Apple,
1:14:26 for example, is that developers are starting to feel like
1:14:29 it’s still a tremendous opportunity to develop
1:14:33 on top of these extraordinarily good global platforms.
1:14:35 But at the same time, the taxes they pay
1:14:37 to the platforms are going up,
1:14:40 the decisions made around their services
1:14:42 are changing in unpredictable ways.
1:14:44 They would like to have a much longer term promise
1:14:47 about how the platform’s gonna treat them.
1:14:50 And that’s another place where I think being able to make
1:14:53 a long-term in some sense immutable promise
1:14:55 over the economic relationship between a platform
1:14:59 that it’s producers, I think could be important.
1:15:02 – In a broad sense, that’s what a constitution is.
1:15:05 I mean, a constitution, the best metaphor for it
1:15:09 is Odysseus and the Odyssey when he’s on his boat
1:15:11 and he’s going to the island of the Sirens.
1:15:14 And he knows they’re gonna be so appealing to him,
1:15:16 so beautiful, their song so beautiful
1:15:18 that he’s gonna jump ship.
1:15:21 And so he ties himself to the mast
1:15:24 and tells his crew members like, “Don’t let me go.”
1:15:26 And that’s in one of the leading metaphors,
1:15:27 what a constitution is.
1:15:30 I mean, the idea is sort of like,
1:15:32 we know that when the ships are down
1:15:35 and we’re in a panic, we’re gonna take away people’s rights.
1:15:37 We’re gonna silence them or we’re gonna take away their,
1:15:40 their right against arbitrary arrest.
1:15:43 And so we try to bind ourselves from doing that.
1:15:45 And we try to create institutional mechanisms
1:15:47 to bind ourselves.
1:15:50 Notably in government, there are two parts to that.
1:15:52 There’s the written rules.
1:15:55 And then there’s the human ability to interpret
1:15:57 and maybe even override those rules.
1:15:59 And there’s a productive tension between those things.
1:16:00 And it goes all the way back to a debate,
1:16:02 believe it or not, I had to be so ancient Greek,
1:16:04 but we’ve done a lot of ancient Greeks today,
1:16:06 a debate between Plato and Aristotle
1:16:07 about which is better.
1:16:10 Whether it’s better to have the rules in the end in charge
1:16:12 ’cause no one’s perfect, which is the Aristotle view,
1:16:14 or whether it’s better to have the wisest person
1:16:16 you can get your hands on in charge.
1:16:18 – Because the laws are king.
1:16:20 – Yeah, ’cause the rules won’t give you
1:16:21 the best outcome all the time,
1:16:23 which is broadly the Plato view.
1:16:25 And they’re both right.
1:16:28 You need some back and forth or productive tension
1:16:29 between those things.
1:16:34 And that’s true in Dao’s, where you don’t want the thing
1:16:36 to lead to a total spiral down.
1:16:38 You need to have some break glass measure.
1:16:44 And it’s also gonna be true in something
1:16:48 like the oversight board, where the company sometimes
1:16:51 has committed itself absolutely to following their rulings.
1:16:54 Other cases, it’s asked for an advisory opinion.
1:16:58 But in the end, if Meta chooses to stop listening
1:17:02 to the advisory to the oversight board altogether, it could.
1:17:03 They would just have to pay a reputational cost
1:17:05 for doing it.
1:17:09 So, I mean, I think, this really is in the realm of art
1:17:10 rather than science.
1:17:13 You always need to have some of each.
1:17:15 You need to have some meaningful constraint,
1:17:19 and then you have to have some capacity for flexibility.
1:17:22 And so it’d be nice to say there’s like a magic solution
1:17:23 to this, but there isn’t.
1:17:25 And there’s no purely technological solution,
1:17:27 but there’s no solution that has no technology,
1:17:30 because rules and the following rules are a technology.
1:17:32 So you need both.
1:17:37 And that may not be the most thrilling conclusion.
1:17:41 Rules always win, or people always win.
1:17:44 Actually, sometimes some wins and some of the other wins.
1:17:46 And that’s real life.
1:17:48 So maybe that’s not a terrible place
1:17:51 to end what I have to say on the topic.
1:17:53 – I want to build on that in one particular way,
1:17:55 which is like one of the things we study a lot
1:18:00 in the history of democracy is this idea of constitutions,
1:18:02 and what we call in political science
1:18:06 and political economy, self-enforcing constitutions.
1:18:09 There’s no external authority
1:18:12 that can bind a country to its constitution.
1:18:15 And so the constitution only has power in the long run
1:18:18 to force people to follow the rules.
1:18:21 If there’s some track record of everyone
1:18:23 having agreed to it for long enough,
1:18:27 that it has some special power that it gets,
1:18:29 that it accrues over time,
1:18:30 because there’s nothing to stop someone
1:18:32 from tearing it up and ignoring it.
1:18:34 And so that’s why we call them self-enforcing.
1:18:36 They only really bind it to the long run
1:18:39 when everyone has a long enough track record
1:18:41 of agreeing to be bound by it.
1:18:42 – This goes back to the Magna Carta
1:18:45 where King John was forced to sign it, right?
1:18:46 And then he was sort of like,
1:18:48 well, actually I’m not gonna pay attention to it that much.
1:18:51 And then there was a battle
1:18:54 over the legitimacy of the crown and all that.
1:18:56 – Tons of history of this.
1:18:59 And the US constitution is somewhat unusual
1:19:01 in the length of time for which it’s proven
1:19:04 to be somewhat self-enforcing.
1:19:07 That same problem exists online.
1:19:09 And I think Noah is right.
1:19:12 He was hinting at this, I think,
1:19:15 that blockchain or ways
1:19:18 of writing immutable agreements into code
1:19:21 do not, in a comprehensive way,
1:19:23 solve this problem of self-enforcement
1:19:25 because you still always,
1:19:27 people have the option to fork and to leave
1:19:29 and to do whatever they want,
1:19:32 or to change the code as long as they agree to it.
1:19:35 And so there’s still something deep in,
1:19:37 like he was saying, art rather than science about it.
1:19:41 However, I do think for the online world,
1:19:44 blockchain provides something pretty fascinating
1:19:47 that I was really blown away by as a political scientist
1:19:48 when I discovered it.
1:19:51 And I think it’s best highlighted through an example.
1:19:54 So there’s a Tao called Lido.
1:19:58 And Lido has been discussing publicly for a while,
1:20:02 its desire to build in a veto
1:20:04 for certain types of decisions.
1:20:06 And what they wanna do is essentially update
1:20:08 their constitution to say,
1:20:10 when there’s this one set of decisions,
1:20:12 maybe the Tao just makes the decision,
1:20:15 but when it’s this other type of decision,
1:20:17 it goes to an external veto.
1:20:21 And when I was talking to them and to other people
1:20:24 about how that would work in practice,
1:20:29 one of my biggest concerns was that in the real world,
1:20:32 if you try to set some procedural rule
1:20:34 that only applies to one set of legislative votes
1:20:36 and not another,
1:20:38 then it’s just obvious that strategic actors
1:20:41 will channel it into whichever category
1:20:42 is favorable to them.
1:20:44 So if they’ll just claim like,
1:20:45 if I don’t want the veto to apply this,
1:20:47 they’ll just claim this is one of the votes
1:20:48 that the veto doesn’t count on.
1:20:51 And because legislatures get to make all their own rules,
1:20:53 you really can’t stop that.
1:20:54 And we’ve seen that in the Senate.
1:20:56 Occasionally the Senate parliamentarian
1:20:58 tries to stop the Senate from doing stuff.
1:21:00 There’s a weird equilibrium where sometimes the Senate
1:21:02 defers to the parliamentarian,
1:21:03 but in the long run,
1:21:06 the Senate has no real deep obligation to do that
1:21:09 and can always change its rules if it wants to,
1:21:10 if enough people want to.
1:21:12 And so I kind of thought,
1:21:13 this veto is not gonna work
1:21:16 because you can just redefine Lido’s votes
1:21:18 into one category or the other.
1:21:20 But I think the thing that turns out that I learned
1:21:22 that’s really interesting is no,
1:21:26 because the topic of the vote is defined
1:21:29 by the smart contracts that the vote actually touches
1:21:31 or doesn’t touch,
1:21:34 you can define in a very deep and immutable way
1:21:36 whether this is one of the votes
1:21:38 that the veto’s gonna apply to or not.
1:21:40 And that allows for a form of commitment
1:21:43 that I’ve never seen before in a legislature.
1:21:46 I don’t think it’s a panacea.
1:21:48 I think there are still these broader issues
1:21:50 of self enforcement that matter
1:21:51 for all the reasons that Noah’s saying,
1:21:54 but I do think it offers something pretty fascinating
1:21:57 in the vein of binding commitments
1:21:59 to legislative procedures.
1:22:01 – Isn’t this the miracle of the US Constitution, Noah?
1:22:04 I mean, you’ve written a biography of James Madison,
1:22:06 the fact that you can set these rules up front
1:22:09 and yet make them flexible enough that they can endure.
1:22:11 – Well, if we had Madison here,
1:22:15 he would agree that it was his perfect design.
1:22:20 You have to ignore the civil war and some other blips.
1:22:22 But yeah, I mean, the aspiration was to create something
1:22:25 that would be able to be both rule-based
1:22:28 and also responsive to change over time.
1:22:30 Some of the design elements didn’t work well.
1:22:34 I mean, the amendment provision, we use it very, very rarely.
1:22:36 He probably thought that was generally okay,
1:22:37 but there have been some circumstances
1:22:40 where we really need reforms and we can’t really get them.
1:22:44 More fundamentally, it’s a system that’s designed
1:22:48 to enable compromise and to enable compromise in the middle
1:22:51 and to push politicians back towards the center.
1:22:53 And that’s why, right now, if you ask
1:22:56 about sort of the big question, the grand questions,
1:22:59 plaguing politics and political science,
1:23:03 I think it’s fair to say that if you have some confidence
1:23:06 that the laws of political science are still true,
1:23:09 then you think that we will migrate back towards the middle
1:23:11 because our system is designed
1:23:13 to push us back towards the middle.
1:23:16 And if you, on the other hand, are really panicked
1:23:18 about where things are going and the possibility
1:23:22 for breakdown, you think that it might take two,
1:23:23 you might still believe in the rules,
1:23:25 but you think they might take too long to operate.
1:23:28 And then in the short term, you could get a breakdown
1:23:31 in overall faith in the system’s capacity.
1:23:33 And the other part of it is looking at the world around us
1:23:35 and seeing that for all of our polarization,
1:23:38 lots of elements of our society are still functioning
1:23:41 and functioning really reasonably well.
1:23:45 – Amazing, this has been a fantastic conversation.
1:23:47 Thank you both so much for all your time.
1:23:49 – Thank you so much, really fun.
1:23:50 – Thank you, yeah, this is great.
1:23:55 – We sort of took it by fiat
1:23:57 that direct democracy has problems.
1:23:59 We mentioned Bode McBoatface,
1:24:01 which is a pretty humorous example,
1:24:05 but are there more serious examples from history,
1:24:08 from the physical world of direct democracy in action
1:24:09 and going awry?
1:24:12 – I mean, the most classic example people point to
1:24:17 is at which the America’s founders pointed to was Athens.
1:24:21 So Athens had a pretty robust direct democracy.
1:24:24 The reality of how it function was quite complicated.
1:24:26 It wasn’t as simple as just like everyone
1:24:28 showed up and made every decision.
1:24:30 They’re actually like pretty complex layers
1:24:32 of decision makers and stuff,
1:24:35 but there was a significant component of direct democracy
1:24:38 among land-owning men.
1:24:44 And one way that people say it went awry
1:24:48 is that during the war with Sparta,
1:24:52 you know, this is just one telling,
1:24:56 but like the people got swept away with passion
1:24:59 and were uninformed about the strategic decisions
1:25:04 that had to be made and essentially forced Athens
1:25:10 to open a second front in the war by invading Sicily.
1:25:14 And that sort of turned into Athens, Vietnam.
1:25:16 It was like a disaster.
1:25:19 And it massively sapped Athens power.
1:25:22 It was super, super costly.
1:25:27 And that is often held up as an example of like mob rule.
1:25:30 This is because the decision to open a second front
1:25:32 and invade Sicily was,
1:25:35 now you said that there are complex layers
1:25:37 of decision-making, but in this respect,
1:25:41 it was kind of a bottoms-up mob-like decision.
1:25:44 – The claim is that the mob was manipulated
1:25:49 by ego-driven generals who wanted to burnish
1:25:54 their reputations by opening a new invasion
1:25:57 and that they manipulated the mob into supporting them.
1:26:00 And if you look in the classical era,
1:26:02 and you could question the motives of all these authors,
1:26:06 that’s sort of the stereotype that arises.
1:26:09 And so post Athens, glory days,
1:26:13 a lot of Roman writing refers to mob rule
1:26:17 and the ability of demagogues to manipulate the mob
1:26:21 as reasons to be very skeptical of direct democracy.
1:26:25 There’s like a really famous passage in Virgil’s Aeneid
1:26:27 where he goes on at length
1:26:32 about how cynical leaders can whip up the mob
1:26:36 and what we really need are seasoned statesmen
1:26:40 who will make decisions carefully on behalf of people.
1:26:43 More modern examples, hard to say.
1:26:48 It’s so unfunctional that it’s really not even tried.
1:26:51 Yeah, it’s really, I mean, you can point to like,
1:26:54 in Switzerland, there’s a pretty aggressive
1:26:56 local referendum system.
1:26:58 I think there’s general agreement
1:26:59 that it’s kind of crazy,
1:27:02 but like it doesn’t work as badly as some other thing.
1:27:04 I mean, you literally at the Canton level in Switzerland,
1:27:08 you vote on who receives passports.
1:27:10 It’s really quite wild.
1:27:12 Yeah, there’s some great research on this.
1:27:15 And you know, in the US people point to California
1:27:19 is pretty extreme on the end towards direct democracy.
1:27:22 And that can go both ways.
1:27:24 In some ways, I think it’s an important
1:27:26 and potentially valuable institution
1:27:30 ’cause it allows voters to surface issues
1:27:32 that the legislature might want to ignore.
1:27:36 And so if you think your legislators are doing a good job
1:27:38 or are captured by special interests
1:27:40 or for whatever reason or not,
1:27:43 sufficiently accountable to voters,
1:27:45 then giving voters this alternative mechanism
1:27:50 to force issues could be quite valuable.
1:27:54 The downside is that you end up with tremendous voter fatigue
1:27:57 ’cause you have lots of these votes
1:28:00 and the California ballot is ridiculous.
1:28:03 It’s many, many pages long.
1:28:04 Most people, including myself,
1:28:07 can’t understand most of the issues being voted on.
1:28:09 And then the second problem,
1:28:10 and this is often a problem in direct democracy,
1:28:13 is interest group capture of the agenda.
1:28:17 And so how ballot initiatives get onto the ballot
1:28:20 is complicated, but it’s actually not that hard
1:28:22 for a well-resourced committed interest group
1:28:27 to force carve outs for themselves onto the ballot.
1:28:30 And so one thing that happens in California is every cycle,
1:28:35 we have to vote on this extremely abstruse ballot initiative
1:28:39 that has to do with whether doctors should be required
1:28:41 to sit in on all dialysis.
1:28:45 And you might think it’s doctors who are pushing this
1:28:47 because you might think like, oh, they make money from it.
1:28:50 But no, the doctors do not want to do this.
1:28:51 And they think it’s crazy
1:28:53 ’cause there’s no medical reason why a doctor needs
1:28:55 to be present for the administration of dialysis.
1:28:59 – Yeah, it sounds like just a ton of time wasted and…
1:29:01 – Yeah, yeah, it’s crazy.
1:29:05 And it’s being pushed by certain interest groups
1:29:08 that are kind of just more generally
1:29:09 in a conflict with doctors.
1:29:11 And they have stated publicly
1:29:14 that the reason they put on the ballot every two years
1:29:17 is to force the doctors and interest groups to spend money
1:29:19 convincing everyone not to vote for it.
1:29:24 So it’s a complete waste of everyone’s time and resources.
1:29:25 And so that I think is a great example
1:29:27 of how these processes get messed up.
1:29:29 – By the way, the Irish referendum that just happened,
1:29:31 my brother-in-law was just visiting from Ireland.
1:29:34 So I was interested in this going on.
1:29:40 It was interesting to see people reject so strongly
1:29:42 these proposed changes to the constitution.
1:29:44 I don’t know if you’ve been tracking it.
1:29:46 – No, I haven’t followed that very carefully,
1:29:49 but I do think that’s a good example
1:29:51 where referendums, if they’re embedded
1:29:53 into a broader process in a healthy way,
1:29:58 can be a really good way to get more signal from voters.
1:30:01 And this happened in California, I think it was 2018.
1:30:05 We had a bunch of pretty like culture war related
1:30:08 referendums or ballot initiatives.
1:30:09 – Yeah.
1:30:10 – And quite consistently,
1:30:12 the voters really signaled through them
1:30:15 that they were in a more centrist position
1:30:17 than California’s elected officials.
1:30:20 And I think that had a big impact
1:30:22 on how the elected officials then proceeded
1:30:26 from then on after they lost these big ones.
1:30:28 So I do think it can be valuable,
1:30:31 but it’s complicated how you enact in practice.
1:30:34 And I think that’s kind of where Dow governance is heading.
1:30:38 There will still be some direct token holder voting,
1:30:41 but there will also be a lot more delegation
1:30:46 to professional experts on these issues.
1:30:47 – You need to protect against mob rule,
1:30:49 but also plutocracy.
1:30:50 – Exactly.
1:30:55 – And find some sort of middle way.
1:30:58 – Yeah, and I think that the delegation stuff,
1:30:59 it’s very valuable,
1:31:02 but it’s not gonna solve the participation problem
1:31:03 fundamentally.
1:31:05 You’re still gonna need the token holders
1:31:08 or the other voters to pay attention
1:31:11 and make sure their delegates don’t go rogue.
1:31:13 And that’s gonna require some pretty careful planning
1:31:16 because there’s definitely a temptation to set it
1:31:19 and forget it in terms of delegating your tokens.
1:31:23 But we have lots of reasons to suspect
1:31:27 that if that behavior manifests regularly,
1:31:29 the delegates won’t have those incentives
1:31:31 we want them to have to do a good job.
1:31:33 So I think a lot of the most interesting work right now
1:31:35 with Dow Governance is around,
1:31:38 how do you build delegation programs
1:31:41 that first of all recruit and give good incentives
1:31:44 to delegates while at the same time, second of all,
1:31:46 still encourage token holders or other voters
1:31:50 to pay attention and to think about re-delegating
1:31:51 their votes on a regular basis.
1:31:55 So that delegates feel like they’re being watched.
1:31:56 – What are people putting in place?
1:32:00 What sorts of new rules or experiments are happening?
1:32:01 – First thing that’s happened,
1:32:02 which I think is really interesting
1:32:05 is a bunch of the largest Dow’s have instituted,
1:32:07 they each have a different name,
1:32:09 I would call them Delegate Programs,
1:32:12 which is a combination of often it includes paying
1:32:14 the delegates, sometimes as a function
1:32:16 of how many votes they accrue,
1:32:20 and creates like a online web interface
1:32:22 that makes it easy for token holders
1:32:24 to delegate to different delegates.
1:32:26 And so you’re putting together,
1:32:30 you’re creating incentives for there to be delegates
1:32:34 and you’re helping token holders find delegates.
1:32:36 And we’ve actually done some research,
1:32:37 there’s some pretty interesting evidence
1:32:39 that rolling out those programs
1:32:44 actually does increase token holder voting participation,
1:32:47 probably because you’re asking the token holder
1:32:49 to do a much easier task.
1:32:52 Instead of voting on like literally changes
1:32:55 to the underlying protocols code,
1:32:56 you’re just asking them like,
1:32:58 find one of these delegates who you like
1:33:00 and give them your voting power.
1:33:03 And part of these programs is also having the delegates
1:33:07 write platform statements and or post videos
1:33:09 about what they wanna accomplish as a delegate
1:33:12 that again helps the token holders find like,
1:33:14 oh, that’s a delegate that shares my views,
1:33:16 I’m gonna delegate to them for now.
1:33:17 Fascinating.
1:33:19 So this is a paper that you’re working on currently.
1:33:20 Yeah, yeah, I’m hoping to have it done
1:33:23 in the next month or so, we’ll see.
1:33:24 Very cool.
1:33:25 And so basically getting people to delegate
1:33:29 their votes more, making it low friction,
1:33:31 making the information available
1:33:32 for people to make reasonable decisions
1:33:34 about who they want to delegate to
1:33:39 and the directions that they’ll take a Dow.
1:33:40 Exactly, yeah.
1:33:42 No, and I think it’s gonna be a really important model
1:33:47 because this came up in the conversation with Noah,
1:33:50 I think a lot of people in other parts
1:33:54 of online governance are coming to the same conclusion
1:33:55 that we need representatives
1:33:58 ’cause we need this kind of expertise in the accountability
1:34:00 that we can’t rely on direct democracy.
1:34:04 But Web3 is years ahead, literally years ahead
1:34:07 in experimenting with how you actually set up
1:34:09 representative democracy.
1:34:12 So I think it’s gonna be, yeah, really important development.
1:34:16 I mean, we’ve got this intense like Darwinian combat going on,
1:34:18 all of these experiments just let loose
1:34:20 and seeing which ones will flourish.
1:34:25 How far away are we from actually determining what works,
1:34:29 what doesn’t work and seeing the kind of best practices
1:34:30 shake out from all this?
1:34:31 I don’t know.
1:34:33 I think there’s probably two aspects to it
1:34:36 that we need to have happen before we’ll know for sure.
1:34:42 One is to get Dow’s to a place where they have
1:34:47 more broader killer use cases for society,
1:34:49 which will in turn make the governance decisions
1:34:51 higher stakes for society.
1:34:54 And then we’ll see which governance structures
1:34:56 can stand up to that pressure.
1:34:59 In some ways they’re already involved
1:35:01 in the high stakes decisions,
1:35:04 certainly in these like DeFi protocols and stuff,
1:35:09 but they don’t have the same kind of global public pressure
1:35:12 on them that other online platforms have faced
1:35:15 because other online platforms are much more mature,
1:35:18 have many more regular users and so forth.
1:35:21 So I think that’s gonna be a really big interesting shift
1:35:25 as it occurs that’s gonna bring the Dow governance
1:35:27 more together with what we talked about with Noah
1:35:30 in terms of like web 2.0 governance.
1:35:33 And then the second is we need over time
1:35:36 to have more Dow’s with broader distributions
1:35:40 of voting power, which is happening over time.
1:35:43 Big criticism of Dow’s historically has been,
1:35:45 they’re clothed in the rhetoric of democracy,
1:35:48 but the voting power is very unevenly distributed.
1:35:52 And that may make some of this like delegation experiment
1:35:55 sort of unrepresentative of what would happen
1:35:58 in a broader democratic system
1:36:00 where there’s more conflict among users
1:36:02 with roughly equal amounts of power.
1:36:06 And so that is the trend, I think,
1:36:08 in token holding over time.
1:36:10 And again, as these become larger,
1:36:13 as they touch on more killer use cases for society,
1:36:15 I think we’ll see that broader distribution
1:36:17 and that will be another kind of pressure test
1:36:19 for these systems.
1:36:22 So those are the two things I’m keeping my eye on.
1:36:24 (upbeat music)
1:36:28 (upbeat music)
1:36:30 (upbeat music)
1:36:33 (upbeat music)
1:36:36 (upbeat music)
1:36:38 (upbeat music)
1:36:41 (upbeat music)
1:36:51 [BLANK_AUDIO]