🫂👩🏼‍💻🔍 215 - Social Science & Collective Intelligence with Brigham Adams of Goodly Labs

There's been a mistake among these primarily Bay Area engineers in properly identifying the problem of our contemporary times. Based on the tools they're building, it seems like they've identified the problem as individuals aren't smart enough. We don't have enough intelligent processes. And I don't really think that's the issue at all.

If I look around at the various problems that are plaguing us, particularly at the societal scale, they're not plaguing us and they don't continue to plague us because there's a lack of ideas about how to solve them or that we haven't had the intelligence to come up with the solution is. The problem is that we haven't had the collective intelligence, we haven't had the coordination to actually pull people together, legitimate the solutions among a consequential set of people who will work together to solve the problem. We're creating these tools that by themselves are these individual agents that kind of model in some ways individual human agency. The goal is to create an intelligence that can look so much like human intelligence or even better that we get confused.

And I think that's just not really what we need. And especially when you look at some of the trends of inequality and so forth, these individuals of intelligences are going to be put to work by the human individual intelligences that have the most power and wherewithal to create 10,000 lots of doing their bidding. It's going to increase the problems of inequality and lack of coordination and techno-futilism and all this sort of stuff. And what we should really be building for is increasing our collective social agency.

I would love to see us building not for a singularity but for a plurality and building tools that do not work unless multiple people are engaged in them at the same time. That's the sort of stuff that we need to be building if we want to have a society that is more equitable that understands how to move together, how to improve ourselves together. And right now it's just solving the wrong problem. The problem is not that we're too dumb, whether we don't have good ideas.

The problem is that we need to figure out how to coordinate on them. You look at every major societal challenge and it really comes down to can we coordinate that scale, not the other idea. Greetings Future Fossils. This is Michael Garfield welcoming you to episode 215 of the podcast that explores our place in time.

In case you happen to dig this episode out of a bacterial genome in the unfathomably distant future, that time is January 2024 AD in the Greek-oriented calendar. A moment at which, as one of my favorite internet memes observes, we have moved on from the 20th centuries fuck around period into a very stark and find out phase. As paths across social, economic and ecological networks continue to shrink due to the increasing connectivity of technological systems, humankind migrates from an Earth on which most events seem impossibly distant and irrelevant to an Earth defined by nonlinear, often exponential impacts of seemingly trivial developments anywhere on the planet. This is the century and the decade in which many of us have no choice but to learn the easy way or the hard way, the consequences of our increasing vulnerability to and power over one another.

And one of the places this is most vividly apparent is in how truths and untruths ripple at unprecedented speeds across the globe, forcing us into a new and intense cosmopolitanism. In the 1940s the message was loose lips sink ships. Perhaps the message for the 2020s is cognitive biases spread mind viruses, which is why I'm so glad to welcome social scientist Nicholas Brigham Adams under the show this week to talk about his work at Goodly Labs to create new infrastructure for collective intelligence, new systems for collective fact checking and sense making that can help us rise to the occasion of our inherently social planet scale challenges. If you've followed me for a while, you've likely read my 2017 science fiction short story and oral history of the end of reality, a peak into our present day post-truth carnival funhouse where AI assisted forgeries demand vastly more nuanced and sophisticated methods for navigating fundamental uncertainty, far greater humility about our validity claims and revolutionary tools for thinking together.

We have to learn to communicate the degree and dimension of our confidence and of our doubt to learn how we can rigorously restore the trust necessary for coordination at scale. And Goodly Labs is, in my opinion, one of the most promising efforts in the world right now in this regard. 2024 is very likely to feel like the end of reality for a lot of us and the stakes are immense. Fair presidential elections, concerted ecological action, and effective AI steering policy are all domains of existential risk in which we must be able to reconstruct some kind of minimally viable consensus reality.

I'd be considerably more worried for our future if I did not know that there are people like Brigham Adams and his amazing team of academics, founders, engineers, and journalists, tilting their spears directly at this issue and working around the clock to help midwife that holy grail of communications technology, a sane and healthy global brain. But before we dive into Brigham's life and work, let me please remind you that this podcast is a labor of love, entirely listener supported, and utterly impossible without your backing. I went through a dark night of the solar the last year and emerged around my 40th birthday this month with greater clarity than ever about how I want to serve the world. I am passionately driven to help facilitate vital discussions on wisdom, power, and responsibility in an age of magical technologies.

I can think of no better way to spend my precious life and to foster more public conversation on how we'll feel our way across the river stone by stone in times like these. Edgework like this isn't easy in the best of times, but thankfully the pace of change these days has made it easier to recognize the value of generalist thinking, wayfinding, ongoing adult co-learning, and open-ended inquiry. The silver lining of living through such incredible disruption is that it's teaching us to work together in new ways and I am immensely grateful that I get to be a part of this. But I have kids to feed, so please, if you find value in the work I do, support this show on Patreon or Substack or become a sponsor.

Patreon.com. And Michael Garfield.substack.com are both found in heads of writing music, art, and generative dialogue where you can find extensive archives anytime you need to shift your point of view. So help bring back the golden age of competitive philanthropy as an infinite game of public goods and provision, and give your money to future fossils, and of course also to goodly labs. Big thanks to my most recently enrolled supporters Rob Segal, Trevor Lyons, Guido Willemson, and Cody Kuiak, and to everyone who welcomed me into my 40s by reviewing future fossils on Apple Podcasts or Spotify.

And to everyone who has commissioned artwork from me or hired me as a consultant or advisor, and big thanks to everyone who hears this and decides they will. One more announcement. The Future Fossils Book Club is back. Substack and Patreon supporters at AnyTear are invited to join me for two video calls this coming month on Iron John, Coet Laureate, Robert Blyes, Smith Wapoeic, Investigation of Manhood in the Modern World.

I'll host two open format safe space-handed and off-record co-learning calls on Saturday, January 27th and Saturday, February 10th, and 12-2pm Mountain Time, in which participants of all genders can find meaning and mutual support through the discussion of this potent and reflective book. And then get ready for two more calls on Michael Crichton's novel, The Lost World, the best selling sequel to Jurassic Park that introduced me and millions of other readers to Complexity Science. More on that soon. First, here's an awesome 80 minutes with the wonderful and inspiring Brigham Adams about one of the most puzzling and inscrutable phenomena in the known universe, human society.

Thank you for listening. Alright, we're here, I have my snacks, you look great but we're not doing videos so people will just have to take my word for it. Brigham, how are you today? I'm doing pretty well.

Life is coming fast, there's a lot of waves to surf and having fun, they're pretty big waves, crashing would hurt, but so far, so good. I want to hear about the waves because you happen to surf a rather large and consequential board on some large and consequential waves, but in order to get to that part of this, I would like to do the part that I sometimes forget to do, which is to solicit your backstory. And I think that anchoring who you are and what you do and why it matters and how you got to care about the things you care about and learn the things that you've learned is the right way to welcome people into the life of your mind. So why don't we start with little you?

Yeah. Okay, we're going to go a bit deep. All the way back. I'll tell you the story of my first encounter with society because I am a sociologist, create pro-social technologies and anyway I can I try to support the ecosystem of people building these pro-social technologies and we can talk about what that means.

But yeah, when I imagine where all of this started, it was pretty early in my age three to age seven in the process of building sentience and I can isolate my first encounter with society when I was seven years old. I'd recently moved to Pueblo, Colorado from Louisiana. And my family had lived in on the Bayou in Louisiana and your Bayou Lacan and Slidell, which is just across like Pontchartrain from New Orleans for those who know the map of Louisiana. And we moved to Louisiana to Colorado.

I was in a music class and the teacher of my second grade music class said, Hey kids, there's a new band that you might all like. They're called the new kids on the block. Some of us might remember this boy band called the new kids on the block. And the teacher said, at first I thought they were black, but I guess not.

This sentence just like rung in my ears because I was like, what is she talking about? I don't, how is a band black or not? What does that mean? So I went home and I asked my parents and they revealed to me the whole sorted history of America's tortured journey with race and the creation of these categories of race and the enslavement of people.

And I wanted to know who did this, why did they do this? And it was incredibly deep for me because I had just spent the last four years in the care of a woman who apparently was black. My mom left the household bear. It's complicated, but she was not in the household starting at age three for me, which is a pretty crucial time to not have someone's mom in the household.

But I spent probably 60 hours a week with this woman. She was running a child care facility out of her home and over half the kids were black. She was black. I was getting raised by her.

She was a brilliant, hilarious, intelligent, warm-hearted woman and just incredibly loving to me and incredible stand-in mother. And when I learned at age seven that society creates these categories that are so hateful, so cruel, that was the beginning of my journey to figure out what the heck is this society thing and how do we heal it, how do we make it better? Yeah. So I can pause there, I know, but I like this grows.

The story grows, of course. Go on. Yeah. So that was age seven.

By that point, I was born in Goose Creek, South Carolina there for two months in Denver, Colorado and then Florida and then Louisiana and then Fudlow, Colorado and this is where I had the music class. And then for fifth grade, we moved to Houston, Texas and then for high school, ninth grade, we moved to Franklin, Tennessee. And so I was constantly on the move. I never lived anywhere longer than three or four years as a kid.

And I think this built up in me some sociological imagination and capacity, just seeing so many different slices of the country, slices of the culture. As a family, we were up and down the socioeconomic ladder and more or less moving up from a passage of the situation near poverty to a situation of upper middle class by the time I was going to college. So that was a lot of comparative data to see that the way humans do life together is contingent. There's not one way to do it.

Every place I go, they seem to think there's one way to do it. But I know, based on my experience, that there's lots of different ways to do it. And so that kind of built up in me even more interest in curiosity and how this thing called society gets constructed and how people get so accustomed to seeing it as the static container that they live in without much awareness that they're actually building it. They're building it and rebuilding it.

Every day as they go, I was encouraged by my father, who was very much a rationalist and had a bit of an anti-authoritarian streak to be very critical, to be very skeptical minded. And that got me into doing debate in high school and mock trial and stuff like this. And I was in a context in Tennessee that was dominated. And I use that word somewhat, intentionally, by evangelical Christian metaphysics and imaginations about what the world was and what our role in it was.

And I went the other direction and read a lot of Eastern philosophy while I was in high school, trying to understand different ways of seeing everything in college. I studied philosophy in political science. And I got into politics a bit, electoral politics, helping to run a political sacred campaign by the end of college. And I worked as a volunteer at the ACLU and for the Sierra Club and stuff like that.

I did a lot of knocking on doors and talking to people. And I worked at Hotel Friend Desk, which was also a place to get a lot of interaction with a lot of people. I worked retail. I worked at a candle shop after college.

So I get that you don't get that kind of experience unless you put yourself in it. Anyway, when I look back, it's just, you know, if someone's dead to me right now, hi, I'm five years old and I want to be a sociologist someday, what should I do? Like, the path I took is not bad training to develop that understanding of how people show up differently in different contexts and how we create our world and stuff on the fly. So when I finished undergrad, I actually had my head on this experience on the campaign.

And I came back after a year farting around to use a phrase from Vonnegut. I came back and ran a US congressional campaign at the age of 22. And I learned that I was like, not bad at that and that it's not terribly hard to do campaigns, but that a lot of people who show up for that don't actually understand how the world works. And naively, honestly, cutely, perhaps I should go to graduate school to see how society works.

And so I applied for public policy schools and sociology departments. And to my surprise, I got into UC Berkeley, which was the top sociology department in the English language. And also a place that in my mind had this mecca-like character to it. I saw myself as this intellectual who was counter-culture from the standpoint of the evangelical Christian culture that I was growing up in during high school.

So this was a dream come true. I didn't know exactly what I was in for. Actually, I have very little idea what I was in for. And it was a mixed bag with some excellent stuff and ultimate success of getting a PhD and becoming like a real sociologist.

So that's at least part one and two. And I could go out and I don't know where you want to jump in here. The vagueness with which he referred to the congressional campaign suggests I shouldn't pry. But I would love to know maybe the way to take this is to understand it strikes me as rare, at least among the people I have met, for someone to have on the ground political experience and then to enter academia.

I'm telling you, the words I use for this are very earnest and it's very precious. But I really felt like I might have a career in politics. But I feel the responsibility to really understand how society works if I'm going to hold the power to change society. And I didn't know enough about what an academic career was and I didn't know enough.

No one in my family had gone on to so much higher education. My dad and my stepmom had undergraduate degrees and they went on actually my stepmom got under some degree and my dad got a law degree that he never really used. But my siblings weren't going to college. We didn't know anything about what PhD programs were.

I just thought it would be like an extension, a more intensive, deeper version of undergraduate. I actually thought that I was going to go to grad school and continue playing Ultimate Frisbee. That I would have time for that. So yeah, I went there to learn more.

Along the way, we also, during that same time in parallel, we saw the rise of Facebook. And this really impressed me. Here I am, I'm like reading Marx in favor and Durkheim and Foucault and Norbert Elias and all of these macro sociologists that explain the gradual accretion of different cultural forms over decades and centuries that produce certain kinds of social psychologies that produce certain kinds of communities and economies and cultures and politics and all this big macro sociology. And I just can't help noticing like this guy Zuckerberg gets a billion people kind of dancing his dance in about a decade.

And that used to take a millennium to get a billion people to be Christians and dance the dance of Christianity or to be Muslim or to be Chinese or whatever to get a billion people. It takes a thousand years and Zuckerberg did it in 10. And not because, you know, I think he's surely an intelligent person, but not because I think he had the very best song or the very best beats or anything like that. But he found with this technology, he was able to scaffold and guide and choreograph the behavior of a billion people.

And so when I saw that, I'd wait a minute, here we are in parallel experiencing all kinds of gridlock in terms of policy. Like the policy changes in the 1980s until now have been in fits and starts and pretty, I don't know, pretty mild compared to what people might have imagined. And as someone interested in helping to improve society, to do it at scale, my head was definitely turned by internet communications technology. I started getting pretty curious about whether that was a better way and started becoming more interested in ways of doing collective governance and collective intelligence through online technologies.

Yeah, and just moving a bit more in that direction and then being molded into an actual social scientist. Like, I actually made that transition from reading sociology to starting to do sociology. And I was turned on by that. And then with my dissertation work, started to move into some very ambitious data science work, which following my motivations there, brought me into working on text analysis tools, these computational linguistics tools for understanding the social realities that are recorded as textual data.

A lot of our reality is recorded in our languages. So learning how to do that. And along the way, I learned all the different things that were available for understanding the meaning and texts. And I was not quite satisfied that they provided outputs that were fully aligned with the complex and intricate theories of social reality.

Like all the things that can happen are recorded across these vast literatures of single case studies. How do we take everything that can happen? Put that in some sort of theoretical schema and then apply it directly to some large set of documents to make sure that your data, your big data, which was the thing at the time, are actually aligned with theory in a way that allows you to ask and answer hypotheses about the relationships between different elements of those data. And yeah, I was like following my nose and not knowing how hard it would be.

I ended up inventing some technology for doing intricate, rigorous, large scale analysis of documents according to intricate expert on apologies via these annotation assembly lines where you break the work out into these simpler tasks that people can do online. I want to pause you there because it's been a while since I don't interview as many social scientists on future fossils as I did for Complexity Podcasts. And all of them tended toward SFI as like a fundamental theory kind of a place. But they brought in people that had very different ideas.

And there is this tension in the sciences, which I know I imagine a lot of listeners know, but this tension between understanding and prediction. And when I hear you talk about following the data and trying to balance it by titrating the appropriate amount of theory into your analysis, I'm really curious how you specifically how you handle this because this is I think this bears on things I wanted to talk with you about later. We'll see. It's like kind of a chicken or egg problem because these are two very different heuristics for the explanatory value of a given approach.

Like some, there's a ton of people working in AI that do not care about theory at all. And they have this weird sort of this bias that's like the view from nowhere of 19th century anthropologists that they believe that the data is not going to lie to them, like they're ignoring the postmodern context dependency turn of data production. But they have an interesting point, which is as long as you're not staring into the ontological abyss of like how your data is produced in the first place, whether it's trustworthy, then you can feel pretty good about letting it tell you what are the meaningful features of that data set. Whereas there's this whole other thing, which is in the sciences, you have, I did not expect to get into this with you today, but this is great.

There's this whole other thing that there's like the other side of the argument related to that, I should say, is this piece, it seems like a lot of scientists I have met over the years have a kind of disdain for theory in the humanities because it is not rooted in a quantitative model as it is in say physics. A lot of the physicists I've spoken to over the years as progressive and open minded as they may be, maintain the sort of residual disdain for what to them looks just armchair speculation among historians and adjacent fields. So I'd love to hear you just go on a bit about the way that you balance all of this stuff or did in your work. Yeah.

I entertain myself with a bit of, I entertain myself with like my own snootiness and towards some of the other disciplines, even though I think there are really excellent scholars who are interdisciplinary coming from a lot of the different social sciences and humanities. So my quick loss on historians is that, to historian, history is just one goddamn thing after another. And for sociologists, we're looking for patterns, we're looking for reliable and repeated mechanisms in history that play out. We're looking for the ways that history rhymes and figuring out how to build those into a theoretical poetry that people really rock and that is explanatory for the world.

Now as to the different perspectives of someone like a machine learning engineer versus a social scientist. I think that the perspective is so different mainly because of what each sociologist versus machine learning engineer sees as their job. And for a social scientist, we're not merely trying to make predictions that improve return on investment by 15 or 20 percent or 30 percent or something like that. We're not really satisfied with an answer that's 80 percent and we don't quite know why it's right.

And one way to think about this, speaking most definitely for myself, but I think for a lot of social scientists, is we are really interested in improving and upgrading the agency of humans individually and collectively. I want to see many more sociological agents. I want to see many more people who understand how precisely we are constructing reality and reconstructing that reality. How precisely the mechanisms of interaction that seem like these kind of micro tiny little scripts that we're running actually compose up into a more or less concordant or discordant symphony of interactivity across social surfaces that in its best connotation we might call civilization.

I understand the criticism of humanities as like a bunch of armchair theorizing and it's because there are different moments in the process of science. You see this in physics too where people call themselves theoretical physicists and people call them experimentalists. They call themselves experimentalists. And the social science is we have a lot where they're talking about anthropology, political science, even some in economics with certain society.

You have a lot of case studies. And case studies are really important but they're incomplete. They're not like everything we want out of science. What they do really well is catalog what can happen.

So when I was doing a dissertation looking at the ways that police and protesters interact during moments of political contention where they're street protest, I look into the literature and there are just like many hundreds of case studies where some sociologist has observed a single protest movement interacting with a single police force in some single city at some single point in time. And we can learn something about the interactions between police and protesters that way but it's what we learn might not be applicable ten years further in the past in a larger city where it wasn't a unionization movement or something like that but it was had something to do with political parties or something like this. You really have to look across all these different case studies to start to see some of the patterns. And it's that bigger data approach that really allows us to tease out the factors that are driving those dynamics.

So with my dissertation work, I looked at the Occupy movement. I saw this movement that had the same goals and tactics. It was happening at the same time and it was happening across 184 different cities and towns in the US. I told my advisors I want to study the Occupy movement and see why the police behave the ways they do towards these different movements that are fundamentally similar.

And my advisor said okay, like four cities, maybe six cities, if you're really ambitious pick eight cities. And I could have done that sort of work written a paper, told a story and I would, there would be a bunch of machine learning folks and folks and physics that are like, what can we really know with the comparative leverage of an N of eight or N of six. And the secret to my success in some ways has been my eVite and an overestimation of my capacity. Let's go right at the top of the episode.

So I'll settle myself these gigantic goals. Because what happens in my mind is I can imagine the thing and I'm like, oh, that thing is really cool. It would be valuable in various ways. And then my mind is, yes, is it physically possible.

And I have enough knowledge of what works in computer science and across a number of domains to understand this is impossible. And then my mind is okay, it's possible. And I'm like, all right, let's do it. And I think because of my experience, probably in political campaigns, I've learned that you can, if there's a lot of work to do, you can rally and it's like a worthy cause you can rally people to get it done.

And that's exactly what I did. I told my advisors, I'm going to go after all 184. And my hack was that I started teaching the method scores. And this is UC Berkeley, just like really undergrads.

And I was able to recruit the top undergrads. These are students who over the years, over a couple of years of working with me, they would go on to win the awards in the Berkeley sociology department as the top undergrads. So I was able to recruit them in and then we went and scraped all the news and television and radio transcripts and stories about all of these occupy movements. And we started annotating them.

And this was my journey into text analysis. And it just so happened that this social science computing facility called the D Lab, where graduate students teach themselves the latest greatest cutting edge computational methods for social science. That thing was coming into being just in time, it's like skipping down the rocks of an alpine creek or river. You have the momentum and then the rocky need tends to appear in the moment you need it.

And I ended up moving into the D Lab and helping them launch that thing, but also learning everything I could from other graduate students about text analysis and then starting to teach it myself to really make sure I fully got it, understood it and we started applying that to all these news articles. And yeah, I just wanted something that was having higher fidelity to the actual theory. And that's where I started coming up with the idea for CAG works, which is this software that makes it possible to turn these very intricate expert analytical labeling jobs into annotation assembly lines of simpler tasks that people can do on the internet in a way that's actually reliable and rigorous that meets the standards of academic research. I thought I'd let you wade deep into the weeds in maybe the wrong direction, but it seems like we have wandered within the eyeline of the work that you're doing now, as far as piecemealing the sort of complicated and rigorous tasks.

And so I would love to note, one can assume that while this whole story you've taken us on, while all of this was unfolding, you were slowly developing ideas about truth and fact and coordination and collective sense making and so on. But at what point did it become obvious to you that you wanted to allocate yourself to the work that you're now leading at Goodly Labs? At what point did helping people devise new ways of navigating the, what I love my mentors, Richard Doyle calls it the info quake, at what point did that come into view for you? And then how did you transition into starting this group?

And I guess maybe tell us a little bit about the group. Yeah. Yeah. I founded an organization, a nonprofit organization called the Goodly Labs.

It's a goodly G-O-O-D-L-Y. It's an ad verivization of good. It's actually used archaically to refer to like numerosity, like a large number of something and or something that's like good or handsome. So I imagine I think of Goodly Labs as a lot of people doing a lot of good.

And that's exactly what we aim to do with the prosocial technologies that we research and develop at the Goodly Labs. So Goodly started, actually I founded it with the same set of undergrads I was referring to and everything we did was just like, is it possible? Okay. Then why not?

Let's go for it. So I think I've talked to people at the time about how hard it is to establish a nonprofit. We might not have done it, but we just, we heard it was possible and we heard there were some forms to fill out and one of our undergrads had some connections to someone who worked at a nonprofit who had connections to some lawyer that was willing to spend a half an hour saying, look, this is how you do it. Here's a template.

I established the Goodly Labs because I wanted to make sure paying as many of those undergraduates as I could for their work during the summer. And I just, my crowdfunding was a new thing and we wanted to give it a try. Our crowdfunding attempt was not ultimately any kind of smashing success, but out of the experience we established a nonprofit. And I moved more and more of my effort and my imaginations about my own future into Goodly Labs as it was becoming clear to me, I just kept getting signals over and over from academics, people that were mentors to me and like very caring kind mentors to me in the sociology department and people at major social science publishers, including Sage Publishing, which also at various times was a very friendly kind mentorship role to me.

But I kept getting this information that I was blowing my career because as a, as an early researcher, I wasn't supposed to be amassing a giant data set. The data set for that occupy project is something that we just kept on a simmer for the last few years because as I'll tell you in a moment, we've been distracted by a project that might be more urgent and we're really excited to launch here soon. But I was getting the advice that like by building up this massive data set, it would produce six big monographs and probably many dozens of papers. I was doing it all wrong.

That's something you do when you already have tenure. And then what I needed to be doing at that time was creating a solo author paper that would demonstrate my capacity to do the entire research process by myself. The literature review, devising the study, properly setting up the methods, actually executing on that, doing the analysis, doing the write up, all of this entirely by myself, which from my standpoint, I'd already done this with a pretty ambitious master's paper. And I don't know, it just wasn't an interesting goal to me to create like a single fiber in a centuries long rope of literature.

In the way I was looking at society and seeing the trends, how things were going and the trends of how things were going, my feeling was like, we need 10 more feet of rope. I'm not going to spend time creating a tiny little fiber. If I'm doing something bigger and more impactful means that I don't get to be a professor, then I guess I'm not going to be a professor. I guess I'll create my own lab.

And that decision was never difficult for me. The carrying things forward and finding that there was a whole set of people who would give me slaps on the back and tell me how great this data science was, but couldn't ultimately shake loose institutional funding. That made things a little harder than I'd hoped it would be, but there have been lessons in that too. So no regrets, even though it's been a process of what I call bootstrapping everything on my shoestring budget, yeah, I'll pause there because I'm wondering where we go next.

Yeah, so we are 40 minutes into this call. And I have not once said post truth or epistemic crisis or deep phase or 2024 election. So it might be worth rather than seeing what the lowest entropy path is to this stuff, just dog-lagging right into it. Yeah, no, I have a story that's a perfect segment.

Yeah, please. So during this dissertation work, I'm at the Berkeley Institute for Data Science at this point. And the director of the Berkeley Institute for Data Science is Saul Perlmutter. He is the Nildall Loryat who led the team that discovered that the universe is expanding at an increasing rate.

And very, very magical connection to Saul because when I was an undergrad, I took a course at Washington University in Singles on the cosmos. And it was just like an intro course. But they taught Saul's study and they showed his methodology, which is his natural experimental design. And I just found it so clever and so interesting that I think it actually may have sparked in me the possibility that maybe I would go to grad school to be some sort of scientist, because it's just so incredibly awesome that you're able to figure out by looking at these supernova type 1A, the standard candle, doing the reverse engineering, the theory and the math that figure out that the universe is expanding.

I thought that was so cool. And then years later, I mean, in an interview with Saul Perlmutter, where he's the director of the Berkeley Institute for Data Science deciding whether I get to go in. And I made it in. And a couple years in there, he says to me one day, he says, I hear you've been, I hear like your tool for analyzing documents is coming together rather well.

And he said, I've been teaching this course for a few years now called Sense and Sensibility in Science, where I teach 400, 500, 100 grads at a time to upgrade their kind of critical thinking skills. And in particular, teach them some of the thinking tools of science that allow us that empower us to be less wrong incrementally over time. And he said to me, I wonder if we can use your annotation tools to pick out some reasoning errors, some of these reasoning errors that I'm teaching in the course, we could pick those out in the daily news articles to give examples to the students. And I said, Saul, that's a really great idea, but let's not just do it for your classroom, let's do it for the whole world.

And that moment was the genesis of a project that's called Public Editor, which does just that. It empowers people with tools to go look at the daily news and label directly onto the words and content. The other words and phrases of an article, we can label directly onto those words and phrases over 50 different types of reasoning errors and cognitive biases and linguistic manipulations that we humans use to fool each other all the time, sometimes intentionally, sometimes unintentionally. And this is giving humans a new capacity to make some sort of sense out of the coffin of information that's flowing through our information environments.

And actually, a lot of that misinformation ends up being rewarded by the algorithms of the different platforms because misinformation is outrageous and interesting. And it gathers a lot of attention. As Deb Roy at MIT showed maybe five or six years ago, false information travels six times faster than the truth online, which of the forces of red blend of an old between the lies halfway around the world while the truth is still lacing at its boots. So, I'm very acutely, I wish I might not know if I want to go forward with this marketing or not, but with public editor, we say the truth just got delcro.

So it can catch up to some of those lies. There's a magic gathering card called Swift Boots that seems to show up in nearly every deck. I'm going to apologize to listeners now because my friends have rubbed me back into this game as a way of maintaining my sanity. And so you're going to hear a lot more gaming references in the year to come, which is like part and parcel with predictions made by both science fiction author Charles Strauss about what happens when a generation of millennial gamers matures to be the sort of lead cohort, but then also comments that have been made by Bill Gates and Yuval Harari about what we do with ours.

And look at Ernst Klein and Ready Player One, right? This whole ball of stuff about what happens to the technologically unemployable and that we preserve ourselves culturally and psychologically through ritual and that ritual is often enacted through game and sport. So anyway, that's a total tangent, but to get it back on topic, you and I have talked about this several times, the need to find some way to augment the processes of truth with a haste ability, if you will. That's one of the only people that I've ever spoken to, and maybe just because I hang out with a lot of cynical, concerned academics and jaded hippies, but you're one of the only folks I've ever spoken to that has this sort of can-do Californian engineering confidence, the constraints that we're dealing with here, they're not necessarily immutable, that we can at least engage in the good fight, that we can have an arms race here rather than just resigning to the fact that we're going to get bowled over by this inherent speed asymmetry in the brain, right?

Because ultimately, even in an individual brain, executive function, you've got that economy in system one, system two stuff, right? It takes a little while to review something critically even as a single person, much less an entire group of people. So this is where I'm starting to recognize your Mark Zuckerberg tie in from the beginning of the episode saying, okay, how did he do it? A thousand times faster, right?

Network properties. Yeah, so let me run on this a little bit. The other thing I was also saying is that we're really interested in upgrading the agency of humans and our individual and collective agency. Right now we're tasked, if you're a normal quote contributing number of society, you're tasked with having a job buying stuff and voting every couple years or so.

And you're really not expected to do too much more than that. Like maybe you have a family, you have some responsibilities there, but from the standpoint of the society at large, that's not a lot. And it's not really an invitation to create or recreate or play some important role in your society. But we're at a moment where there are some real societal level challenges that we're facing and we need to coordinate to confront them together.

The situation of misinformation flowing through the platforms is very much a consequence of this attention economy where the more people's eyeballs are on a screen, the more advertising dollars are earned by those platforms. And it's an absolute good, I would say, that folks like Zuckerberg and the Googles of the world and Twitter and et cetera have created these tools that allow many more people, I don't know how many 1,000 or more people, like 10,000 or more people, I don't know, to have a voice that could potentially reach a million people. That is absolutely good. It also creates the possibility of a cacophony.

And then when your algorithms promote the stuff that's most outrageous, you get that sort of cacophony in there. There's a great need. And I'm not nearly the first to say this. And Tristan Harris and folks at the Center for Humane Technology have put up this wonderful kind of documentary, this very slick documentary, the Social Development a few years ago, pointing out these dynamics and how there needs to be a solution to this.

But we haven't really seen a solution because mostly we've been hoping that the platforms would voluntarily tweak their algorithms in the direction of better, saner, more rational discourse. And that is against their imperative to maximize quarterly shareholder returns. That's at least one reason they haven't really done it. With the tools like public editor, we can actually scale up the capacity of humans to look at this content and do the close analysis.

Even Sol and I have been working on this thing with a team of great postdocs, cognitive scientists and journalists and librarians and other social scientists. We've been working on this public editor, this kind of social epistemology engine for a number of years. And even he and I, if we read an article, we miss some of the stuff the public editor would find. It's just too hard to hold 50 different models of these different cognitive biases.

We've got like a peeled authority and ignoring selection effects and correlation, and anecdotal data and the false dilemma could go on and on. And to have a model of each of those in your head while you're just trying to intake the information about what's going on in the world today, it's pretty much impossible to apply all of those models simultaneously. It just swaps the capacity of the humans and the smartest people. But together we can do it.

And the thing that's really neat about this system is that we can actually do it in something like real time where we're intervening quickly in the news cycle. So there are a lot of fact checkers out there, and we really appreciate the work of fact checkers. It's important work. So there's some claim that this state of affairs exists in the world.

That claim can be checked by fact checkers and sometimes it'll take 20 minutes to confirm or just confirm that claim. But it might take 20 weeks. It might mean that you go to some world village and talk to someone who says, actually, you need to talk to this guy over here. He doesn't get back until next Tuesday or something.

It could take a long time. What's pretty neat about public editors within just the four corners of the article, these 50 plus cognitive biases and rhetorical manipulations and reasoning errors, they end up being really great signals of slanted or biased content. There's a very high correlation between the appearance of these manipulation tactics and these reasoning errors. There's a very high correlation between the appearance of those and the appearance of false facts.

But we don't have to wait to get the reporting on whether the fact was true or false. We can within half an hour of the publication of an article, we can turn it around with a bunch of labels on it pointing out to the reader exactly where they need to apply some specific doubt to a particular claim. We're not coming through with a heavy hand and saying this entire news site is junk. It gets a red flag like an entire news site or even an article.

We're not putting a big red flag on an article and expecting the audience to discount all of the information in the entire article. We're giving the appropriate specific doubt to the appropriate specific words where we're not necessarily saying the author's wrong. We're just saying they pose this idea as a false dilemma. So just be a little aware.

Learn to look out for these things. So we call this the, I have to get my mom credit for this one. This is a woman who can turn a phrase. She said, it's like Google Maps for navigating the biosphere.

I was like, that's great mom. The biosphere. She's going to be that up. Okay, I want to linger here because the question for me then becomes if in Google Maps on an actual terrain, you can, I think about semantic spaces, right?

The distance between one limb of a three part analogy and the other limb, the relative lengths of those, like ideas, the networks of association in one person's mind versus another person's mind, one idea might take somebody two leaps across an associative network and it might take somebody else five. But that's different from sort of statistical tendencies, more like an insurance evaluation, like an actuarial thing where it's okay, yeah, generally if you catch somebody in a logical fallacy, they're human. They're going to get doctor point relative to somebody who's just totally impeccable in that regard. But like, how do you assign relative values for the frequency of association between somebody who, for instance, might have argument to authority being more likely to have certain other kinds of fallacies that they tend toward versus, but they're not all created equal.

Some of these are much more pernicious and harder to spot. I think about there's this great episode of Stuff to Blow Your Mind where they looked at politically motivated and numeracy. So it's funny that you just went on this whole thing about how you don't even catch it all the time. They were saying that the research on being able to bullshit yourself with reading the numbers the way you want to or being selectively ignorant of mathematics seems to be more pernicious and advanced in very intelligent clever people.

Like there are certain ways in which the smarter you are, the more capable you become at lying to yourself. Yeah, I'm just curious how you, what ways are out of map all of this? How do you know drive time between one set of logical, false and another? Yes.

Okay. There were two main things I wanted to say in response to your question. The first is that for the first round of products we release, we're really just focused on giving that daily news reader the tools to see these different forms, these different forms of error fallacy bias manipulation. But in some future products we will also release something that provides a what you could call a bullshit signature or a misinformation signature for an author or even for an entire news site.

Like this particular news site tends to lean heavily on the appeal to authority. And then we also see as you suggest that with appeals to authority there are correlated these other types of manipulation or fallacy. We're extremely interested in seeing how different maybe political ideological communities or just like organizational networks are display or fall for or appreciate these different sets of errors and fallacies together. So there will be some great research out of this data that dig into that more.

If we're imagining that this exists across communities, I think the next reasonable question is okay, how do you make sure that your community of annotators is not biased by just coming from one of these communities? And there are a number of ways that we address bias or try to squeeze bias out of our system. In the presentation of tasks to our annotators, they are looking at just one small chunk of text at a time. So we actually set them up with a task where they're not reading the entire article, which gives them a lot of information about the point of view of the author and kind of starts to trigger their biases relevant to whatever kind of political or policy or ideological domain is discussed in the article.

Instead, they're just focusing on a small block of text. And then they're answering questions about it one by one. This is a lot like a reading comprehension task that we've all been subjected to thousands of times if we have any relationship to the British Imperial School and system. And for some of the questions, the annotator is asked to highlight the text that's supporting their answer.

And that's actually how a label gets applied. So they might be asked a series of three questions that finally ultimately get to the point of, oh, what you're seeing here is actually an appeal to authority and they're checking some boxes of appeal to authority and then they're highlighting the text that is the appeal to authority. And then they get a follow-up question that asks them to discern how misleading is this appeal to authority. And this is interesting because sometimes the most ridiculous appeal to authority is actually less misleading than one that's more subtle.

And we want to bring in a bit of human judgment on that. Like, do you think the typical reader will, for instance, and here one of the things we look out for in our language modules, so we have these different annotation modules that make up the assembly line in the language model. One of the things we look for is exaggeration. And if there's some line that says, wow, that insect was as big as the state of Texas.

That's an exaggeration. Very easy to spot. It's also not really likely to seriously mislead someone. There are very few people that would believe, oh crap, we're like getting invaded by alien insects that are as large as the state of Texas.

So they can rate that as an exaggeration, but they can actually recommend that no points be taken off for it. They're going to do that task, but at least two other people are going to do that identical task. And then we have an algorithm that's looking for a consensus on both the categorization of the error and on the severity of the error. So if one of the people says this is not misleading and the second one says it's slightly misleading and the third one says, yeah, it's slightly misleading, then we will take the average across those and maybe a point eight points will be deducted or something like that instead of a full one point.

Another thing that we do to make sure that we're properly contextualizing the error in the article is that the job of one of the annotators is to pick out the different arguments in the article to identify the main argument of the article and also to identify arguments that are like either minor arguments or arguments that might be included in the article because the article is refuting them. This way we can show the error and the arguments being refuted, but it doesn't actually result in any credibility points being removed from the article because the article is actually appropriately raising awareness about that reasoning error. Aha. Okay, so there's two things that seem we're bringing up to folks that want to click to expand and learn more.

One is back in the day, a couple of years ago, I'll try to find the link and throw in the show on this. Simon DeDao of Carnegie Mellon did this talk at SFI on mathematical proofs as networks of arguments rather than linear chains in which each statement builds upon the previous. We have this way of thinking about arguments in which you make one point and then the next point rests on that one and he's actually more like a table where you can kick a couple of legs out from under the table and the table will still stand. It might wobble, but you can put a coffee on it.

You might not want to put a hot coffee on it. You put a one diagonal. Right. Yeah.

So I'm thinking what you're saying that rings with that because we have this tendency and this also goes back to your point about waiting the assignment of the context and the relationship between if this person said this, they're less likely to be believable in some other way. I think most of us, I might be doing it right now. Fault to this person is untrustworthy in a certain way. Therefore they are blanket untrustworthy and especially as society gets bigger, there's that great line in Sensei where they're like, yeah, you meet more people in one day now than you used to in your entire life.

Maybe this is consistent through history. We don't give strangers a ton of the benefit of doubt. Meeting somebody online is like meeting them in a dark alley. You don't have a lot of time to get to know somebody and break bread with them and then you end up with a lot of unnecessary violence.

And so I like in this approach about helping people not throw out statements of value simply because they are uttered in proximity to statements that are less valuable or simply harmful. I think this is especially important in the broader context of what I call the narrative wars. We have all of this technology for communication, for many to many communication. You can broadcast things like we did in the 20th century with television radio but we can also reach out through Twitter and all these other platforms.

And there is so much value it seems in winning some narrative war where your side is the hero and the other side is the bad guys. And we really don't want to be caught up in that. We also want to give people this, I don't necessarily want to call it a middle ground but we'll call it a higher ground for people to coalesce into. So we see things like, I recently saw that there's some movement against various different groups that have taken a crack at solving this problem and got some government funding and others this little pro movement with Jim Jordan on Capitol Hill and a few others, kind of moral entrepreneurs who are trying to say that this is the censorship industrial.

Whether or not that's true, I think, and I don't think it is, but maybe there's some concerns that are worthy of our attention here and there. You have groups like NewsGuard that are just going to try to say for each news site Red Yellow Green. And there are a few different orgs that are providing these sort of radio and reading ratings for articles that are basically just telling the audience like, hey, we are the authority on what is true and we've broken it down for you in a simple three part categorization and if it's green, it's true, if it's red, it's false, and if it's yellow, be careful. That is not the hand holding that people need.

We need to give people more granular and specific and specifically applicable kind of advice on what kind of appropriate doubt to apply and what sort of circumstances. And this is, I wanted to pick up on, you mentioned system one and two, kind of fast and slow thinking. I think we have a kind of a system one and two that operates at a social scale as well. And the slower, more careful conscious system two thinking of society is science itself.

It couldn't be slower. It couldn't be more careful either. And so when knowledge, when some set of claims has undergone rigorous scientific scrutiny and process, when it comes out the other side, we're like, hey, this is probably really useful. I'm going to go apply it in reality.

It's not really useful. Like this little set of knowledge shows me how to a mechanism to do this thing, but this tool and it works exactly as predicted. This is going to be more true in realms like physics where objects are dealing with don't have their own agency. In the world of humanities and social science, the objects themselves are social constructions.

And they are at any given moment, they're in some quantum superposition of I'll show up like I was yesterday or something completely new or some weird mutation of my normal behavior or whatever. But all these things are explicable by some manifold structural equation model of that person in time behaving in roles in various different roles from how they show up in their family to how they show up at work to who they were as a child, getting reprimanded and developing a behavioral complex to respond to reprimand to their visions of the future and who they want to be like, we are really super complex. But the goal I think is to be able to take that sociological imagination that helps us understand how we are constructed and be constructed on the fly as individuals and together, how we're doing that together and empower more and more people with that vision of this is what's going on. What's going on a regular basis.

And so what we see back to system one system two, we see this possibility that we can be collectively agentic, collectively intelligent if we can really use our system to social science to reflect upon ourselves and create the interventions that are going to bring us into that better world. The other system that we use to think together socially, the system one is this like crazy madhouse of the media news cycle. It is not careful. It is not slow.

It's incredibly fast, but it's also incredibly powerful. It's connected to political and economic agents that are doing shit every day that is constructing the reality in a way that the scientists aren't. The scientists are in that ivory tower. They are on the sidelines.

And I think what we're proposing, one way to describe what we do at Goodly Labs is that we're trying to create infrastructure that is in some way a bridge between that system one and system two. Now we're going to take some of the careful, less wrong approach and thinking of science and scale it up and speed it up so that interfaces with that media news cycle that's driving the relations of power and driving the actual actions and building and rebuilding of society that's happening every day. Because I have a dark mind and non-trivial inheritance of Scottish and Russian-Jewish gallows humor, I have to say this, and I don't want this to be taken either wrong way because I endorse this work. But I have to wonder, since we're talking about it in terms of the collective nervous system of society, if I think about the body as an encoding of evolutionary intelligence, you learn a lot about the world we live in by looking at bodies, the way the stable features of an environment are shaped in flesh.

And this difference between reflex and cogitation, like executive function, it's there. And it's wrong to assume that because it's there, there's a laundry list of arguments from what seem like design flaws in the human body against intelligent design. We get sore feet because bipedalism wasn't even like an afterthought, it just happened. Whoops, but yeah, when I think about what it means to connect the slow patient rigorous processes of science which are without question entirely too slow to keep their eye on the ball sometimes with rapid change, the wave we mentioned earlier, and it's funny, like watching, there's a great paper about this, about social science as a crisis discipline led by Coleman came out while I was at SFI and it's fascinating, it's 17 authors.

But yeah, to talking about how we need to come up with better ways of doing science faster, being more comfortable with the fundamental uncertainties, like we have to be able to act before we know for sure that we've done the right thing. So science is under pressure to think in a more sort of policy kind of way, right now, just because of the pace of things. That said, okay, but I think of the internet as fundamentally psychedelic. If you look at the map of the world wide web as a network and you look at the, I think it was Imperial College London study where they put people on LSD inside of a brain scanner and they showed increases in functional connectivity between parts of the brain that are normally inhibiting one another.

They look very similar. There's this weird thing that I really want to write about and explore more. I know Doug Rushkoff has written about it also and Alexander Biner and many others, Markoff, the history of psychedelics in the West and the history of computing technology in the West are basically the same story. Doug Engelbart, who was involved in ARPA and behind Xerox Park and the internet as we know it, personal computing as we know it, the grandfather of all of that stuff, Engelbart was one of the 350 people at the Institute for Advanced Study that was an LSD subject on Steve Jobs and Stuart Brand.

Anyway, what am I getting at? I'm getting at the fact that a lot of the epistemic crisis that you're trying to address seems to come from this all connected to all strategy that is writing on the network bonuses to performance and economies of scale, like Brian Arthur's preferential attachment. All this stuff. There's this whole thing where it's we think that, and so something happens like 2008 or cascading bank failures or COVID or something.

We think that the idea should be like, oh, these things aren't connected. Let's connect them. But then what happens is that when you couple oscillators, one of them, I think it's the fast one where there's like, well, the fast one ends up driving the slow one and the slow one ends up driving the fast one. They're not able to maintain sort of their unique strengths and partition functions.

And that's what I'm getting at. I don't know that this is a problem with the system that you're making and frankly, we may not have time to figure it out. No, I'm enjoying the exploration. Even before you brought in like the potential challenge, I was noting a crucial difference, I think, check my thinking on this, a crucial difference between the connectome of the brain and the connectome of the internet, which is that, as far as my understanding, goes with the connectome of the brain, you have at some point everything gets boiled down to a single audio visual projection that is processed by a single neocortex audience, and then some sense is made of it.

And not so with the connectome of the internet. There's billions of nodes just like in the connectome of the brain, but there's also separate audience members. And we're getting this bulk conversation of the understanding of reality because there are some audience members witnessing one rabbit hole of reality and others stuck in some different rabbit hole. And so I think it's plausible that in that more pluralistic network of functioning executives, it can be, it's both a different phenomenon to consider, but also it's plausible that an architecture that kind of pulls some of the features of system two into system one or bridges between the two or puts the two in service of each other could actually be, those two systems could be running simultaneously.

I don't mean to have derailed you here. Maybe this might be a performance, very problem. I think about, again, in terms of reflex versus slower, more methodical, integrated kinds of action. I was lying in bed and my daughter got up all of a sudden, she's a very fast person, got up and elbowed me in the eye and I immediately shouted obscenities.

And then it took half an hour to walk her back down off the proverbial ledge to realize that this very fleeting explosion of intensity was not something that she needed to register in her youthful plastic brain as daddy doesn't like me. And so when I hear it, when I think about it, you're right. So there is something very, maybe we just don't see the unified theater of which everything in the web is integrated in some sort of higher structural thing that we just can't perceive because humans were too dumb. I don't think so.

I don't think so. Not too dumb. It's just a matter of what we do to that. It's just we have our input bandwidth.

Fair. Yeah, you want to take a little bit of time. Right. But I think at the same time we can look at the network structure of the internet and at least make hypothetical statements about whether or not it has a central working space and it doesn't seem like it does.

Right. So this is reminding me of the question of is there a center to the universe and where is it? And you could plausibly, if you can find the boundaries of the universe, you could possibly find a point that seems to be the average of all the perimeter points of the universe. And you could say that's the center, but then you're making this like plausibly ridiculous assumption that everything has expanded outward at the very same pace.

And this is bringing me into a conversation about AI and about monoculture and about the psychology that I would love to just riff with you on the other day. I was speaking with some people about not to get too deep into it, but about some of the stuff that's going on in Israel and Gaza and the resurgence of anti-Semitism and how the Jewish people are seen in a particularly unique way. And then how actually the Jewish culture is somewhat unique in its introduction of monotheism to the ancient world or maybe not introduction, but it's popularizing of monotheism. And this is truly important to the later development of what we tend to call Western civilization and in particular the psyches of the people who live in Western civilization because ultimately to the extent that someone has a model of the cosmos is something that large and powerful and luminous they want to be in alignment with it.

And if we have this notion that the cosmos is governed by some singular entity, not by a committee of gods, but by a single god, then we're looking for this like reduced dimension vector of the preferences of this one god. Instead of having a notion of each god as their own preferences for how things should be ordered and I can be in more or less alignment with any of them at any given time, but that may or may not be able to satisfy all of them simultaneously. If that's your view of the cosmos, then your view of how you are in alignment with the cosmos really gets into your psychology. And I feel like with monotheism, but then also I think we're seeing it with this version of AGI, this singularity vision of AGI, we're creating another sort of impulse toward this monocultural psyche and just monoculture generally, which is something I'm really concerned about because I have a sense that monoculture could really reduce the amount of beauty that's represented in the pluralistic kind of diversity of human societies and cultures.

Oh, wow. Okay, I didn't expect us to talk about this, but I'm glad that you went there because I just saw this, you and I are in this AI discussion group together. And the other day, one of the folks in that group, I think you saw shared this news item about Ilya Sutsgiver, as I say, the chief AI scientist at OpenAI. I can actually read the item, so yeah, I'm curious.

Oh, yeah. So it was about how he was leading internal meetings there with these kind of, what almost struck me is like revivalist Baptist Church, or like an ecstatic field AGI hands up in the air, trying to get people in chance about this stuff. And the other thing I wanted to mention earlier was this thing that you're doing about mapping the arguments, and I know you're collaborating on this with Jamie Joyce on a level that I'm not even fully aware of. But when I had her on the show, we were talking about what she's doing at Society Library to offer debate mapping as a service.

There is this thing where it's thinking in terms of planetary science, like getting the orbital point of view on a conversation because we can do that now with our technologies is intention with what's working against the other knock-on effects of these digital technologies, which is the erosion of reason, right? And ultimately, that's the question here. It's not about whether one approach wins. It's whether they're in the right balance.

Like this notion that we can rationalism as an ideology is itself irrational in its refusal of the overwhelming evidence that reason is built on the cognitive infrastructure of emotion, which is built on even simpler modules. And so before this call, you and I were talking about this view that is, I'm starting to hear more and more people espouse and I'm really glad to. It was espoused by Josh Shray in the Emeralds podcast. It's like that episode went viral and it's interesting to see so many people backing this position that in an age of generative AI as a mature technology, we're basically done with the problem of getting what we want with our tools and are now into a new kind of problem, which is are we wanting the right things?

Have we been initiated into the essay, what has been throughout history and esoteric understanding technologies and traditions shared among magical traditions of the source of his apprentice, right? Okay, you don't get to play with fire until you've learned how to wield it. And as somebody whose father is a card carrying NRA member, I came up seeing the entire entirety of that of even dangerous tools. If we're talking about the end user agency, and I'm not staking a personal opinion in this dogfight, but I do have some sympathy for the distinctly American position that rather than just banning guns, what we actually need is to teach people how to reckon with dangerous tools and how to use them safely.

It'll be really interesting to see how those things get remixed in the debates around AI alignment and safety in the years to come. What I'm saying is that it's one thing to advocate for a cultural retrieval of the language and forms around magical traditions and wizard schools and people being granted like driver's licenses into higher levels of utilization. Like we're just pouring lighter fluid on the fire right now by letting everyone adopt these tools. But then on the other hand, so many of the people that are pushing these tools into society are doing so out of what is rather blatantly a kind of religious zealotry that is being clothed in the language of modern secular reason.

Okay, take that whichever way you want. Okay, so the thing that I think is really important is to take even one step back and just ask what are we solving for and why? Yeah, sure. If it turns out we create these tools and they're extremely powerful, I'm all for some kind of initiation process that helps people responsibly hold them.

But I think there's been a mistake among these primarily Bay Area engineers in properly identifying the problem of our contemporary times. Based on the tools they're building, it seems like they've identified the problem as individuals aren't smart enough. We don't have enough intelligent ideas or we don't have enough like intelligent processes. And I don't really think that's the issue at all.

If I look around at the various problems that are plaguing us, particularly at the societal scale, they're not plaguing us and they don't continue to plague us because there's a lack of ideas about how to solve them or that we haven't had the intelligence to come up with a solution. The problem is that we haven't had the collective intelligence, we haven't had the coordination to actually pull people together, legitimate those solutions among a consequential set of people, a set of agents who will work together to solve the problem. So I think in a lot of ways we're solving for the wrong agency. We're creating these tools that by themselves are these individual agents that kind of model in some ways individual human agency.

The goal is to create an intelligence that can look so much like human intelligence or even better that we get confused into thinking it is a human individual intelligence. And I think that's just not really what we need. And especially when you look at some of the trends of inequality and so forth. These individual intelligence are going to be put to work by the human individual intelligence that have the most power and wherewithal to create 10,000 lots doing their bidding.

It's going to increase the problems of inequality and lack of coordination and techno-futilism and all this sort of stuff. And what we should really be building for is increasing our collective social agency. I would love to see us building not for a singularity but for a plurality and building tools that do not work unless multiple people are engaged at the same time. I had this Burning Man art project fantasy of a giant beautiful gorgeous earthship that does not run, does not move an inch until 200 people are sitting and starting to pedal together and then the whole thing will move.

And I think that's the sort of stuff that we need to be building if we want to have a society that is more equitable, that understands how to move together, how to improve ourselves together. And right now it's just solving the wrong problem. The problem is not that we're too dumb or that we don't have good ideas. The problem is that we need to figure out how to coordinate on them.

We need to look at every major societal challenge and it really comes down to coordinate that scale. Not the other idea. Okay, two things. And then I want to give you an opportunity.

I got to wrap this by the hour. So I want to give you an opportunity to – that was a beautiful sort of grace note with which to sum this. But I want to speak to you. Like the vision that you're holding is a vision that I read first in William Irwin Thompson's book coming into being where he said that the planetary culture, the new planetary selfhood, you know, because he was working with Lin Margulis who founded this theory of endosimbiosis for major evolutionary transitions that the cilia on the surface of your intestinal cells were once free living bacteria that are now obligate.

They've become a part of this larger thing. And so he was saying that basically like we have to go through this kind of transformation as a species where we stop seeing ourselves as the – whatever spirochetes used to compete for resources with these larger bacteria and with each other and literally row together the way that you see the little rips around an amoeba doing or a perimysium as it's moving through a drop of water. And then that's what we are. That's spaceship earth.

That was his – he went on tours with Bucky Feller and stuff. And so that was like – the image that he was giving was one in which the way that we finally get together and steer this thing is through precisely the kind of coordination that you just described and it's precisely the kind of coordination that you and your team, which includes a lot of great people we haven't discussed and to which credit should be given. That's what – I really – if I'm going to place chips on a horse in this race, I really see what you're doing as crucial in the realization of this cosmo vision. So I salute you, the collective and each of you as individual Celia.

Yeah, thank you. I personally really appreciate that. But let me also talk about some of the partners and organizations and like allies I see in this ecosystem. And you mentioned Jamie Joyce earlier at the Society Library and she's been doing phenomenal work.

She's an incredibly charismatic leader and a genius mind in her own right and I think props to her on building up this capacity for mapping deliberations. This is what we need more of, not like one off debates every four years where people who are trained to verbally joust, try to tear each other down and win points through demagoguery. We need 24, 7, 365 deliberative capacity and earlier on in Jamie's project we met and I had been working on some of my own kind of deliberation stuff and we still are and we were able to compare notes and see that we have some similar approaches, we have similar deliberative grammar I call it that makes it possible to create these big maps at scale while still making them traversable computationally so you can easily have different displays and zoom in and out. And Jamie's been doing excellent work similar to me in the sense that they're kind of boot driving it on a shoestring budget for quite a while.

But there are others out there too. There's folks in this newly forming collaborative technology alliance. There's the Open Civics project that Benjamin Life and others are leading to help build up some of the ecosystems around this. One of the initiatives that I think will probably collaborate with Open Civics on is a little incubation cohort we call the governance burden because the people who are working on this coordination technology are, it turns out, maybe no surprise, they're rather cooperative people and a lot of them are quite smart and competent and we've seen each other and found each other at conferences and while the philanthropic ecosystem, the economics of it would encourage us to compete in a zero-sum way for funding as we spoke to each other about our imaginations of what online deliberation and governance technology could look like.

We kept having this experience of seeing in the other a very sharp and capable mind, good heart and interesting hypotheses for how to solve these problems that we weren't actually holding us hypotheses. So we can look at each other and say, I want to know the answer to your questions but I'm focused on my own model, my own architecture for how this will work. And when enough of us had recognized this in each other, it became clear that together across all of us, we had hypotheses that really gave a comprehensive map of the solution space. There are not an infinite number of ways to skin this cat apologies to cats whom I love, but there may be dozens and the traditional model of that we says we should each be competing zero-sum, one or two of us will get funding and if those projects work out then they will go ahead and dominate the marketplace for governance technologies for the next 20 or 30 years and for all we know we'd be stuck on some local maximum in the solution space, something that's like that works well enough but we could have done much better maybe.

So we've pulled together as this incubation cohort called the governance garden so that we can do rounds of R&D in front of each other and share our findings and plausibly after each round merge some methods, learn that this didn't work and that didn't work, maybe take some feature set and hand it off to a friendly cooperative team that we could see as a rival but instead we're all looking for this omnivin for humans where we get the very best and deliberative governance tools to the most humans as soon as possible. I'm looking for philanthropists for funders to say oh yeah duh this is the way we should be doing things it's not about me the philanthropist picking the winners and losers from the outset if there are a few teams that recognize each other's competence and could will and they want to work together and make sure we get the very best solution let's get behind that. I like it it induces hope in me. We have to grab onto that hope is a fuel and I know it's been misused before we all got a lot of hope change in our years but there is a real there's like an incredible latent space of pro-social tools that we can be building we've hardly begun as a species we've barely even tried to use these internet communications technologies in ways that create super intelligence that create more output than input literally we've hardly begun.

So this to me is like the great technological revolution that is unsung and under-resourced and can actually provide a really helpful future for humans where the super intelligence that we build collectively together can actually contend with the super intelligence of AI and our collective intelligence systems can feed into and train and tame and parent the AI so a concrete example with our public editor project I was telling you we're aiming to do the top 100 or 200 news articles in Europe during their campaigns and elections in the same in the US and that's about all we can do with a human volunteer crowd is the top 100 or 200 articles which is pretty good it's like the over-to-window for society but once we've done that for six months or nine months we'll have enough examples of each of these different types of recent years but we can feed them into supervised machine learning models train AI and we can be labeling all content in English language that's nine months from now if we get this thing up and going and everything's ready to go we just need some rocket fuel so we can be living in a world starting a year out from now where we are demonstrating to ourselves that we had this capacity for the last three decades that we haven't used to be super intelligent in some cases beyond the capacity of AI and then make that AI into a tool that we can really control and revise and improve and keep on a leash and keep it safe because it's there for us in common it's not there for a single individual and which always is going to be redundant benefit of the individual who's already most powerful we don't have to build AI for billionaires and kings and autocrats we can actually build a social AI that's there for a broader or pluralistic service. I feel like this vision liberates us from having to sick King Kong and Godzilla against mecha Godzilla but then now it actually sounds like that's exactly what you're suggesting but they won they did it they did it and then with mecha Godzilla neutralized then they can choose to go their separate ways or in a nice way where like Godzilla and King Kong like gently parent mecha Godzilla and train it like how to how to be a little more open a little more caring to himself and others and I'm gonna get too far. Can I get Jim Rutt's GPT4 script helper to write the fanfic that will have runway ML make next year as a user-generated feature that we can this is Kong hearts Godzilla. That sounds like a silly art project that I could fully get behind.

Yeah we need we need new sacred texts right? I'm gonna count this conversation as one of them thank you so much for being on the show. It's been a total delight as I hope you know I need a product before I really love the way your mind works and it's always fun sharing mind space. Likewise sir thanks again for listening you can join follow-up conversation on this episode and its themes in the future fossils discord server check the show notes for a link and join the fun over the next few weeks we have some really exciting episodes in queue including a conversation with Psychonauts David J Brown and returning guest and dear friend Sarah Huntley as well as a very fun playful talk with Jing Laiokaner the field museum's curator for vertebrate paleontology I'm also gonna decant a really interesting and subversive panel discussion I had last year on the creative misuse of technology for the next museum in Amsterdam and much more.

Once again go find the show notes for this episode on sub-stack or patreon I hope you have a most wonderful aeon.