#43 - Jamie Joyce - Democracy, Digitization and Debates

37:44I feel like a database just does better when it comes to representation. And then there's deliberation. So the Society of Libraries ontology is a deliberative ontology, right? They're debate maps. There's many ways. Ontology is just like, it has many meanings, but in like, you know, data science, it's like the shape of the data itself, like the rules that you've created about the relationship between pieces of information. And the ontology that we chose was deliberative, debate-like. It's in part because we're interested in answering questions and modeling the different points of view. So we're going back to the classical definition of debate as in trying to pursue truth through argumentation, right?

38:19And I'm going to build an argument. You support it, refute it, whatever. it's a way of going back and forth trying to figure out what's so um but it's also what congress is supposed to do the representatives come containing their you know constituent information they go to a place and they're supposed to debate about it but debate is a skill arguing in a logically valid way is a skill and these human representatives have limited bandwidth they have limited time attention ability to speak like relying on their like in real time performance is absurd they may have limited subject matter familiarity, cognitive capacity, patience.

38:53And then that's not even talking about like the corruption incentives, like all that other stuff. It's a little bit darker. So yeah, I just feel as though there's certain functions in the legislative branch, which could be replaced with technology. And it may just so happen that what the Society Library is working on is kind of that. How, in order for that such a thing to happen, you would have to assure the politicians that they would still have a role 100 and i i mean i'm sure that they would because there's a gazillion things they you know as yeah that's the thing they're just they're not they're not able to do do these things so that this is fulfilling a job that you know there's still plenty of other like things to go do like kiss babies and and actually listen to meet people yeah exactly and again exactly it's just it's giving them a source of info you know a much more robust body of information from with from which to actually then make better decisions for their constituents totally um like now you mention it like it seems crazy to me that they wouldn't have some kind of database like how can you possibly i mean it means everyone's just doing it i guess that what are they just like licking their finger and holding it in the air well there's like polling and stuff like that like there are tools um there are companies that provide tools but i've never seen anything that goes that many steps far right it's just like you know okay this is what your constituents think here's like a graph showing a sentiment analysis of what they believe but like Like the responsibility then still is on the legislator to make sense of it, to like actually piece it together.

40:19And it's an incredibly like mentally and cognitively laborious task to make sure that you're not only hearing from your constituents, but speaking to all the right experts. Right. And they'll bring in people for congressional hearings and, you know, they'll, you know, get documentation from the Library of Congress, CRS reports like they're just intaking so much information. So it would be really helpful to have technological augmentation, which cleans, collates, labels, links, organizes, pre-organizes as a deliberation for them. So that's the thing that I'm really excited about. And in no way I'm saying they should be replaced in this entirety, for sure.

40:53I think that humans still serve many useful roles. But that is the part of government that I am most excited about because it seems the most outdated. The thing I would love to see is some kind of digitization and just more transparency in the lobbying process. Yeah. Because technically, lobbying is meant to be relatively transparent for everyone to see, like, where has money been flown? You know, where is money coming from? Who is it going to? Whose interest is it likely serving? And I mean, there were there was the like Lobbying Disclosure Act in 1995. uh the honest leadership and open government act of 2007 but from what i can tell these have basically made no difference um there's only like loopholes a lot like a lobbyist only counts as someone who uh spends at least 20 of their time lobbying on behalf of an employer which means that most deliberately like keep themselves under that threshold um and then like organizations frame their activities under like, you know, oh, it's for public education or issue advocacy rather than direct lobbying, et cetera, et cetera.

41:59So there's just like all of these different loopholes. Do you think there's some way of using your kind of model to, like, how would you go about like categorizing and recording these flows of information when it comes to lobbying? It's not one of the problems that we're looking to solve specifically, but, you know, in a hypothetical sense, let's say if the society library was our deliberation graphs were a tool that were included in the legislative deliberative process and relied upon, it would kind of limit the need for lobbying in a way. Now, I don't think there's any technology that's really going to overcome human nature unless humans are entirely constrained by the technology.

42:42So a lot of this is just kind of wishful thinking. Like I think we would need a really noble champion in government and a lot of like coordinating actors to change the system enough to kind of create a system that actually allows for technological augmentation in a way that actually changes the incentive structures and the gain that's being played. But like hypothetically speaking, you would not need if you were a nonprofit organization and you want to send a lobbyist off to Washington to have face to face with a representative who's going to be casting a critical vote on a piece of legislation that you think would harm, you know, the work that you're doing instead of you needing to send off a lobbyist, a human being to have, you know, use their connections in order to get that FaceTime, you could just submit your information, your argumentation directly to the system.

43:29And instead, the system's not like overwhelmed by like the number of people inputting to it. It doesn't matter how many people say a thing. Any claim is just one node in the database. Now, in that node package, which we talked about being rather complex, we have the various phrases in which we, you know, if we find a claim over and over and over again, they all get packed in the same node. We have a feature called variant phrasing. So we just keep track of all the different ways in which the same claim is uttered and it just gets packaged up into that point. So, you know, cybersecurity and like DDoS and stuff like that aside, it doesn't matter how many people say a thing.

44:03One person could technically make a difference in argumentation in that deliberative process in a system like this. But sadly, I don't think we can solve the problem in actuality of preventing lobbyists from having an influence. And I have an anecdote to explain why in a way. So obviously information by itself, in and of itself, doesn't really do much. You have to take actions on that information. So we, a couple of years ago, a city came to us and they had a mediator because the city council, which was like all white individuals and this group of black activists were head to head and there was a lot of high tension and emotions about the issue because it was arguing about the institutionalized racism of the way in which the city goes about electing its city council at large or single member single member means you break up the city and elect one person per district and then at large means the top five or however many most popular candidates get on the city council and so it was so tense it was just after the george floyd you know protests and that whole time and the mediator reached out to us with the city and said, listen, this isn't just a matter of people being civil with each other and engaging in nonviolent communication.

45:12It's also a matter of like fact and like argumentation. Can you help us make a less biased decision? And so us just gathering all the knowledge, right? We interviewed the activists, gathered statements from the mayor, the newspapers dove into academia, you know, and we were steel manning. Like one of the activists made a point and we found a dissenting opinion from Ruth Bader Ginsburg that supported that point. So we're linking all this stuff together and creating a robust set. But then we had to move it into action. So we invented a new type of decision-making model. It's a micro voting decision-making model where people actually have to go on record as to what argument in what dimension of the decision they were stating that they agreed with.

45:49And at the end, it aggregates up to like what their vote is. So instead of it being a binary yay, nay, where the logical calculation is happening in the black box of their minds and not able to be audited or scrutinized by anyone else, we made all of that explicit. So it was like, here's your one binary decision. Do you pass this ballot motion or not? But here's the like 55 different steel manned arguments with links to evidence and like the 24, however, different dimensions you have to consider the economic and the human rights and all these different things. And then we asked them vote on each set of arguments underneath each dimension.

46:23And then all that aggregated up, which means they would have to go on record with what they believed. And the city council members who participated loved it because it was like, oh, my god this is so complex you made it so simple the activists were super happy they're like we feel so heard but then what was really interesting is that some of the city council members refused to participate because then they would be on record with what they actually believed so i think there's no it's like the it's a constant problem in civic tech we can build amazing technology it's the adoption that's the problem and it's it's because people don't want to they don't actually want the transparency they don't want to be put on the record right what they say well yeah that's and that's that's the thing with all of these you know even with the lobbying thing um most people i would expect think that there is a big problem with corporate lobbying affecting you know it's creating regulatory capture etc i mean my issue very dear to my heart um fact you know how to improve the conditions of farmed animals it's just like it's all because of the big farm the big agriculture lobby just pushing through the most heinous corporate cronyism basically yeah um through all these lobbying practices and on the flip side though there is this tension that like if everything is tracked every every every flow of penny you know penny is it it will create the ability and especially nowadays we're in this sort of era of like mob justice and like ideological attacks against someone just because they did actually fund something that you didn't like but maybe you know there's a sometimes people do fun stuff that you don't like but they have good reasons for it that you don't understand right and it's like how do we navigate that tension because i think i would still rather be in the world where we have more transparency you know and give up some of those it's essentially privacy right and like freedom you know it's it's freedom of expression i suppose is what would get curtailed but then yeah i don't know like if everyone could just be very reasonable then it wouldn't be a problem but because a lot of very unreasonable people out there it's just like there's like bad actors on both sides that make it really hard to like find this yeah you got any thoughts um this is like the ultimate win-win question by the way it's uh it's really hard to manage big complex groups of people i would say in general um i do feel as though like maybe one foundational thing i i for me a lot of it boils down to access and comprehension of information right i do believe that which means that there needs to be some transparency and access to that information so you know some of the problems with mob justice is if everything is transparent in terms of what government is doing except for like you know national security issues or something like that um and maybe there's a watchdog to make sure all that's valid yada yada um people are only getting a fraction of the information which explains what's happening and we already see that different media outlets are taking something that happened and filtering it down into a headline or a little quip and then that's what someone can has the time to ingest understand and walk away with and so that's such a crux problem i think in human civilization is our relationship to information and how poor and distorted and twisted and careless um you know the information that we're receiving can be managed and um where do you coming back to this topic of information hazards um have you where do you draw the line between informational completeness you know in one of your libraries uh and essentially withholding certain concepts in the interest of perhaps public safety has that come up in any way have you ever had one where you're like ah that is definitely an info hazard or shouldn't be in there or is it not necessarily like the types of topics you're covering it's just it's not relevant um there are relevant things i don't think we've come across a serious issue like that but has like stopped us in our tracks or anything.

50:15If you go to our website, we have a whole page about wicked knowledge management issues and some having to do with like policies about what words should we use. What if someone's trying to like insert dog whistles into the database? Like what do we do about that? And what do we do about hate speech? What do we do about privacy and all these things? So we have some rules about them. And it's been pretty easy in part because we do a lot of detachment from identity. So I would say, you know, one information hazard would be doxing the people where we get information from because we're not just collecting it from media sources.

50:44We're also collecting it from people. So we are mapping like, you know, AI safety and AI accelerationist arguments. I would like go into these communities, and I would go to their conferences, and I would stand up before a talk and say, does anyone oppose me like recording this? We will transcribe it, delete the audio recording of this discussion, and like give no attribution. And that's just how it will appear in the system, right? It'll be like society library collected. You know, we're not saying who. And some people agree, some people don't agree. Yeah, so it's kind of easy in that way. We've been able to find policies to skirt around these issues.

51:15And a lot of it has to do with just the way that we're treating the information. We are enriching it in a way because it doesn't really matter who says it. It matters like what kind of evidence and argumentation support it. So by creating that space and like cleaning it up and refining it, we're removing a lot of the harms. So no, no information hazards about like the necessity of outlining, you know, how to build a bomb or anything like that. Yeah, right. I mean, that's the thing. Some of them are like just pretty obvious. But again, it's more these like edge cases. And obviously part of the sort of issue with the woke era was that it would like conflate harmful with offensive.

51:52Yes. And people be like, I am offended and therefore what you said was harmful. Right. And it's like, no, that's not how this works. But it's difficult because the term harm is ultimately a subjective thing. Yes. And in a way, information can harm you, right? like i'm i'm i'm not one to like uh i won't go down that road i'll just say um it's an interesting experiment because i think that information doesn't need to harm you it is just information but like practically speaking information can harm so you could walk up to someone and say you know hello and brightens their day or you could say oh my goodness you know someone you love just died and suddenly their whole physiological state changes it captures their mind like if they happen to trust and believe that person it can be completely untrue but like the way in which our minds interact with information and reality um can be harmful right i don't think we should like pander and accommodate that like our personal orientation towards it is like just get used to ingesting more information and improve your relationship with information so you interact with it more like an object you know if someone gives you something you're like this is spiky i'm gonna put it down i don't want to interact with that or like oh i like this i'm gonna save it for later and maybe keep it for a while because that's my point of view now so i'd really love to eventually get to that point and oh yeah i mean it sounds like what really what you're advocating for is a separation from information and identity yes definitely it's one of our virtues and values yeah okay tell me about your virtues and values because you've got a really nice in fact i'll list them off quickly and we can maybe dig into a couple yeah because you've got them listed on your page which is always a very good sign of anything when they when you people like list out these are our values that we are trying to adhere to yeah um generally a good sign uh so yeah intellectual honesty intellectual independence intellectual inclusion earnest service love of information and individuals yeah so going back to the point we were just discussing i think one of the like subtitles um for one of them is like if you have an opinion it has you right and so we work to like let me think about that if you have an opinion it has you yeah i see because it's like you are captured by it you're captured by it you've made it into your identity etc yeah yeah um yeah so everyone can go and read our virtues and values but what we are hoping is like to bring these virtues and values to new audiences so i really feel as though something missing from our education is again improved relationship to information and what i mean by that is like you could imagine kids learning to interact with information this way a set of balls that they could play with right so it's hot this one's cold i like this color i don't like this color but literally learning that they can just pick it up and have it or pass it on to someone else they can put it down when they're done with it it's just these are just things that exist we can generate them internally or we collect them from the outside and if we if we have that kind of like physicality around ideas then maybe we're going to be less like pearl clutching and offended and angry and it seems to be like so good for our own well-being as well as good for us as sovereign like individuals who want to work and collaborate together in order to manage civilization in our own lives well yeah it's supposed to be constantly reacting i mean they often get a lot of flack but that's what i love about like the rationalist community um and all the sort of ones adjacent to it is that you'll often hear just like the craziest conversations because people are just comfortable playing they realize it's like oh we're in a lab of ideas yeah and someone can throw out some thing that is a little left field and people will tell them no that's probably ridiculous for this reason this release and this reason but they they it's like a i hate the term but it's a safe space to actually mention these things um you know a lot of like moral philosophers often like that's a literal job is to explore the like boundaries of morality for example i was having a conversation with a friend who um you know a moral philosopher who he really liked uh he was talking about peter singer and then he's like but then i read this thing he wrote about like disabled children and i realized that he's just like really fucked up in the head i was like well let's read it and it's like oh well he's he's talking about this weird edge case yeah like and he's just like talking through it and it's like i get that it's like kind of a an upsetting because he'd like this person had um a friend who had a disabled child etc and it's like it was like this taboo topic but like their literal job is to talk about like what are the edges of taboo yeah and why are they um and any community that sort of cultivates that out.

56:19that environment essentially where you can explore these edge cases that's where progress comes from Once upon a time, it was considered like taboo to say that keeping slaves was bad. Right. You're in the South. Yeah, exactly. You would have been either like laughed at or even like looked at with like, oh, how dare you? You're upsetting the order of things. Right. People will die because of you. Right. Like if you upset our economy, what will happen? Yep. You know, and yet that's where moral progress comes from. So again, yeah, it's such an important thing to get people to stop seeing ideas as these like things that can attack.

56:54And even though it's true that they, we can't get over the person, like personalizing it, right? Like we can look at an outlet and see a face. Like it's just like human nature to just like personalize things and like project onto them. So it's, it's hard, but. Okay. So I mean, you're someone who's obviously very good at doing this. Were you born this way or is it something you've like cultivated over the years? Cultivated over the years for sure. Yeah. What are some like methods that were helpful for you? Childhood trauma. Really? like most founders and entrepreneurs yeah there's usually something that happened that like completely changed them forever and uh it's funny because i'm not wearing my glasses today i've told this story before on stage um i normally wear glasses but uh how i got glasses was really what began this domino effect of like like just becoming more and more aware of our relationship to information just to be able to viscerally see it and see how people engage and interact with it um yeah so I can tell that story if you want please okay so I was a very small child and on tv at the time different cartoons and things a lot of the smart characters wore glasses and I really loved that I was like oh I want to grow up to be a smart character but because of these stereotypes I thought I needed glasses and so one day I'm walking home from somewhere and I see broken glasses on the side of the road and I put them on my face I'm very excited.

58:16And my mother says, take those off. You're going to ruin your eyes and you'll really need glasses. And so I was like, okay, so I need to get glasses to be smart. But in order to get glasses, I need ruined eyes. So I need to ruin my eyes and then I'll get glasses and I'll be smart. So it was, you know, unsound logical reasoning. Right. One of the premises, not quite true. So I walked outside and I stared into the sun until I blinded myself. Yeah. It's a true story. What? I mean, as far as we know, right, it's family memory, right? Right. How old were you? I don't know. I'll have to ask like my father, but just it's the family story.

58:49My brother swears he saw me do it, you know, but then I get glasses and I don't like immediately become like a super genius like I thought would happen. And it was like this horrific epiphany and like dawning on me of like, oh, information's powerful. And if you interact with it and you get it wrong, you can make a decision that impacts the trajectory of your entire life. Like I'm, you know I didn't know about LASIK back then so I was like I'm blind now like I need glasses to see and it didn't make any difference to my intelligence you know and it just like shocked me and so then I became like weirdly hyper aware of information and I would like notice like on the playground like someone would say something and then someone would go tell someone that same thing and then later and observe that person who heard it like restate that opinion as if it was their own and I'm like that was given to you you didn't know like that didn't come from you and so just kind of seeing the memetic pathways of information flowing and getting distorted just became like a weird obsession yeah are your eyes okay now in terms of like how how did it i mean can you presumably you can hear i mean you're looking at me you're not yeah i'm wearing contacts okay yeah both my eyes are otherwise extremely healthy they are dead the same in terms of how damaged they are and they degenerate at the same rate so which is also unusual it's unusual for my eyes to be as healthy as they are um you know not like bent and distorted or anything and then just to have the same prescription in both eyes degenerating slowly well I mean you're in the era of exponential medical right technology yeah I mean they look amazing I do love your glasses um wow what a story yeah and it's it's it's like weirdly metaphorical too because you know light is often associated with knowledge seeing the light you know so it's definitely teed me up to be a character who's like obsessed with knowledge and the pursuit of truth in like the most earnest sincere way it's like a superhero origin story yeah i've heard that before honestly it's you are misinformation

1:00:53oh that's good oh how have i not thought of that before so i'm i'm i'm i've got a like a beauty pageant sash that says misalignment that i made that i'm gonna make a video about so i should make you one for misinformation and we can be like a you know we can do like a fucked up beauty pageant thing super down gotta figure out like the logic of it of like are we pro you know is misalignment like whitewashing alignment you know because she's like a beauty pageant saying i'm aligned but right right yeah i don't know there's something in there anyway we'll unpack that later um so like one of the biggest sort of issues on my mind at present is how it feels like our information ecosystem which has always been under the influence of whack incentives of course um but it's you know like the tragedy of the commons that's going on in terms of like this almost like people polluting it with low quality information um and just generally like playing the attention game you know is this kind of race to the bottom scenario going on it feels like at least and i don't know i would love to just like hear your thoughts of someone who just like lives and breathes information and like passing it is in a high quality way as possible like what what possibilities are there that you've seen that are trying to like stem that so you know turn the cycle positive yeah well i do think we need to invent new platforms i i do think that there are so many distortions happening on top of platforms, right?

1:02:26Individuals are distorting information. Individuals are also adding a lot of context, an important context, veritable context to information. So we definitely want human input and interpretation being added into the economy of ideas, marketplace of ideas, but there's so many distortions. And so I do think we need to invent new platforms. And one of the Society Library's goals is to help provide these data sets that can be used by different platforms so that people have an easier time finding these primary sources if they exist. Yeah, I would say for individuals, a good habit. You know, you can use your heuristics and defer to the interpretation of trusted friends, especially if they happen to be subject matter experts, yada, yada.

1:03:06But also if there is not a person that you trust to be reliable, always try to go to the primary source. I love to watch a bunch of interpretations i'll go to youtube and like type in like whatever press briefing and just watch all the talking heads like distort and interpret and layer on top their own lens on everything yes and then i'll go and watch the primary source and like luckily i'm trained to kind of be able to parse the difference and keep track of the multiple interpretations some people may not be able to do that they just continuously update their priors like they perceive something and then that changes when they get new information but if you're trained to like track many threads that's a very um but yeah just it's a good rule of thumb just to try to look for the primary source whenever possible because there's usually a lot of distortions happening sometimes it's really additive context and information but it it has an impact on your perception how would you design take the x platform um yeah for me once upon a time that was one of the best sources of sense making yeah that existed um that said i was probably not privy to the level of censorship that was actually effectively going on on that.

1:04:14Like COVID was like the eye opening moment when I saw like, wait, there are all these, you know, all the lab leak discussions seem like they were just down weighted and shadow banned, et cetera. So in many ways, like it's new ownership is, it's been great. And then it's like lifted that. But then there are other things, there are other downsides that have come. Like for example, you can't really check the sources very easily because you can't have any external links. So presumably that's pretty deadly to the tool in terms of it's, sense-making quality right what would you do if you could wave a magic wand to fix that so it's funny because i actually put together a proposal and sent this to people at x and xai you know when elon bought it i think he referred to twitter now x as like the town hall right and so i came up with the proposal to actually create a feature to make it more like a town hall um in a more like formal sense so like you know on the side panel of x they have all these features and now they have Grok summarizing debates where you can see like here's a summary paragraph of like something going on and then it lists the different posts that people are making about that issue so you can see the individual posts so we put together a proposal of like combining a few features that already exist with our technology and saying when there's a topical issue you can use our ai which predicts like one of the ways in which we're working on automation is we use ai to predict the possible permutations of thought um based on like training data of having mapped like the differing points of view on complex issues.

1:05:36So it predicts what they are and then generates a series of arguments to support for and against. And then we actually query an index that's also AI like constructed by like, you know, a sorcerer code that goes and collects and deconstructs and builds an index. And then we like prune it backwards. And so we propose that like, why don't you use our AI to predict the possible spectrum of beliefs about this issue? And then what you can do is, you know, link the tweets of what people, you know, the positions that they're taking um so you can kind of like draw them in but what they're seeing is like not just a list of all the tweets that other people are making in their own interpretations but here's the list of like possible points of view and if someone's arguing to support that it's like in a nested hierarchy it's in the comments or the threat like the threaded comments of a post um you know we can i can give you the actual drawings of it if you want to put it on this video yeah please yeah so we like sent it in and we were just like yeah you could it's like cleaning it up and organizing it you know people are saying a lot of things but you could organize it these people are presenting arguments to support this position that position and then of course yeah adding links and i also think adding the quotes so chain of provenance is really important and you can kind of what does that mean chain of provenance means that you are tracking where something is coming from where knowledge is coming from so what is the url of the the whole entire source what is the paragraph or the embedding that you are commenting on or that you're extracting from so in our nodes we have the references.

1:07:00We usually create a backup in the internet archive. But then you can unpack the node and see the exact paragraph that is either exactly where we derived a specific claim from, or is an example of an expression of someone making that claim. So there can be a bunch of them. And maybe we derive from one of them, but there's other quotes. So you can just flick through and see the exact thing that we pulled it from, which would be really nice to have chain of provenance on x it's like someone is making a comment about a piece of legislation or something that someone said it's like you have that exact media clip because we also have media you can actually like open up and see this is the exact part of the video that we're referring to um and i don't see why like the comment little node in an on the x platform couldn't be a more dense metadata package because don't sub stack do that i'm not sure i'm not a sub stack person uh well just because you can like directly quote because obviously that stuff is more long form you can And like that, I'm going to quote that exact thing and it will link to it.

1:07:58So I feel like they kind of do. Is it just incredibly hard to incorporate this on the back end? Like, is that the reason? Is it very expensive? They didn't give me a reason why not. Everyone I sent it to at X and AI, they were pretty stoked about it. But I could be a number of reasons why. It could be difficult to implement. It could be that they weren't aware of who they would need to ask in order to implement it. maybe just not on their list of priorities right now because they have other things that are more demanding of their attention and time but i hope it gets to that point i hope we can actually have a formal town hall and instead of just a bunch of people like screaming at each other screaming at each other we can say okay there's an issue here are the many takes on that issue do you want to argue for and against any one of these positions let's link to evidence and like it's a collective intelligence system at that point you could have like the current thing tab yeah yes yeah so we imagine that like the this is the current thing that is trending i mean because they have trends right but i never look at them and like that kind of the hashtag thing is no one does that anymore right right but so you can go to trends but then like have that mapped out yeah society library style would be so sick and you can put it into the existing threads right like the the existing way that x is like you can kind of do that we we did our own experiment where we tried to do that as it is like there will actually have to be modifications but i mean just making it like linear and like nested in the threaded comments like you can kind of coerce it into that structure a little bit So it seems like feasible.

1:09:23Yeah. And I do think it would help with diversity of thought, freedom of thought and retrieving like quality information that that will start to change the way people interacting with with the platform instead of trying to like say something that's so based or so rhetorically catchy that you like skyrocket and popular and everyone's retweeting you. maybe people would optimize for i'm making a like a good contribution like here's i found some meaty evidence and here it is like maybe it would change the game of how people participate by trying to build like the best like series of arguments to support different positions it's i mean it's like sort of community notes feature but on steroids yes yes far more robust far more robust yeah i mean yeah it'd be incredible and that's a way of like injecting nuance back into it because that's what's now disappeared from that platform yeah it was never fantastic for it because like by definition you know by you compressing ideas down into 280 characters or whatever it's like a lossy thing and it's rewards you for but i mean it used to reward decision being concise and precise with your language and now it seems to reward just like as you say being based being pithy who can slam down who can say the who's funniest yes yeah you know it's it's it seems like it's targeted towards are like the lower emotions or the lower values or something like that and it's fine to do that sometimes but when it's like on the like a current thing topic where there's actually like meaningful sense making that you know can and should be happening you want that to be as nuanced as possible and like that would be one way of like yeah it's like and then you can have like rewards of like who has contributed the most to it right yeah even at like the end of the year or something it could be like a weird social gathering thing where just like this person like contributed like you know the most like of this type of evidence or something like that okay in the interest of steel manning because it's what it's all about what's a steel man against doing this uh it's probably difficult and maybe it is not uh rationally engaging with human nature maybe people actually do not want to do that extra work but the counter argument to that is maybe grok could do a lot of this right you know well so that leads me on to my next question which is like i feel like your whole life and work must have changed so much in the last two years yeah since we've, you know, we're now in the LLM age and these things are getting better and better and better.

1:11:37And they're so useful. Like every time now I'm doing an interview, you know, my research, how has that changed your work with Society Library? I imagine it's just like, what's your like speed up? Yeah. So it has completely changed our organization because instead of having human analysts, it's now a very small team and we're just playing with AI all day and building technology all day. Cause there's so many stages. We're not just figuring out like the mega prompt for like an LLM or anything. There's a lot of like non-AI technology that we have to build in order to execute our methods because we're looking to, you know, try and find as much information across different forms of media as possible.

1:12:14So there's different like search, custom search techniques and things like that that we're, you know, building more traditionally. But AI definitely helps a lot. And I'm like our chief prompt engineer. So it's a lot of like me ticking off everyone by never being satisfied because I'm like, you know, generating for the parts that are LLM, we like I generate prompts and then they run it a whole bunch of times and I'm like looking at the data and I'm almost never ever satisfied and like constantly finding edge cases which is like really fun um yeah my poor engineers though what impact have prediction markets made with what you do or do you think there's any way that you can incorporate them into sort of general better sense making uh yeah so prediction markets are so interesting because I feel as though a lot of people think as though they are the sense making that is like much better than debate mapping like why have debate maps not taken off why are prediction markets so popular like oh it's because they're more useful obviously no they're just more fun i think that's what it is exactly it's a game it's a game that people can win at right yes yeah so there's that and like prediction markets are meant for answering a different kind of question because prediction markets are asking the question like will this thing happen and debate maps can be used to ask those you know answer those questions but debate maps are like more useful towards answering questions like should this happen what should we right it's more of ought questions as opposed to is questions exactly right or like will be questions yeah yeah totally so i think that a lot of the it depends on the platform right but the data that may be generated uh could be useful if there's like threaded comments and people are arguing about it and upvoting and downvoting which i think like manifest does um and it's just like the prediction market like graph or whatever in and of itself is a claim right someone may argue well we should do this because you know people think it's inevitable anyway and then the graph is like a piece of evidence that people think that so um i think there's a lot of interoperability between the type of data that we're like gathering and organizing that could be useful to different systems but ultimately it's for different questions entirely yeah is versus ought is a good distinction between them do you see them as a competitor to what you do not really because it's entirely different questions i did lose an engineer to them though i'm like luckily a friend of mine um is the ceo of metaculous and he ended up uh getting one of my engineers because one of my he was a younger junior dev brilliant talented we loved having him but he was enamored with prediction markets and i to try to sit down and be like they're different questions we're trying to answer so i lost them to metaculous we miss we miss you very much um yeah they're like the the catnip to ontologists i guess right so i don't think they're a competitor um uh again interoperability between them they're useful for sure they're they're easier to grok um easier to meme and i just hope that one day like debate mappers knowledge graffers have their day in the limelight as well of like people not being intimidated by the look of a graph and seeing oh this is actually something useful maybe a heuristic that's useful to defer to to try to understand what civilization is thinking have you tried making any of your like visualizations of your graphs or maps sort of three-dimensional in a vr space uh i want to do that so badly i like see a world where i'm in there it's just some you know kind of like matrix where i'm just like pulling you can open up concepts and they just i super want to do that because you know i've been in denial about it but i am an artist and it's funny how much in my work with the society library we've just like put off art and i have to constantly explain to people like it's it's the debate maps are just a way of visualizing this data it's just structured data it can be visualized in many different ways like i explained our papers interface earlier you know we have ai agents now there's different search.

1:16:05We've had this cool like, you know, galaxy type of representation before. But that's what I really want is like, I cannot wait for the many beautiful, interactive, wonderful, adventurous ways in which this knowledge could be represented and just like kind of unleash it on artists, because like that is the art of comprehension and engagement, you know, like and so many of these platforms are like limited, linear text based. So it's like, how can we use art to make the exploration of knowledge enamoring again. And I think we are missing this in the digital age because obviously not everyone's a bibliophile like me.

1:16:39But if you go into a grand library, like if you go to the Library of Congress, you should know that if you're an American citizen, you actually are entitled to a library card and you can go into all the underground tunnels. If you just wander in there as a tourist, they'll be like, yes, look around. Okay, now leave. But if you know your rights, you can get a library card and you can go everywhere. and it is a marvelous adventure. And just seeing, like just seeing a rack of books and knowing that like that is just a brick of knowledge of a bunch of stuff maybe you've never even heard of. And there's a wall of them and there's stories of them.

1:17:11And you can go through underground tunnels and see them. Just that feeling of being in the physical manifestation and presence of the instantiation of knowledge is just awe-inducing. And digitally, I think we miss that because everything's flattened, flattened and just like put in one feed. so yeah vr and you know many different types of like 3d visualizations i think could bring that like majesty back which i think is so important because no matter what the age is humans have a relationship to information and the media format and we can have deep reverence for it and respect for it and offer it and know how to use it or we can kind of just be like bombarded by it and at the effect of it and it's just like a part of our lives and we don't even like have a metacognitive thought about it and why it's there and how it was made and what its purpose is and how we could use it yeah because the the limiting factor of sort of the current way you know even like a debate map or any of these sort of representations of the cluster of arguments and the trees and so on is is that they're very sort of like system two you know that are logical you know they follow logic yeah right literally yeah for those of you who don't know uh system two is the name for like your slow logical mind um like if you were to try and add up 41 and 72 you think it through that's your system two whereas system one is like if i ask you what's five plus five you instantly know the answer it's like your intuitions and so on and yeah this kind of like data representation is ultimately a very system two e thing um it's the kind of domain of the nerds yeah let's be honest here right and rightly or wrongly it feels like that the world is moving away from that people are going there or at least they're wanting more in people feel the need for like symbols and like quick understanding they want to grok stuff it's like everyone's there's there's a greater desire for system one in part because it's easier and people are lazy so i wonder if like art can be the bridge yes between those because sometimes we need to just feel something so we can like embody it and we can often understand stuff better without having to like bring online the slow calculator yes so i'm so glad you mentioned this because you know we talked about uh educating kids in the relationship to information earlier right and so one of the things i was having a conversation with one of my colleagues he's a i don't know if he wants to be doxxed maybe david mccraney he's great um don't worry we won't get your name and address it oh your address he's a best-selling author he's like a pop cognitive science author yeah he also has a wonderful podcast too and um he's a dear friend and we were talking about trying to make cognitive science like sexy like who and that's actually how we met is i tweeted out we need a bill nye the science guy for like cognitive science and like uh you know logical fallacies and all these things for a younger generation and immediately i think it was like the most engagement i got at the time ever you know in the all previous tweets and everyone's like david mccraney's your guy and i hit him up and we talked and we like totally hit it off and so for a while we were entertaining the idea of a kid show and it's like how do we make it like sexy engaging like like and sexy as in like poor choice of words okay a reality show a fun reality show i can i can restate it no it's fine like i'm sorry we can take it out if you want but i think this is great i like i didn't mean like that poor choice of words how do we make it like um just like yeah enlivening engaging like just like exciting um for like younger people like right you know um and um well i mean a game is one way a game is one way but we thought of like we we developed this kid show um and it's really meant for like you know pre-teens teenagers where it's like these cool sci-fi characters like from the future and like they each like represent like some some type of information dynamic so like one of the characters that I designed uh was going to run the episode teaching kids about logical fallacies and it's all about the mind like the idea we had was calling it mind your mind so like you know like how does information like get generated by your own mind how do you get how does your mind get affected by information how do you engage with information that is coming to you externally and so one of the episodes in logical fallacies there was one character can't remember her name now um but the big reveal was that she's actually like a hive mind she's like the 300 logical fallacies and she's always popping up in the show everywhere and then when it's her turn to have an episode she's like constructing the sets as you go and like it's because it's all about deconstructing an argument and showing where the errors in the construction of arguments happen so um yeah i do think art i love this i'm i love it too i'm so excited you know obviously he's had many i think he's had like another bestseller since so he's been busy but he's he's down if we can find a producer but uh we sketched out like episodes for two seasons And I feel like that's critically important.

1:22:02If we can teach kids, teenagers to just understand like the difference between information and like self and really unpack and explore that, I think we'd be a much healthier society. And so make it like cool and sci-fi and exciting and shiny could help with that. You had a meme token made about society library? Because I was going to ask you, I was like, how are you guys funded and so on? But this actually, this meme token isn't made by you. Nope. what happened there yeah so um so shaw walters from ai16z which is a parody of a16z he uh is well known in the crypto space he's an amazing 10x engineer he's a dear personal friend and uh he's he was a volunteer for the society library for many years he would hop on help us build some ai stuff hop off we've known each other for quite a while and he blew up in the crypto space with his like you know parody meme tokens and then one day he tweeted out or x posted out everyone you should donate your cryptocurrencies to the society library they're a wonderful non-profit and you know i love them you should give them you donate to them and so we've got all these like random meme tokens which was really fun and exciting and then in our wallet appeared sl token and by the time it hit our wallet we were given 50 of it it was worth half a million dollars already so someone was like selling circulating and it was a pump fund contract so like very limited there's no way you can like edit it and like add malicious code or anything it's just like you can make it in like 30 seconds so someone was like circulating it but it has our name our logo and that can't be changed right so the moment sl was created it was our name and logo but people already like bought enough of it to increase the value and 50 was sent to us and this was a very rich substantive experience in like ethics and leadership and trying to navigate something that is inherently zero sum.

1:23:59Because meme tokens are inherently zero sum. If you are winning by like extracting value, someone is losing that value. Someone's holding the bag. Someone's left holding the bag, right? And so what was wild was, you know, we acknowledge this, that we were given this gift. And then our whole world kind of blew up a little bit because it is like an international player versus player game. These factions kind of group up and they try to nudge and manipulate and coordinate together to make sure they're the ones, you know, selling at the top and not at the bottom. And like, you know, I think the SEC recently released a statement about how meme tokens are like trading cards.

1:24:37They're collectibles, right? They kind of just represent culture or ideas. SL literally has our name and logo. It was not made by us, but it represents the idea of us, a charity, you know, doing like work, like finding, you know, organizing information, fighting for truth, yada, yada. um and so how is that legal do you don't you have a trademark uh we don't have a trademark i guess not yeah you're a non-profit yeah it's i mean we could we could have a non-profit but i just uh i don't yeah i don't think we did a trademark we could i guess but yeah um even so it doesn't matter we there's nothing we can do it's forever in the blockchain now right that's it's indelible it just it just exists so at the same time it raised funds for you guys we never sold oh yeah because so this is where it gets into like the ethical unpacking right if we took it that's removing value from other people and we didn't feel comfortable knowing that you know people would have the information to con to know what game they're playing and so we were trying to figure out and it kept raising in value it became worth millions and millions of dollars our 50 stake and we were trying to figure out a win-win game out of the inherently zero-sum game And, you know, some people, obviously, we used our own system.

1:25:49We used a tool called Perspectives that we built, which, you know, predicts the various perspectives. And so we were modeling different things out. But what was struck me personally was that I actually spent the time to get to know some of the holders because people are just bombarding you. People are trying to scream at you, accuse you, beg you, manipulate you, but also cheer you on and help guide you. And, you know, all these things. It was it was absolute chaos. It was an overwhelming time. But I like spent some time getting to know some of the people. And I was like, oh, some of these folks like live in countries where I used to do humanitarian work in.

1:26:19I used to work in international sustainable development. And like, you know, that kind of work was like building schools and outfitting refugee camps and fighting hunger and stuff like that. And I'm like, a lot of these folks like live in those countries. And, you know, the thing with crypto is when there's trading happening, there's a lot of big players. They're throwing around hundreds of thousands of dollars, tens of thousands of dollars. And there's a lot of small actors. Yeah, just the little people. yeah and so we were trying to find a game that was win-win um i presented this at south by southwest i don't think that game is legally plausible but we were trying to figure out how can crypto not be zero sum yeah but unless it's like something like bitcoin where just continues to gain value you know over and over again um and so the thing that we thought of was okay if there was like a charity token model where people knew this is a charity token you buying into it means the charity is going to take out that value to build projects uh up to a certain threshold so people can keep buying in at a certain level the charity keeps taking out but then you find some crypto philanthropists who maybe would not otherwise give to the charity because they instead have an incentive and interest to see crypto succeed like a model that actually succeeds and they agree to buy at a certain threshold like okay once this thing is worth like a penny we'll buy a certain amount of it we'll spend five million dollars and so you like it aligns incentives in a way where people who want to gamble on crypto are incentivized to get in early because you know they'll they'll make the most return and they may be first in line to get bought out right like the order of operations of you jumping in and buying this means the order of operations you get bought out by these crypto philanthropists and then the you know philanthropists tend to want to invest in things that are successful now by the charity like taking out funds to do the projects it means by the time the philanthropists come around the project's already done so they can see am i going to buy this out are not and people who are buying the token are assessing will this charity succeed so just kind of aligning like you know people who want to gamble on crypto to see if they can have a return the charity wanting to do projects um and the philanthropy was going to give anyway or not give anyway wants to give but maybe wouldn't give cash because it's kind of like a little like why would a charity do this whole morass if they could just get a donation but there's plenty of people who would not otherwise give to the charity unless they are just retroactively you know compensating everyone for the project having been done and done well and i yeah i mean technically the money nets out to zero the the money itself does but yes the other values are also that's why often you know where we draw the boundary around the game is ultimately arbitrary right and like clearly the value being created there is mutually benefit like yes it's positive sum it was the only thing we we could like we could figure out like this is maybe the only way we we thought it could work and then unfortunately doesn't seem like it's legally allowed um so we're just a mean token what's the i i think that if i mean we did so much work too we consulted with a lot of lawyers because we're just like we didn't make this this isn't ours but like we don't want to get confused as being like a launcher of it um so my understanding is that like if you if anyone is engaging in trading crypto and they have an expectation of return then it's a security so like you could do it but you would fall under like security laws which is a lot of like regulation and like so it just seems like more it's less easy of a model so i guess it's possible but it's much more difficult i mean it sounds like it's something that should exist because again like if it's aligning incentives sufficiently such that these important charity projects get done yeah you know it's like there's all this energy there right let's flow it somewhere better than just kids playing these stupid games who can like out meme this one and like right the moon etc like the dgen culture which is fascinating by the way well listen i i talked i for 10 years was like how do i harness this dgen culture of the poker community who i love to bits but like it's a lot of it's you know we all call ourselves dgens and it's like how do we make this very zero-sum game of poker have positive excellentities more of a win-win yeah let's encourage people to donate a portion of their winnings to the best possible charities good job and it raised like you know millions of dollars and people loved it and there's like other like organizations doing it now um and so like it's yeah that's to me that's kind of what life is about it's like finding looking at these like classically zero sum or in many ways like probably like meme coins are a net loss to society in their current format so they're technically a lose-lose games and like how yeah like this this is one way of doing it um so i love it yes well you know in our pursuit in our quest of a win-win uh you know we just didn't touch it we still haven't we still haven't touched any sl and then the trump meme token was launched and then the melania meme token was launched and the value of sl just yeah zero sum did its thing yeah trump sucked the money out of society library out of a lot of meme tokens you know yeah yeah so well that's what you get for playing silly games everybody yeah it was a fun experiment and i just want to encourage anyone in that space to really like think about like who is likely to lose in those games you know because it's very easy to rationalize like you can imagine how easy it could have been for us to rationalize taking millions it was meant for as a donation like they did it for us like there were so many rationalizations but it's like we took the time to do some due diligence and like actually engage with people and just like this just seems totally antithetical to our being as a non-profit and so we're just gonna like i really respect that yeah yeah yeah that doesn't happen often i mean i mean frankly i don't know if i'd be able to make that not push the button and get millions like that yeah for that if it's for a cause i care about right i wouldn't do it for myself at this point but like if it was truly for a cause i like cared about i i think you do make that i mean but it is actually at the same time like it's an interesting ethical dilemma because maybe you should still take it like society library is really important i agree it's very important yes but i do at the same time if you can source the money from elsewhere that would still be better yeah i agree everyone knows what's going on and it's like yeah it's it's the eternal dilemma all right well so to finish up uh one of the things we always do with guests okay at this point is rapid fire predictions it was so funny because um i saw a podcast um where an episode of yours where you did this and i was just like oh i don't want to put numbers on things like i'll just do 50 50 every single time jamie don't you dare you can you can do however you feel okay so i have a question instead if you don't mind mixing up the format slightly go on instead of putting a number on it because it is that seems so arbitrary it kills my soul a little bit because when i when i think about like actually trying to quantify the likelihood of truth across so many permutations of like possibilities and the arguments claims and evidence just saying a number is just like it just kills me a little bit because you are you are the queen of system two ontological examination as you should be and so i'm asking you to do something as your soul yeah But I have a solution.

1:33:21Go on. What about Likert scale? What's that? Somewhat likely. Oh, likelihood scale? Yeah, just like, yeah, just not putting a, like a hardcore number on it, but just likely, not likely, very likely, you know, Likert scale. We can do that. It's funny though, because I, my first ever Ted, it was a TEDx talk was on how a number speaks a thousand words and that many, a blunder of history has been made because someone refused to put a number on a probability and instead used a word like um a fair chance or oh this will probably happen um and when people do that then there's always a interpretation but you're right that the downside the flip side is is that you often give a false impression of accuracy when you're not right so you know what i'm not going to force the rapid fire predictions upon you okay because this is ultimately you know when it would not be a win-win thing to do uh fine uh i did have some fun ones though okay let's do it let me let me just throw them at you and see if you like we can cut all of this i'm just i'm curious i'm just curious how you how you like if it resonates at all so likelihood that by the end of 2028 ai generated summaries of new bills are automatically posted on official government websites can i answer twice both ways i would say very likely

1:34:44um 55 percent interesting so 55 percent is you're very likely it's like yeah it's like basically it's like slightly more than a coin flip damn okay likelihood we get digitized voting by 2028 well in a way we already we kind of do already in a way you mean like on a platform as opposed to having to go to a physical place yeah with a machine i'd say 45 percent unlikely it's all gonna hover around the 50 range likelihood we get digitized searchable database of bills by 2028 sorry say that again the likelihood that we get a digitized searchable database of bills if i have anything to say about it 90 there we go i just contradicted myself now you're getting into the spirit now i'm just like so my you know i i tend to deliver and make things happen so uh i'm not sure if this is a good one uh likelihood that over half of u.s states adopt uniform standards for digitizing their legislative records by 2030 uh in theory in theory likely we'll say 60 percent uh in actuality 40 percent likelihood that the society library meme coin its market cap goes back above a million i would say oh sorry i need to give you an end date uh by in some point in the next five years i would say it's likely we'll say 55 okay uh likelihood that by 2050 there is a single universal debate map that covers all major human disagreements in some kind of shared database or hive mind of some kind 20 percent with a huge fucking asterisk a girl can dream a girl can dream a girl can dream because i feel like that's what you want right something directionally there like a hive imagine that just this pulsating hive mind of all information that just knows i mean it's that would be omniscience right that would be god it's it's my favorite episode of rick and morty when he dates the hive mind i'm just like oh that's the best i feel like that's your kink information yeah having it access to it yeah being a bibliophile is a very real thing like um i got a gift of a pamphlet printed in benjamin franklin's shop so it was you know printed in the 1700s um and it's all about the inquiry into spiritual truth and you know it was given to me in this box it's like hundreds and hundreds of years old this delicate pamphlet printed on benjamin franklin's printer and i couldn't stop bawling i like hit the knees just crying like there's just a class of people out there who are just lovers of knowledge and information and books and pamphlets and things it's a real it's a real love i see it i feel it like all of my knees and it's contagious yeah i get like actually it was so funny one of i have many editions of benjamin franklin's autobiography and uh the one that was the was that presented itself to be the least like adulterated like manipulated and changed edited all that kind of stuff um i found in a vancouver bookstore and by the time i found it and i opened it and i read it and it said you know like this has been unedited um i just could not stop weeping so much so that the bookstore owner came to me and he's like I'll give you a discount like you should definitely own this book like and I was just like it was many years ago and I was like oh it's like pretty expensive I was like I don't know if I can and the people in the bookstore all threw in to buy the book for me because I was like losing it and loving it so hard so it's a real thing do you have many books at home are you do you actually have a big library I do how many books roughly oh I don't know god I don't know over or under 500 surely oh I don't know about 500 because I'm a digital you know book right of course of course i would say i definitely probably have a couple hundred yeah you should see my you should come over yeah i'd love to it's very library-esque last question what's one of your favorite win-wins however you choose to define it that maybe people aren't aware of favorite win-wins it could be a thing like an event or how about an aspiration yeah yeah i feel like like oh because there's kind of like this opposition between truth and democracy my favorite way of expressing it is in the story of socrates of course right famous truth seeker so much so he was deeply annoying and then sentenced to death by athenian vote a democratic vote to kill him because he was very annoying and corrupting the youth even though he was like this radical you know truth seeker he was corrupting them how by giving them just stories or uh well it's it's you know hard to say i don't know the story oh you don't want to tell you yes please yes okay okay so soncrates uh is like the father of western philosophy uh the socratic method was named after him and it's this technique of asking people questions to cause them to like articulate their premises of belief like what they're resting their belief on and you keep going down and down and down and down until you deeply embarrass people by revealing to them that they don't know what they're talking about like like at the end of like every like like penetrating socratic inquiry is like i don't know or it just is and it's like deeply embarrassing um so allegedly like we only know like the story of socrates through plato and like uh another writer because he famously never wrote anything down but apparently he was teaching this philosophy of like you know using argumentation to investigate truth and asking questions and um yeah apparently that was corrupting the youth and uh making everyone very annoying and i did read the um socrates alleged like speech at his trial but basically if i'm recalling correctly his argumentation was like you are all biased in various ways and like there's nothing that i can say to overturn what you think in your minds because you have these pre-existing beliefs about me but like honestly you should be paying me for this good work um and there's some legends around socrates that he was like the wisest because he was the one who said like he knew nothing right socrates is the wisest man who walks the earth because he knows that he knows nothing that whole story um so yeah he went to trial and was sentenced to death by athenian vote um he had to drink hemlock poison and there's this beautiful painting of him like you know doing this gesture and like drinking the hemlock and apparently upon his death he was all the students are crying and weeping you know plato's one of his students and he was not upset and everyone's like, well, why aren't you upset?

1:41:29You're about to die. And he's like, well, I don't know that when I die, I'm actually gone. And if I'm not gone, then I will not be sad. And if I do, you know, disappear, if I'm gone when I die, then, you know what, I won't be there to be sad. Yeah. So famous characters. That story is one of my favorite articulations of like this tension between truth and democracy. A lot of people think like democracy is inherently good. And we're actually in a wave of it right now of people trying to create democracy tech. Let's have direct voting let's just hear from the people about what we should do about ai it's like well are like mass groups of people voting and being like constrained in their voting actually the best heuristic or way of figuring out what's true or optimal i don't think so i mean it's better than what came before but it doesn't mean it's the best end state right there's who knows what better evolutions there are of such a system right so instead of directly digitizing that maybe we should figure out what is like this optimal relationship so i think the ultimate win-win is when the relationship between truth and democracy really comes into harmony where people who are participating in the decision making process have like not only the cognition that's primed to interpret and process information well but also the tools and the information itself so i feel like the ultimate win-win of human civilization and humanity is when we essentially revere the sincere inquiry into truth and democratic principles at the same level.

1:42:55I think that's the sweet spot. That's like my hope. It's an absurd idealistic hope, but that is my hope. I love it. Beautiful. Thank you so much for coming on this. Thank you. Yeah, really good to have you. Same. It was a pleasure being here. Thank you so much.