A Protocol for Thinking Together with Gordon Brander & Michael Garfield

Transcript

foreign [Music] guacamole what an episode that I've got for you here so we are talking about scaling sense making with Gordon brander and Michael Garfield Gordon brander is the founder and CEO of subconscious previously worked on web platform and browsers at Google and Mozilla Michael Garfield is a digital media strategist at the Santa Fe Institute and the host of the complexity podcast and the future fossils podcasts I wanted to have them on the podcast to talk about a sub stack post that Gordon wrote and it's entitled thinking together steps to an Ecology of mind so basically what we're talking about in this episode is this idea that

society's hit and information information scaling threshold and then they start to stall out as they reach this sort of collective Neurosis caused by the complexity of their information environment and the inability to rely on institutions for sense making this sounds familiar to you maybe it's because you've spent time on social media and you know that our institutions for sense you're going through Rapid change in the 21st century and so I wanted to talk to Gordon and Michael about this opportunities and also challenge space and what it would look like for us to move our information environment towards the planetary challenges that we're facing things like climate change Global pandemics mass extinctions geopolitical tensions things

like that what are the ways in which we can increase our sense making and get away from a collapse trajectory for society and more towards an Awakening trajectory or a sustainable regenerative trajectory for society and so Gordon's really a beautiful thinker on on this and I know that Michael has informed a lot of his points on as our problems get more complex our ability to meaningfully coordinate breaks down and what are the ways in which we could start to address that so in the second half of the episode we talk about how can we scale our thinking can we think with the second brain like an AI assisted brain we think together at dunbar's scale do we think together with an entire network

of humans and what would have to be true to build Technologies in order to sense make better in the 21st century so I don't know if you can hear in my voice but I'm just buzzing about this conversation with Gordon and Michael I think it's one of the most deep and fascinating conversations but also maybe there's a little bit of complexity there that uh I found myself Googling stuff through the episode so so maybe you will as well but one of my favorite episodes that we've done so far I think that you're going to enjoy this conversation on scaling sense making with Gordon and Michael enjoy coordination Athena Dao is addressing the underfunded areas of women's health such as menopause and endometriosis Athena is a Community Driven Dao giving women a say in decision making and funding the

scientific breakthroughs that impact their lives with Cutting Edge technology like molecules IP nft framework they invest in research IP that ensures exclusive licensing rights Athena Dao is Raising awareness to women's health research working alongside existing organizations so join Athena Dao today to help change the landscape of women's health research hey Gordon hey Michael how you doing doing well yeah really excited to get into this piece that you wrote Gordon thinking together steps to an Ecology of mind so can you tell us uh what was in that piece and what was going through your mind as you're writing this this piece yeah maybe to situate this a little bit uh I'm founder of a sort of decentralized tech startup subconscious we're building this

protocol for thinking together I guess you could sort of qualify it as like what if notion or wrong but a protocol um so that piece is my attempt to kind of situate what we're doing in a larger sort of picture of the world um and for me this all started with this really interesting paper scale and information processing thresholds in Holocene social Evolution so this is a real mouthful but yeah we're gonna have to unpack that I think yeah so so here's the back story uh there were these researchers who were are coming through this database called seshot databank and this is basically a huge collection of data about historical civilizations dating all the way back I believe to the

last ice age and as they were coming through this huge pile of data they started to notice this disconcerting pattern which is that uh civilizations in this data set tended to scale until they sort of saturated their ability to make sense together they were basically as they scaled overwhelmed by the information environment that they created for themselves um so I sort of think of this as like you know we have to make sense together we have these kind of civilizational CPUs uh basically ways of figuring things out like tradition education markets democracy bureaucracy and and science right and every civilization has their own sort of collection of these ways of making sense

um each of them kind of has a capacity to make a certain amount of sense and um and what we saw in this paper with the author saw in this paper was that you sort of scale for a while until you kind of saturate your set of of tools for figuring things out together and then you get stuck so uh then the authors pause it something really interesting which is that uh there are a number of civilizations that seem to just blink out in the data set just disappear and they actually posit that um maybe what's driving these collapses we don't understand is is a sort of uh missed opportunity to invent new ways to make

sense together so you sort of hit that plateau and reinvent yourself uh or you don't um so I read that and I took a step back and I was thinking like I think we've hit a civilizational scaling threshold you know I think the internet has really massively scaled up the complexity of our information environment uh but we haven't really figured out yet how to how to make sense in this new environment so that's sort of the launching off point for the piece beautiful Michael anything to add yeah I mean this is I I talk about the stuff with Gordon all the time you know or whenever we're lucky enough to speak with one another because this was a theme that came up again and again and again in the conversations I was hosting

for complexity podcasts at SFI and my own inquiry into this really started with a talk that Simon De Dayo gave at SFI in I think January of 2019 where he took a process uh from the digital Humanities called atopic modeling you know where where they they tag large corpora with a machine learning technique and find out sort of statistical relationships between uh machine generated topic clouds in historic texts and so he took the proceedings of the Royal Society which is the oldest scientific journal and

examined how different constellations of topics evolved over the 350 years that that journal has been around and what he found is that there is a kind of a a stock market curve in the confidence of scientists uh that science is going to arrive at a unified understanding of all knowledge over time that about in a cycle of about 150 years that our our Collective faith in conciliency or you know or unification goes up for about a century and then there's a profound collapse in in the the sentiment that we're going to actually be able to unify

knowledge that lasts about 50 years and then there's another major scientific breakthrough that grants us an ability to unify this knowledge and we we pursue unification again for another Century until the amount of information generated by those new techniques overwhelms us again and then we find ourselves back in an age of pluralism so like you know you and I met at the plurality Institute conference in in January and I think you know it was it was great to meet in that setting because what I see uh you know Glenn Weil and and vitalik and Puja olhaver and that piece that you know that uh on soul-bound non-transferable tokens and

like you know how to you know quadratic voting and you know these these are some of the most exciting uh developments coming for me out of the web 3 space precisely because they seem to be in my estimation the most promising Innovation that's going to allow us to start to build back together functional social epistemology in an era where the profusion of of new and this what what my my mentor and my friend Richard Doyle calls the info Quake generated by digital Technologies uh we can actually start to to organize that in in some way that is more effective than the the processes that were put in place after the you know the

30 Years War and the the social epistemic crisis generated by the printing press so like if you look back to you know the uh the Protestant Reformation and and the emergence of modern universities and and modern corporations and modern journalism we're in a very similar kind of moment historically and so yeah like you know this is the conversation Gordon and I are in all the time is like how do how do the kinds of tools that he's building and the kind of tools that the the folks in the plurality Institute seen are building start to help us patch new forms of collective intelligence together that keep us from Vanishing into history like those those

communities that just disappear in the seshot database yeah I think this is this is key too because we've most of our most of our civilizational CPUs our ability to sort of make sense together the tools we have date back like to the 18th century to Michael's Point like these are the tools that sort of I guess allowed societies to scale to sort of State levels of complexity um but they're just getting ddosed by the internet right like we can sort of see like people spinning fractally off into like algorithmic reality bubbles and just weird selection pressure for for memes that are kind of like social scissors and uh you know and disinformation and and in general it

just seems like the complexity of the environment is like at least an order of magnitude uh too large for for those institutions I want to pinch and zoom on something you said uh social scissors uh scissor statements is something that balls used to be Rawson talked about when he was on the green pill podcast and basically the idea is that if you can come up with a statement that is maximally divisive a scissor statement then that's what will be upregulated in today's social media environment because the social media environment is maximizing for engagement because these are for-profit companies that want to keep you on site longer because you're more profitable when you're outraged and so it'll operate great up regulate scissor statements in in order to keep us on site more because we're more

profitable that way disregarding the fact that we're not as healthy when we're enraged in that way and so I think that you know one way of I think looking at the culture wars right now is that we have a media environment that up regulates division because it's more profitable that way and so I think that this is just one example of what you call the information Bob Michael the the problem with dealing with sense making when there's an information abundance because of the internet not not scarcity so um yeah I mean I'm wondering if you could take us forward Gordon into the piece and and and how you think this all plays out if we don't solve for social sense making and we aren't able to keep up with reality like what are the

trajectories forward yeah maybe to build on your last Point too I think um something we see in evolutionary systems is each time there's a major transition he sort of resets The evolutionary clock and favors like really simple sort of forms of life and I think in some ways this is happening in our knowledge environment as well like we had these really complex or still have sort of they're creaking but like these really complex ways of making sense together like democracy like the scientific method and they're just kind of getting completely ddosed by this new information environment but but what does thrive is uh just really sort of like sticky scissorsy memes like basically the the information equivalent

of a virus just kind of like ripping through the landscape and replicate 80. so I you know I think what what happens if you're if you're not able to kind of evolve the next step like evolve multicellular memes like more complex thoughts than than these sort of basic memes is you end up with uh kind of uh social disintegration like the complexity of the society can't be sustained so you end up sort of breaking down into into smaller kind of units but I you know in some ways you could say hey this is just kind of evolutionary pressure doing its work like more communities means more strategies being tried in the new environment but I kind of take a look around and I'm thinking

like this is sort of an unfortunate moment to be hitting the information scaling threshold like we have some some planetary challenges right like yeah climate uh kind of a big deal um we've got like Global pandemics we're still sort of like struggling our way out of one of those mass extinctions geopolitical tensions Financial crises maybe looming nuclear threats I hope not but like these are sort of planetary skill challenges we need to figure out how to um think together over the network pretty quick to cope with them yeah and I was gonna yes and what your point I mean it feels like we've got these planetary scale challenges but

what web3 is up regulating right now is like Dogecoin um or nfts that are very simple and so um it gives me actually a lot of it calms me a little bit to know that when we hit these thresholds the things reset into their most simple memes that will spread like viruses and then we can add complexity and we can kind of add from there build into solving the more complex problems on top of that what is it Gauss law that any any sufficiently complex system evolve from a simpler system over time yeah exactly yeah and um you know it in a sense to I I do I think about things like Dogecoin and I'm like well you know so let's sit back and

wait because like this is the maybe a first pass on something bigger um do you I do sense that maybe growing some something more complex requires kind of getting into into the the loop of actually solving sort of fundamental allocation problem like if we think about what What markets are doing and have been doing in our society it's it's about sort of getting things where they need to go it's like running a distributed social computation to sort of figure out well you know I grow potatoes who needs potatoes you know I I make software who needs software um that's actually that's a pretty mathematically difficult kind of thing to figure out in a market kind of mostly if you squint does it you know with some problems here and there at the at the

extreme ends but you know I think we we sort of maybe saw some of the beginnings of this with algorithmic trading and and now with crypto like when you layer a network on top of the market mechanisms that we built up over the last few hundred years again there's like this impedance mismatch so this is just like the network moves much faster and the spikes are much larger and you get funny things that happen like you know Bank runs on Silicon Valley Bank that happen nearly overnight because of the the speed at which a mean can spread on Twitter is just much much faster than the speed at which like let's say a financial um system Financial Regulators can can cope with it yeah and you know just to stack on that I think

this what Gordon's calling the impedance mismatch is something that for me became really really obvious at the beginning of the covid pandemic right because or or even before that when Hurricane Katrina hit New Orleans and people were left stranded in a disaster area to which uh you know the help that they were expecting from FEMA just was not coming and then the the sort of federal protocol for how to address this ended up directly interfering with people's ability at the local scale to respond to the disaster with mutual Aid networks you know where people were being locked into their homes and arrested if they tried to come out and actually help

their neighbors and you know so uh when you know episode 139 of future fossils was a monologue in which I got into the the mismatch between the latencies at which this kind of uh you know sense making and strategic decision making uh is uh available to us through these uh Global Systems or even you know Nation scale systems uh which have evolved due to the efficiencies of an economy of scale right so like there are certain ways that you know just in time Supply chains uh are very very handy but then when it comes to you know uh they end up

undermining a lot of the resilience or or even anti-fragility that's afforded to us through uh you know diverse cultures and diverse strategies from you know a kind of a patchwork of more local scale organizations and so you know we we're at a point now where the problems generated by like the externalities that are a natural consequence of these economies of a scale you know um the the way that these systems quote unquote course grain or you know address things at a very low resolution uh and so a lot of stuff gets missed I

mean this is this was like the I'm writing this book right now about the lessons from Jurassic Park and and the legacy of chaos and complexity theory in in the 21st century and one of the things that you know that Jurassic Park popularized this notion of the the butterfly effect which is not really about the uh the influence of small-scale perturbations on macro scale phenomena it's it's about the inability of of us to model the effects of stuff that's falling below the layer of of resolution in in our models right it's not that the butterfly is causing the hurricane all the way across the world it's that we can't predict the Hurricane by creating a model that this

coarse grained enough that it misses the influence the butterfly and so there's a there's a great piece that I'm citing in in the discussions around this book Jurassic worlding by tink's org uh tanks.org.wordpress.com called Welcome to Jurassic Park which is all about how Jurassic Park can be read as a kind of Prophecy of supply chain cascading failures in covid you know and so I mean this is just to like make the stakes clear in why a a more Mesa scale organization and you know restoring this sort of you know the relocalization of sense making and decision making through

Technologies like uh you know Bitcoin and and like subconscious matters in the world is because the the the larger systems that we now depend on as a global civilization are creating monsters that cannot be addressed at the scale of top-down decision making yeah so increasing resolution is good is my reductionist way of uh yeah of of repeating your point although it's really like there's a certain irony in making a reductionist point about increasing the resolution uh but we'll save that for another time well we're getting we're getting towards the the second half of the the episode so maybe I'll pass it back to you uh Gordon and um I think that you have three trends

that you see as as finding a path forward and and I'd love to hand the mic to you to to take us through those and and uh to guide the conversation forward yeah that's right so there's this Insight from cybernetics that if you find yourself in a complex environment and you want to have agency in that environment uh you basically need to kind of be as complex on the inside as the world is on the outside um so this is called Ashby's law of requisite variety he says only variety can absorb variety or if you want to make sense of a complex world you have to have a model that's equally complex so you know I I would say hitting the information scaling threshold is is our society no longer having a model that is complex enough to cope with the

environment I think we see this reflected in in social media in sort of broad kind of spread of conspiracy theory which is kind of like a broken epistemology it's like the inability to navigate your environment and and make a coherent model so the first Trend that I think I see that's maybe kind of a reaction to this environment you know people just trying to keep their heads above water is this rise of uh of what I uh what what people call tools for thought so tool for thought this kind of I guess you could consider it it's like a hey it's a fancy notebook uh but really it's it's tools like notion tools like Rome tools like logsec um we sort of saw a Renaissance of these tools a couple of years ago and um

people basically building huge kind of personal libraries of information that they found on the web um making sense of of the world around them through these tools I think what these tools do is they let you level up your understanding they kind of let you drink more of the fire hose and actually make useful sense of it um so you know I think it's maybe not an accident too that the last time we had a huge social transition like around the um the enlightenment era um which is sort of early early era of maybe industrialization uh social change upheaval uh there was a huge craze for what is called like a commonplace book

which is basically just like a diary right but people like write everything down and there's some really like funny Twitter accounts that like replay some of these commonplace books like a window into that time but it's sort of same thing you have to externalize your brain because the information environment's like too complex to shove into one head right so you sort of you create a second brain um and and there are great tools for this um but I think you know there's maybe a limit to what you can do at the individual scale like you might be able to kind of keep your head above water start to make a little bit of sense of your surrounding environment that way but you know I I start to ask myself in what ways might we sort of start to

rebuild the ability to think together actually and um this this brought me to Robin dunbar's work um Michael I I'm sure you've you've probably got a bunch of footnotes on Robin Dunbar who's done a bunch of interesting anthropological work but uh one of his more famous Notions is the idea of dunbar's number so this is the idea that the human brain is able to basically keep track of about 150 people um and and what's interesting is that within that scale of 150 or less we tend to actually kind of fall into cooperating really effectively without thinking about it because this is actually what we did for most of our

history like hundreds of thousands of years on Savannahs or Tundras hunting mammoths you know in small groups like just and dancing around campfires all that kind of stuff so I look at the internet and I'm like Hey we're we're seems like we're doing a pretty ineffective job of thinking at scale um we start to see you know sort of in keeping with this notion that evolutionary Landscapes sort of start simple first and then work their way up we're starting to see people think at an individual level with these tools for thought but you know maybe actually the next step is thinking together at kind of Dunbar scales and um you know I I kind of think the rise of

the group chat the rise of the Cozy web is maybe a reflection of this like the rise of the little like sub 100 person Discord of people making sense together um and this doesn't necessarily even need to mean like like arriving at facts right or or pursuing science or anything like that it just means like literally kind of vibing together and figuring out a way of life that works for you and your group um so I I see like this kind of bottom-up exit from global social media and toward these smaller forms of sense making is maybe like a hopeful signal um the back half of my article is kind of exploring something that I think is is unsolved yeah as of yet and that's

how do we actually get Beyond thinking at an individual scale thinking at a group scale to start you know thinking across the internet thinking over the network at larger scales and maybe even planetary skills um I don't think anyone has an answer yet I think our our project is sort of pulling on some of these threads um but I do see two maybe maybe hints here um and one is this notion of of an aggregor I'm I'm kind of one actually maybe have has this come up before with one of your previous guests we haven't gone into Edgar goros yet let me summarize what you what you

said and then we'll get integrated yeah Edgar gross how's that sound so um basically uh there's a challenge with How Much Information Society can process we've had an information bomb in which we can't make sense of all the information and you're proposing three modalities in order to solve the complexity of the problems thinking with a second brain thinking at Dunbar number scale and then thinking within the network which I just kind of think of as if you look at dunbar's number of uh it's like the the with dunbar's number is a proposed cognitive limit to the number of people that we can have relationships with with um you know starting with myself which is one person to my immediate kin which

is five people all the way up to 150 people which is I think the game theoretic limit that is dunbar's number you're closing scaling by thinking with the second brain for me and then thinking together at Dunbar scale with the relationships that are closest to me and then outside of dunbar's number thinking within the network right and so scaling sense making within those Ways by let's not get into the means but like those are the directions that you're proposing and I think that brings us to sort of like what is an aggregor at that point yeah that's right and so I want to also note the leap from personal and small sort of Dunbar skill group to something larger there's like a big Chasm there because the first two we know how to do like instinctually as

part of our our evolutionary makeup as social creatures um but the the last one it's like actually most of our problems this information scaling threshold that's caused by our previous sort of like bolt on social Technologies for dealing with the information environment failing and so now we have to invent some new ones so like what does this look like I I don't know but I see hints um and this notion of an aggregorithm one so so an aggregor is kind of like a meme creature it's we were talking earlier about this this notion of scissors memes like memes that sort of create a trench warfare right between two groups and it's just like so like you you say something and to one half it's like infuriating and to

the other half it's like galvanizing and then they're just like lobbing things back and forth and the the meme is getting replicated because some people are quote tweeting it to dunk on it and others are replicating it because they love it um if you are a virus like that's who you want to be that's where you want to be right because you're getting replicated by people who hate you and you're getting replicated by by people who love you and if we just think about information as a kind of almost like an evolving creature right like just setting aside people's actual beliefs or what they think they're doing um like the the memes that sort of are able to create those conditions are going to out-compete the mutes that don't

um in many cases so the notion of an Edgar Gore is like a kind almost like a collective behavior that emerges from the structure of the network and the memes that are being sent around so you kind of you you repeat an idea often enough and the network will start to behave in a certain way uh you mentioned certain words on Twitter and you're gonna get like people swooping in to kind of like dunk on you or like you know high five you it's almost like this Collective behavior is is like a character itself it's so this notion of an aggregoric kind of Mike you actually might know this is drawn from the occult

is it not yeah yeah it's it's uh the you know Western esoteric magical Traditions you know that that came out of this I mean if you look at like uh even like Christianity with you know we're three or more Gather in my name there I am this is a an articulation of something like for me this is the most interesting question that comes out of uh or or this is something I've been pursuing since I discovered complex system science in 2005 because the math that governs the evolution of syntactic language is generalizable to the math that explains the origins of multicellularity and you know complex social organisms and you know the human technology relationships

it's every major evolutionary transition is governed by the kind of drivers that you're you're discussing here and you know what I think the most kind of optimistic thing about what we're living through and the way that it's changing the the frame through which we understand all of this is that we're no longer seeing ourselves as you know discrete individuals but we're seeing ourselves as members of a kind of colonial organism and that's a natural adaptive response to this this thing it's you know so you know just as a as a point to your uh your uh remarks on the hundred person discard Discord server right there was a great piece that came out a

couple years ago from the UC Santa Barbara team that included my buddy Albert cow a former SFI postdoc on how in response to mass capture phishing fish Oceanic fish schools are actually getting smaller because they're no longer desirable targets from these huge net capture ocean fishing operations and so I think that you know that uh you know Discord and mighty networks and these other kind of more bespoke uh social networks online are kind of like these small fish schools that are evolving in an Adaptive response to uh you know the the emerging Universal surveillance architecture and the attention economy

and the way that it's preying on us you know so I mean it has a lot to do uh with more than this issue of of uh sense making it scale and a lot to do with you know the the sort of balance again so how do we how do we backfill the kinds of social organization that we have to to include you know something that's not only uh able to respond rapidly you know it turns quickly in response to a threat but uh but you know all this other stuff and you know and so just to get a little trippy about it you know this is a big question that I've been exploring with my friend the tea fairy and some other members of the the burner sort of psychedelic immunity

which is you know how can the experiences that people are having on on psychedelics of this sort of uh Collective you know this we this like a phenomenological direct experience of participating in an egregor right how does and then and then how does this uh how can we how can we start thinking about using psychedelics not only as the objects of psychedelics research but as instruments in coming up with new questions that we can ask and and new uh experimental designs that allow us to inquire into things that can't be really understood or inquiries that can be addressed at the level of an individual human intelligence and that's what your

tools are are doing for us and you know I think of the internet basically as a psychedelic substance that's creating the kind of problems that you have when you realize that you know you are uh you know just one corner of this this emerging weedness that you know what William gives to say consensual hallucination um I you know so eager voice we got these like we got these sort of this concept that was borrowed from you know sort of esoteric Traditions to talk about this weird phenomenon that emerges on very large digital networks of emergent behavior that that repeats and almost has a kind of personality like these kind of mimetic entities like why

is this important why like to to sense making I like I look at this and I'm like well an egregor is in a weird way a large number of people uh indirectly cooperating together to produce Behavior like I would say most of the Intercourse that we've managed to bring about on Twitter or Facebook are are not are not benevolent like they're not really very healthy agrigors but I think there's this open question like can we actually tame our agricors or maybe co-evolve with them or or like work with them uh in order to produce something meaningful like some sort of meaningful uh Behavior over the network that's coordinated I suspect uh for that to work we're gonna need to probably

incentivize things other than engagement but there there is there's I think it may be another interesting thread to pull on here a sort of second hint at like what it might mean to to think together at large scale and that's um what's happening in Ai and llms like if you if you squint you could sort of look at a large language model which is trained over basically the entire Corpus of the internet and then some as almost like a crystallized egregor or like a crystallized colony of agrigors like somebody took a snapshot of that whole internet with all of its weird interconnections and that latent structure of communication and you're able to kind of like summon aspects of that out through a prompt right uh this is interesting it's like it's as if

we've we've taken a snapshot of our collective intelligence and and provoked it into speaking back um I don't know what this means yet but I think it's important you know I think a lot of people who've been indexing on the the challenges that llms bring about uh with regard to disinformation sort of mass generation of spam uh but I kind of look at it and I'm like wow we've we've sort of we've captured the human Zeitgeist and we've we've taught computers to dream like this is like machine intuition artificial intuition and uh and and I wonder in what ways might we actually think more effectively uh with these sort of crystallized

aggregoris than than without them I kind of Wonder you know sort of following Marshall mcluhan the answer is inside the problem not outside it um this is part of our our project at subconscious like part of my hypothesis here is like you actually want to think with these llms who are building a tool for Thought powered by AI and also by a distributed Knowledge Graph so this distributed Knowledge Graph is you could think of it as a massively multiplayer um almost like a roam graph but but but like as a distributed protocol with many many people collectively contributing to it uh one of the things I think we're going to need to to do to get to you know sort of better memes multicellular memes is um is apply like uh selection

pressure to new information that actually promotes like sort of the the emergence of more complex thought um funny enough I actually think this is something that early internet did really well like Google had this notion of page Rank and um you know pagerank the basic ideas uh the more things that link into a piece of content the the higher it gets ranked in the search um and this is sort of a transitive property now uh the funny thing that this does is it sort of selects for the evolution of almost like little memetic genes like little little things that are worth linking to and then the act of linking to them you're sort of like

stitching together with links like many other genes and creating almost like a multicellular concept so you know what we're building with subconscious is going to be basically using AI to help you dream and sort of game mechanics to help you fold your thoughts together with the thoughts of others you can follow others and every followers and you synapse in your second brain um so there's a kind of like I hope a gradient from Thinking by yourself to thinking in small groups with the people you follow uh sort of thinking at scale with these these AI agricors very cool I'm going to try to play back some of this just to say it back to the audience and

hopefully ladder up their understanding so um in a lot of ways I think of what we're talking about here as there's a duality between individualism and collectivism I am a person who has an understanding of this conversation but as a collective we've got multiple perspectives on this conversation um and thinking with the network as an individualist-centric way of leveraging collective intelligence um and that aggregor which is a shared thought form is is almost like a collective oriented way of thinking about what we see within the network so um so thinking within a network and in aggregor are both different ways of saying that we have to figure out a way

to internalize the complexity and to sense make about our information environment and understanding that we're in an environment right now where those who replicate the most information are the ones who are going to win The evolutionary game and we're in this place right now where social media optimizes for scissor statements it it uh it creates uh it creates division it creates outrage and and that's how we propagate the memes and and I think that like Nathan Schneider always talks about how social media right now can only destruct power it can't construct power and this is why you get things like cancel culture but you don't get like sense making of like to the level of like Society could depend upon and so you're proposing Gordon that AI could be

a way for you to manage complexity to create an aggregor that uh has your best interests at heart but also can manage that complexity AI that helps you dream and sense make and that's what you're working on at subconscious so that's my way of playing it back is there anything I missed or got wrong there that's about right so I broadly think of this as you know in in design science literature you have this concept of the the Diamond model like basically creative work being sort of you have Divergent phases where you're kind of increasing the rate of mutation creating new ideas and then in a sort of convergent phase like the other side of the diamond kind of narrow down your winnowing your polishing um most kind of creative processes that produce useful outputs end up kind of

broadly falling into something like that if you can't squint social media it's like purely Divergent right like it basically there's all these pressures to kind of like fractively fragment off into into little kind of filter Bubbles and it's sort of a fork only information environment like if we think of this as like a git metaphor right uh you're basically you're always forking someone else's statement but you're never merging there's no incentive to merge you're always quoting to dunk but you're never really quoting to stitch together uh like a larger work of understanding uh like in scientific literature they have they have quote too they're called citation and yet somehow the you know

there are selection pressures like peer review that cause like knowledge to emerge from that process and and on social media you know our selection pressures are like liking retweeting you know doing the little angry eyebrow face on Facebook like somehow these mechanisms are not producing you know Common understanding or coordination or or they're not even really producing like uh like um superstitions that we can agree on that would help us coordinate right they're just sort of producing a lot of splitting and anger so you know I I think I'm not going to claim to have all the solutions here but I think we we need to balance these sort of generative

uh uh Divergent forces with some other mechanisms that are convergent and cause us to actually not just Fork but merge together to produce something uh of worth together um I I probably I see like maybe only two examples of this on the internet where it's going well like GitHub is one Wikipedia is another um I don't think that we've exhausted the design space of possible ways that we can make sense together on the internet yeah we got to pull you down the Dow rabbit hole at some point yes right yes there's lots of people trying yeah yeah actually that that's a great sort of like maybe third or like 2.5 kind of like people are are are we're

rediscovering uh Democratic mechanisms across the network and and we're sort of I think speedrunning a lot of the history of democracy too which is kind of fun to watch right like we have to like re-experience it so that we understand why these things were put in in place in the first place in in the democracies that we inherited so like I see that as a broadly like really positive kind of uh happening well we got to get got to get to our Federalist Papers stage of democracy in which uh I think we create a um Michael I want to pass it over to you and hear what's how this is resonating with you and and what comes up for you oh yeah absolutely so uh I recently had my old friend Kimberly Dill on future fossils and she's an

environmental philosopher in California and she reminded me that there's this great most people are familiar with this this idea in some form or another you know the blind people feeling up an elephant you know somebody thinks it's a snake somebody thinks it's a tree trunk this is the jainous principle of anacondavara which is you know that that we need multiple perspectives in order to understand things and this is you know just to Anchor this in some more web 3 stuff this is one of the things that always struck me as inherently problematic with uh Universal consensus model for distributed ledgers right because what you end up with in this sort of original formulation of of

Bitcoin and these other Universal consensus protocols is exactly what Gordon just said which is a system in which it doesn't actually look like science it doesn't look like the bottom-up construction of knowledge it looks like the top-down enforcement of a reality tunnel whereby if you just agree then we're then we're just going to run off down our own little tube into an Ever more practice you know fractured uh you know population of incompatible forks and so you know one of the more promising uh protocols or projects I've seen coming out of the web 3 space is a hola chain which I don't know how many of your listeners are familiar with with

holochain but you know it's based on uh something that's you know a much more sort of mature post-modern knowledge construction philosophy which is one of locally generated bottom-up knowledge that is then permissioned voluntarily up into you know an Ever Ever more comprehensive structures so you know knowledge and and I basically it's like their security protocol is one that is is very lightweight and minimally energy intensive because now it doesn't require consensus on the part of the entire network in order to sign a transaction you know and so you end up with something that looks in in practice

considerably more like the way that third person verification of uh like the of the results of a scientific experiment can then be you know replicate more and more and more and so this is this is an interesting thing about you know some of the more uh confounding puzzles presented to science right now with like the replication crisis are because there is something in the modern formulation of scientific research that's attempting this view from nowhere where we we pretend as though the the situated knowledge of the people running the experiment doesn't exist that this is this is purely objective which is of course nonsense because there are all these factors that

that we have left out of the design of our experiments you know my favorite being uh you know this question of why you you can run a Horticultural experiments uh at different times and get very different results and and it turns out that it has to you know it has to do with the phase of the moon and tidal forces pulling on these plants that you're growing in a basement somewhere but if you're not thinking about all of the stuff outside of the context like basically you can't ever reproduce an experiment perfectly because there's always something Beyond you know you're not dealing with closed systems and so you know any kind of uh scientific pair word I am that addresses the world through this sort of fiction

of closed systems is going to end up with replication problems and this is the same thing that we end up with you know problems in in governance that attempts to solve things from this sort of you know uh omniscient view from nowhere you know it's just nonsense and so you know again like I think uh you know how how we balance these you know these attempts to get everybody on the same page with the reality that that no truth claim is purely objective uh you know how do we how do we instantiate this understanding in the technological platforms moving forward is is really interesting the draft off of Michael's thoughts on on sort of this uh notion of plurality epistemic plurality uh this is

something we wrestled with a lot in the early stages of doing protocol work for um subconscious so we we developed a kind of distributed knowledge graph protocol called new sphere and I guess you could think of it as like a sort of massive shared Rhone graph made up of smaller public and private graphs that are signed with cryptographic keys um so in in new sphere like the sort of Base protocol uh there is no sense of like absolute truth actually like every kind of concept is rooted to some key which is self-sovereign it belongs to some user um so it's not uh this is what uh the

concept of git means it's like this is what the concept of git means to Gordon or like this is not like it's not like this is what the the concept of uh money means absolutely it's like this is what the concept of money means to Gordon um actually the way this is reflected in the in the protocol is you sort of are always saying at Gordon slash something at Gordon money at Gordon cats it it kind of reminds me of how like there's this idea in physics that everything's interrelated and there is no absolute sort of state it just depends on the perspective of the Observer and especially when you get into general relativity uh this is kind of like general relativity but for our information space that it you need an

observer there's no absolute knowledge yeah exactly actually this is an interesting tension in distributed systems uh one of the things about decentralized systems is you can't really appeal to a source of Truth uh so you actually you have two options if you want to do some sort of source of Truth you can create some centralized uh repository that is controlled by an authority or you can use a blockchain um so if you want to you know sort of defer doing either one of those things you are stuck in the world of relative statements which is actually kind of an interesting world to be for a little while like our our protocol is sort of of pure distributed protocol built on

top of ipfs and kind of Embraces that limitation you can actually layer kind of um I guess absolute truth claims on top if you want using a chain but we actually thought it was kind of interesting to try to solve the problem of like bottom-up shared meaning and and just sort of rooting claims in in a community or in an individual as represented by like a secure cryptographic key um so I guess broadly you could think of it as like uh you know in crypto you have wallets and you have your key pair in new sphere there's you have your graph and that also is represented by keeper private key allows you to write to it public key allows you to address

it it's an interesting Construction um Michael anything to add not really I I mean I feel like you know again just moving forward uh you know I have a lot of uh hope for what y'all are working on respectively you know because I think I guess the last thing I can say is that I have found generally speaking to Gordon's earlier comments on large language models I found them in a way moving against the current uh of you know what's required here and I I don't know if there's actually any substance to this but apparently a couple weeks ago there was an ostensibly leaked

Google internal memo on how you know the The Clash of the Titans between Google and Microsoft was getting lapped by uh open source language model development on much much leaner smaller models that could be run on a phone or something like this and I see this there's a great piece that's inspiring the book I'm writing by wjt Mitchell who's a media theorist at the University of Chicago it was a follow-up to Walter Benjamin's essay on the work of art in the age of mechanical reproduction and Mitchell wrote this piece called the work of art in the age of biocybernetic reproduction where he looks at Jurassic Park and the shift in emphasis from the Tyrannosaurus as the sort of

principal villain of dinosaur fictions to the Velociraptors the Raptor yeah yeah which again like in Star in Star Wars you know you have the the snow Speeders versus the atats like there's something coming forward in the Zeitgeist right now that understands lean agile Nimble stuff to be out competing you know these these enormous cumbersome energy intensives and and uh you know very a ponderous systems and so you know when I look at you know again like the problem that I think we're facing with like seeding control to stuff like Barden and chat GPT is that it's allowing the the the the the mean to govern here it's it's

allowing that the you know this the average of everything that is the you know reflected in the corporate the cultural Corpus of of humankind to be you know what's giving us the answers and we all know that the average is is uh you know a path if we're relying on the average we're going to get a path to mediocrity we're relying on the lowest common denominator you know we're not really invoking uh human intelligence in its finest hour right and so you know I think the this you know this co-pilot development or you know the uh you know systems like uh like uh which is coming out of a a a pair of developers from Copenhagen that allow you to build

author-specific uh authorial brains that write in your personal style you know I think that this is this is where this stuff is head did because uh you know I'm not getting what I consider to be useful personally tailored information out of chat GPT you know it's and it's that's why it's writing sucks that's why you know that's why you can't get really really beautiful Pros out of this thing because the vast majority of writers are not that good you know here I'm I'm actually pretty hopeful I wrote a piece back in I think it was January basically hypothesizing that uh that open AI didn't have a moat uh and wasn't an aggregator which is the sort of extreme

kind of defensible locked down business model like that Facebook or social media leverages Network effect to kind of keep people locked in and my hypothesis was like well yeah they're a large costs that go into creating these models but they're kind of like one-time costs and then the marginal costs are much lower um and and really what this looks like is more like a traditional industrial Monopoly where you drop a bunch of money and then you're sort of like sit on your your factory and nobody can compete with you because you have this big Factory the the problem is that with software that doesn't really work because software is basically zero marginal cost copy once you've created this llm it's it's relatively easy to clone it and and this is exactly what we've seen happen right like we've basically seen open

source clones uh based on Llama Or on other things like just months after the the closed Source One so I think you know for good or bad we're going to end up with a I think a proliferation of Open Source models I think that this opens up some hopeful possibility for uh for like user-owned AI for AIS that are basically acting on behalf of individuals and communities instead of like top down for you know for open AI or for Microsoft or for Google um this is part of you know I I think the bet we're making or the world we're trying to build toward with subconscious is that um you know over time like everyone should have their own AI right

it should be you and your super intelligent subconscious that like riffs with you and thinks with you and kind of like scales up and multiplies your own ability to make sense and to and to think um and and just aside from what we're doing too I I think I see this as hopeful there I I saw this tweet a few a few days ago by Professor Sarah and Mary Walker who I'm pretty sure you've interviewed Michael but uh she said the G and AGI should be replaced by global because whatever intelligence emerges which surpasses our own is likely to be just that emergent as a collective property of our own intelligence and globally integrated Tech not explicitly designed by us in a lab to be General

I thought that was mind-blowing and it being basically summarizes like the threads we're pulling on and the direction we're trying to head and when you think about it in that way AGI alignment takes on A New Perspective where it's like oh this is our Global intelligence that how much is all of humanity aligned right now we've got social media the 10 stores outrage and scissors statements and so it kind of brings it back in to me into how do we create more sense making for humans before we start layering AI on top of it way easier to align AI to the needs of your neighborhood than to the you know diverse needs of everything in the biosphere well maybe a decentralized well this I I feel like I'm teasing what

might be another episode between us and the future but uh decentralized World in which we all have AIS in our pockets and how do they coordinate with each other how does that create new opportunities but also creates new problems is kind of where my mind goes yeah and I think um we're gonna need those AIS on our side to cope with the information environment that the AIS generate because if you thought the internet made things complicated just wait until like you can sort of mass produce spam and disinformation that passes the Turing test but you know it it's it's it's back to this notion of requisite variety again like the complexity of your model has to match the complexity of the environment so if AI is generating a lot of noise I I really think probably the

only way for it is for AI to be incorporated into our own personal sense making systems to be able to cope with that well um this has been a fascinating episode I think we a lot has fallen out of the idea that Society has hit an information scaling threshold and that could stall it out and we need to scale our sense making through thinking with the second brain thinking together at Dunbar scale thinking with a network and Gordon Michael I really enjoyed the conversation I'm wondering last question how can people get involved in your work how can people help you with with with your respective projects and missions maybe I'll go to you first Gordon yes uh well you can keep up with what we're doing at our sub stack which is subconscious.substack.com our project's

called subconscious the protocol is called New sphere it's all on GitHub licensed under MIT and Apache licenses open source we welcome contributors we have an active Discord um we really just we think that the tools for Thought should belong to everybody and we want to basically build a new kind of distributed content web that everyone's free to contribute to one that's more decentralized and and maybe a little more web 3 or web 2.5 flavored uh so yeah any any listeners who are interested in engaging with us there we're more than welcome yeah and I'm a big fan Gordon I love your next blog it's superb and I'm I'm glad to be in the alpha for subconscious I feel very privileged and then yeah as far as

I'm concerned uh you know Michael garfield.substep.com is where I publish future fossils I have a very large uh Facebook group and a very small and and intimate Discord server but you know having just left the Santa Fe Institute and uh you know being adrift like link on my glider above the landscape of all of this stuff you know I I not only welcome people into the conversations that I'm facilitating but invite listeners to uh you know solicit me if you're interested in in you know bringing me on in an advisory or consultant role because I'm I'm completely unclear about what it is I'm doing next other than you know continuing with this stuff and

and would love to you know sit in and and help people think through these kinds of questions in their own projects and organizations well um I will personally attest that your writing and thinking is fascinating and that goes for both of you so there'll be links in the show notes where people can check out thinking together which is the essay that this podcast was based upon subconscious which is Gordon's um sub sack and then Michael and your and your links will be in there as well so thank you both for joining me really enjoyed the conversation and the sense making that we're starting to do together so thanks again thanks Kevin you're awesome keep it up yeah likewise foreign [Music]