deterministic thinking has kind of had its day right we live in a probabilistic world and we need to have excessive cognitive humility hey I love technology I'm like incredibly Pro but I'm worried about a loss of intellectual autonomy if you transfer your moral decision-making to a collective you become like a be in a beehive and it doesn't really matter
as much what you're doing individually I look at dogmatism as brain death there was a lot of like philosophy to law pipelines what we're really interested now is philosophy to code rather than simply producing papers and having arguments we want to translate those into math into algorithms and we want to translate those into open source
software well hello everyone it's Jim o sha with yet another infinite Loops I have been looking forward to today's guest for quite some time my guest is Brendan McCord the founder of the cosmos Institute a philanthropic effort to explore the intersection of AI and philosophy but he's also a little bit of a superhero in my opinion Brendan your
your life like I I I as I was getting ready I'm like I want to make a movie about Brendan's life you've had an amazing career you were a naval officer who spent 610 days below water I mean oh my God uh you your degrees from Harvard and MIT you're now a visiting fellow at St Katherine's college at Oxford I mean wow I'm I'm very intimida but welcome thank
you thanks for having me so let's jump right into to why the cosmos Institute I love the play with Cosmos ver taxes right Cosmos is the Greek for from below and Taxas is planning from above but but why did you land on what is also one of my favorite topics the intersection of AI and philosophy yes so it kind of started for me two and a half years ago
I I remember having this moment of reflection where I just sold two AI companies and I just had the second of my two kids um the older of whom did these back here um love and um and for people who have had either those events or others like them happen although maybe there's nothing like having kids um you know that this
kind of thrusts upon you this moment of uh of reflection for me the question was how do I spend the rest of my time how do I do something that um benefits humanity and is worthy of the rest of my professional years and also so what do I model for these little kids that put me into reading because I realized that I had this very large gap between a kind
of zeal for you know public Improvement like wanting to help my fellow man um and the actual groundedness like intellectually that I had about what the principles of human progress really were so I started reading um 15 minutes a day was kind of all I could muster I had basically giving it up during my startup phase but that grew and grew and grew to
three four hours a day had a great intellectual Mentor here in Austin um who helped me with a wonderful syllabus and this allowed me to discover philosophy after doing that I uh it kind of gradually took over my life until I felt like I had to do almost nothing but read philosophy and I turned kind of back towards the work that I had done
previously in the AI Community but with a new lens I sort of said you know how's the AI Community thinking about this technology which is arguably the most important I mean maybe not more important than language but arguably let's say the end of the modern tech project how's the community thinking about it in terms of philosophy and I really didn't like what I saw so this
led to the creation of Cosmos Institute you know I'd love to talk lots more about it but it was it was essentially out of this um uh concern for where the intellectual incumbent movements were and what the quality of the kind of intellectual debate was in the I feel yeah and I I completely agree uh philosophy they make a lot of jokes
about philosophy majors and everything else but when I'm asked uh when I was still in Asset Management people would ask me what should I study and that was high on my list because it taught you not only how to think but also how to think critically how to take apart an argument how to understand and and your reading list is amazing you start with
Plato and Aristotle you go all the way through this really surprised me Murray rothbard's uh for a new Liberty uh which uh I I have read and you know to me philosophy is basically if you look at it in in the terms of your reading list when you start out with the Greeks and you and you get up to our modern philosophies it's really a question of
around power Justice Liberty uh the nature of human society and it gives us a lot of great Clues on hey this was a very popular philosophy that didn't really work out very well to wow this philosophy is working out really really well and then kind of a synthesis uh view uh I think becomes more nuanced than the All or Nothing we were talking before we began
recording and and I said you know it drives me crazy with the world out there being so deterministic in its thought I I often joke that deterministic thinking you know zero 100 yes no black white um often leads to uh hilarity or tragedy because we live in a probabilistic world but let's talk a bit about the reading list uh you've got like many of my
favorite things on there and and then why did you craft it the you did which I find really intriguing and really interesting in terms of synthesizing an overall view yeah yeah so it's a on philosophy I think philosophy is a kind of quick sodic Pursuit especially in a very utilitarian technological World um the way I see philosophy is that it
gives you it doesn't give you Godlike wisdom it doesn't give you the answers and so um it's a quest for those answers it's a quest for the truth but it will probably never yield them and so why do it well I mean you you gave a couple of really good reasons the other is that it gives you an awareness of what are the The Perennial questions what are the
what are the questions that Plato and Aristotle dealt with 2500 years ago how did they treat those questions what answers did they stumble upon and what kind of errors did they identify this you know thread of kind of constant question asking for things that don't really admit of hypothesis testing fundamentally like questions that we're
still going to be debating 10 years from now 100 years from now probably a thousand that's the that's the realm of philosophy I also think it's the only way to kind of get Liberation from the half considered truths um from the kind of more dogmatic thinking like Twitter makes for um you really brings out some of the most um dogmatic and
factionalized ideas or factional you know um Tendencies and I think philosophy is a way to kind of get across that but on the reading list where that came from was I just I asked a few people that I really admired what rabbit Hol should I go down and which one should I avoid and you must allow for some open-ended exploration like you just you read a book you see a footnote
you see a quote you go there and you pursue right there that has to be part of the journey but I think there's a lot to be gained from from downloading as well the kind of experience of others so part of it was crafted in that way I mostly did a chronological path starting with the Ancients who thought very profoundly about these questions they
had very genuine profound Reflections in part because they didn't inherit a lot of the you know complexity we have today or the assumptions we hold and then I moved forward to the Proto Enlightenment the birth of the modern technological project I moved to kind of the the enlightenment proper and then I went to the American founding I went to some of
the modern economic ideas in new institutional economics public Choice Austrian economics other other facets um to kind of then you know segue into contemporary debates and it isn't because I think philosophy is a progressive science in other words it's not that I think that you start chronologically because then you get
like the you know the baby steps followed by the sophistication I actually think there's a lot that we've forgotten about the the past um and that by going back to these kind of radically alien Minds we can get better insights on where we stand today agree completely and I I've always kind of looked at uh the philosophical uh don coyote Quest as
basically a fundamental dissatisfaction with the status quo whatever the status quo happens to be right so you know we we had uh medieval institutions and we had Guild systems and and then we had the enlightenment right and it was like whoa and then this continues and continues but it also doesn't always lead to uh improvements like for example
we compare and contrast collectivism which you have a big reading list on Marxism and other schools of thought uh vers individualism um you know I I have a point of view there obviously that I don't try to hide I think that there were some really interesting ideas for when Marx was writing at the time uh but that we have progressed
Beyond them I also have a lurking suspicion that in most of the collectivist texts or or Phil philosophical treat seat that I've read there there's a lurking zero sub mentality there's this idea that we everything is scarcity we must divide that p uh equally we must make sure that everyone gets their fair share versus the positive some ideas that I think
more individualistic free markets uh uh uh Embrace without missing all of the problems that free markets can have so again I always add that caveat but what do you think um I also I if you've read Hawkins power versus Force I often kind of bring that book in as well because he's really talking about the way most most societies were organized basically
by force human history and it was like do this at the end of a spear then gun then you know morass of regulations and and he chose I think kind of unfortunately the word power many people because we're label thinkers in many regards many people have negative connotations with that word but what he means by power is no no no influence you don't command someone else to do
something you influence them through argument through example Etc what do you think yeah so I I love reading about collectivist psychology and sometimes Jim when I say that people will say I do too like I you know and I'll realize they they're saying they're saying it positively um I uh I love reading about the the horrors of the 20th century experiments
um I never want to repeat them and they seem to occupy this sweet spot between they're so close in historical proximity yet they're so um forgotten and there's there's an idea that I like so so a couple of the books that I would recommend to you know your listeners I love wild Swan and Mal the unknown story both written by Young Chang she was the
first Western PhD earner from Ma's hermetically sealed China and she wrote these they're very long but they're wonderful well swans is about her her mother and her grandmother um so you get a really good cross-section of of time in China or longitudinal piece and then ma the unknown story is kind of like um I think the best book on um on Mao and
then um Soviet era things um Gulag archipelago Darkness at noon the these are um great great books but an idea that I particularly like that ties to the reading list is the man of system and this comes from Adam Smith it comes from the theory of moral sentiments and it's this idea that the man of system thinks that he can control Society like a machine it's a
deterministic you know as you said gym um controllable machine and this man of system often starts with good intentions but in practice he resorts to tyrannical methods um basically the the the idea is that one can be so self-satisfied that they think they control humans like they pieces on a chessboard or organ stops you know um but the problem is the
pieces have bition you know humans are not Newtonian humans move in directions that are outside of the control of those planning the system and so the man of system is confronted with a dilemma what do you do you know you start by saying well things like privacy are cool you know great but if the system doesn't work um with privacy in place well then
the rules have to be bent you know this super structure once You' set it up if the pretense is that you're going to create a system that works in a beautiful way what happens when people don't cooperate and the system doesn't work is you kill them or you imprison them or you reeducate them to some varying degree you have to manipulate
the soul uh to work to have the system work and so I guess the message here is that whatever the system claims initially um you know in in in Soviet system it's uh we're going to socialize the means of production whatever it claims initially the super structure can't control just a piece it needs insight into everything
it has to control what is read what is thought and that's where it becomes this totalizing system she president she um says that the Soviet experiment didn't work in part because um there wasn't great enough access to the mind you know in other words like the indoctrination was a was a failure um and that that's why it didn't work so this is not the only kind
of insight on collectivism there are other really juicy I think epistemological insights on why the central planning model doesn't work there are good um moral arguments but I think one of the Practical because you brought up like kind of the the lessons of experience is that the systems become totalizing the other thing I don't I don't want to get off uh Jim is you
talked about like probabilistic and deterministic I think the reading list gives really good insight to this and this is something I did not appreciate I actually I will say I spent I've spent time in government um and I've been guilty of some of the man of system behaviors uh in writing Grand plans DOD strategy uh I I was primary author of um
the chips and science act I wrote a very early and important um prototype of and um and both of these were very top down I think what the reading list shows is that this human reason this incredible superpower that we possess the thing that kind of distinguishes us from the other from the other animals among other things um that we we need to
be very careful about the claims of Reason like what reason can do David hum was excellent I always think about this phrase that his project was in in a to in in a sense to you reason to whittle down the claims of reason so in other words as compared to some of the more you know you could say ambitious I'll put it in quotes reformers mostly French
the philosophes the encyclopedist the rousos of the world dard although he was a little complicated um Hume and Ferguson and Smith and the Scottish side had a bit more of this epistemic humility where they thought that really what reason tells us is that we can't make these predictions with a high degree of accuracy um you know we can't simply start from these
abstract first principles and think that we can use deduction to arrive at a perfect society and then design it tabular rasa like that gives you the French Revolution which is a failure yeah so um another one of my favorite thinkers on this is uh Lewis Carroll who famously wrote Allison Wonderland I wrote a lot of college uh term papers
about how that was really Charles Dodson the logician trying to demonstrate to his Victorian uh neighbors and fellow citizens that there were deep limits to logic and he wrote a very short piece called what the tortoise said to Achilles in which he argues that in every logical system if you look at them as axiomatic in other words in no need
of proof right uh when you when you keep going down when you keep going down and all the way down to the base what you're going to find at the base is a assumption that has no proof that simply was a human asserting this is fruit and then that brings you to agrippa's trilemma where the uh where the uh limits of logic are kind of decimated
honestly uh um and so I think that again back deterministic vers uh probabilistic I think that as we get closer as a society to understanding that deterministic thinking has kind of had its day right we we live in a probabilistic world and pragmatically as you as you mentioned we need to have excessive cognitive humility in my opinion because we are always learning one of
the things that I wrote down and put on my screen for a while was most of my opinions are probably wrong and and the reason I did that was to remind myself you know a ay these these are not facts uh these are for the most part conjectures thesis you know all of those things let's continually test them and let's upgrade and update our thinking as
we learn more and more and more so for example uh just looking pragmatically as you say uh vers collectivism verse uh individualism and free markets objectively it's very difficult to make the argument that free markets just didn't work better right with all of their problems they really delivered on many of their promises where most of the
old philosophies didn't and I kind of think I I look at dogmatism as brain death because essentially what you're doing is just repeating something that you have absolutely no uncertainty about and to me that is catastrophic because we we live in a world that is uncertain yeah I'll um this relates quite heavily to or quite directly to the discourse around existential risk
and AI um and I think there's a lot we could talk about there I'd love to you know get into that but um I will say that you know David Deutch one of your I know intellectual uh uh Heroes uh and he he talks persuasively about this that you know really we have these um these subjective expectations about the future and they come in various
strengths and we you could say you know Jim are you sure that this is going to happen are you very sure you know I'm very very sure that sort of thing but there isn't a very good inter subjective comparison that we might do you know if I'm very sure and you're very very sure um how do we figure that out we can quantify it and we can try to kind of
like harmonize um neither of us should say that it's impossible or certain because the laws of epistemology tell us that you know if we discover something later that it would alter that that probability like the uh the zero and the one are kind of untenable positions but um the problem with existential risk among many things is that if it if if
the likelihood of some bad thing happening does fall within that zero to1 range um not inclusive then it's um and it has an infinite disutility like a very small probability times a negative Infinity utility that's still a negative Infinity number and so um it creates a situation where anything one might do in light of this takes on kind of cosmic
significance it's the most important thing possible I think you have to kind of reject that premise and you have to say actually we have we must maintain this idea of an unknowable future subject to an unlimited number of forces we have to be scenario agnostic that doesn't mean we don't con we're constructively on risk it doesn't
mean we don't Point them out and work very seriously on perceived risks whether they're the kind of Bal ordinary risks that I care a lot about or the species ending ones which you know I also do care about but I think are um I think there's an under-provisioning of the former thought so that's there's lots more on the on the kind of um risk
side but I think you mentioned the the free market which is a topic near and dear to me one one point I wanted to make here is that um before I put together the cosmos Institute the very first instinct that I had after being kind of taken over by a love of philosophy was wow there must be more people like me you know I actually
didn't write um term papers on uh Lewis Caroll and things like that it sounds like you have an interesting like academic background I was like stem in business basically and I never really had exposure to these great thinkers so I realized there are many technologists who are like me and they're building the future they're they're like the John Stewart Mills of
today but there there are like you know translating ideas into code that allocate billions of human Minds 247 I mean it's incredible uh High leverage formative position in society where where are they engaging with these big questions and so the first thing I did was I got introduced to Tyler cow and I knew right away that it was a you know
it was a brilliant fit and we put together this thing called The Mill Fellowship johnst Mill Fellowship to try to cultivate a new breed of philosopher entrepreneurs um people like Patrick Collison who's um you know leading stripe or uh Reed Hoffman you know like people who engage deeply with philosophy right in a way that you know I think Jim
you embodied this really really well um because the thought is that they're going to be much better at deliberating regarding the alternative ends of their Creations so you know we we we put that together and what we what we walk through together is what are you know what are the key ideas behind markets um what are the nature what is the nature
and the limitations of political action and what are the possibilities of Bottoms Up Solutions in society it's kind of three flavors we hit you know lock monus Mill you know those kinds of thinkers as well as Hayek on the markets and it's just really beautiful but I I'd love to talk more about the market because I think it gets um very little
love I went to Harvard Business School and I like to say now that you know it's kind of a Bastion of capitalism but at no point do we actually talk about the the promise of markets or the role of markets the only focus is on reimagining markets through things that I later you know kind of come to understand as like syndicalist you know socialist ideas
like very very interventionist kind of um uh ideas to try to fix perceived uh perceived deals yeah and uh the the takeaway I have there is that you know human OS human uh operating system uh is I think primarily an emotional one not necessarily an intellectual one and I think that the ruling human emotion is fear and that makes sense if you look at Evolution right
uh for for most of our history as a species fear was a very important thing to have especially fear of Novel threats right I I I joke a lot about how the fact that we're all descendants from the cowards who ran away if we're out on a savannah in Africa and there's a bush wrestling we're the descendant of the guys who saw that and ran as fast as
they could in the other direction and not the Curious one who said oh I wonder what this is and then he gets eaten by a saber-tooth tiger and and I think that we I do believe that we are able to demonstrate cumulative cultural Evolution so I don't believe that we are tied simply to our biological evolution um but I I do believe that we have to take them into
consideration when when talking about these ideas uh because you know they they pull us in a direction that um is is almost unavoidable uh if the emotions get involved but you you mentioned that uh you want to talk a little bit about uh the various schools of thought around AI which I I think let's do that right now I it came later in my notes but uh since
you brought it up look there's AI pessimism which is it's G to kill us all let's nuke the GPU plants and you know be done with it then we have long-termism we have effective Al Al alism and we have the accelerationist uh Bros and listen if if I have to put myself on that Continuum I'm I'm closer to the accelerationist Bros than I m to
the AI Doomer but with a big as trick which is whoa let's be moderately optimistic but always aware of the fact that we are going to create a whole new set of problems but hopefully there're going to be better problems but all all progress comes with problems and again with a nod to Deutsch we we have to understand that
we're going to get a lot wrong too so take us through your views like on on each of those AI pessimism all the way through accelerationism and and I know where you land but I want you to tell our audience where you land yeah yeah so when I first put together a cosmos Institute which um by the way tries to kind of conceive of Technology as a Triumph of human
agency so a lot of these topics we're talking about come back to that for me technology is a means of human flourishing and I I'll touch on why that's important as it flows through this um and then later we can talk about some of the things we're actually doing to kind of realize that but when I first you know put it together the first thing
I did was try to look at the landscape like any entrepreneur would do you know you're starting a new business in this case it's a nonprofit figure out where things stand and how you kind of what the white space is and I kind of group got together a group of brilliant people you know um you know AI folks as well as like forensic accounts and try to like
Trace out the money flows just because I feel I felt like there was when you're trying to understand like an a graph of uh ideas you need to you need to understand the money piece as well and so we came up with some pretty interesting conclusions um but I'm going to walk through it at a high level I think you got the categories right uh
Jim the the pessimism side in its worst form becomes kind of an apocalyptic almost eschatological like end times view of risk it's very Quantified that's what animates it um it's very popular in the practitioner circles people who are very mathy um and it's kind of like I I watched I finally got around to watching Oppenheimer it's kind of like this um
you know denouncing of of of a creation you know in a kind of fistic a of power you mentioned fear fear is actually really important in this camp and it's kind of ironic because to have an atavistic fear uh you know that we sort of uh had in our common ancestry gym when we were running away from the uh from the scary you know Predators it doesn't
it isn't really compatible with the idea that we're kind of rationalist uh almost disembodied Minds in other words the the cold rationalism uh ironically creates the the most fertile soil for this fear emotion that is uh part of our prerational uh selves so it's really I think the pessimist position has three entangled philosophies I think there's
rationalism there's long-termism there's effective alism very briefly on each of these each of them is kind of like what they say on the tin if you will so like rationalism is about being rational it's a sensibly that but what's interesting is that it's not a kind of rationality that Plato or Kant would recognize or
even modern thinkers like uh Nicholas TB it's a kind of basy and updating on a value function so there's a lot more that rationalism you know and reason entails um this is a very narrow Slice on that I don't want people to be confused that that's the only way to be rational um but it basically posits that preserving life is Paramount and precondition to I
don't know everything else fulfilling any other needs and desires so it tends towards a so-called extropia ISM which is you know kind of a defeat of death this actually comes out of bacon and dart you know it's a very long-term idea um but it's given new life in the form of yow's Harry Potter and the methods of rationality I'm sure most people know
but this is like the greatest fanfiction of Harry Potter it ends with Harry Potter defeating death this is the uh the rationalist stream the second Camp is long-termism which is very future oriented again as it as it seems It's a self-sacrificial philosophy it's something that's very focused on the existential threats existential threats can be natural they
can be artificial this tends to focus more on the artificial ones um but it's kind of like reaching into an earn of innovation and then picking out not a white ball not a gray Ball but a black ball that's going to destroy us it focuses a lot on the extreme case it's going to like if I'm teaching my kids to ride a bike and they're like just laser
focused on the pothole on the side of the road you're like what's going to happen they're probably going to drive into the pole or at least they're going to have trouble but it's a very excess focus on the extreme case so so the Bol threats the natural threats that lie before us the things that might happen in the next you know few years or or or
you know 100 years are not really a concern um for this school it also blinds us to kind of what I call the Tok villian threats which we can talk about later but I think I think teville is like an underexplored voice in this and I would I just think he's he's brilliant that's that's long-termism the third is um effective altruism which is the most
effective or sorry I'll say adopted uh academic theor since Marxism it's a very popular Theory um it is a consequentialist philosophy that has us judge any behaviors that we do that are other regarding by their fruits or you know by their results not by their intentions so um it's a kind of one of these things where I love phrases like
this because who can disagree with effective altruism it's like effective altruism these are great words um but there are a few radical premises here one is that um basically proximity and time shouldn't govern this kind of behavior so a util is a util whether it's yours or a child in you know uh Africa or a future human who's 17
Generations away from being born or even a shrimp um these are all kind of all of this can be reduced to utils and um saving Humanity in the infinite future is the most important task if you're committed to altruism I mentioned this problematic equation if you kind of take the all these things together all these premises you have you know small probability of Doom times
negative Infinity utils equals negative infinity utils and some people challenge it some people say brenon you're being too simplistic there's going to be like a heat death of the universe that will cause it to not be a negative Infinity util event or whatever I I was just like two weekends ago with like some of the top people in this field and they're
very committed to this equation like I I I think this is straight up the equation that drives it um the other difficulty with EA affective alism is it's fundamentally not individualistic you must take the perspective of the universe you know Jim you probably have some projects and plans that you think are going to you know make the world a
better place but you have to kind of surrender those to what the calculation says whereas uh Aristotle and a lot of other moral you know philosophers would say that any moral decision inherently Springs from the IND ual so it kind of alienates us from our moral choice I am not surprised and this is kind of a low blow but I'm not surprised to see moral
failings in this group in in areas like sexual assault or Integrity Integrity in business precisely because if you transfer your um moral decision-making to a collective to a mass calculation you become like a be in a beehive and it doesn't really matter as much what you know what you're doing individually so again probably will get a lot of flack
for saying that but that's that that's the risk school and what I try to do there is to always take the arguments pretty seriously I try not to say oh these are people who have like you know bad vibes or something like that I think uh the moment you start to say that a a position is just like immoral because of these kind of like ad hominum on ad
hominum basis I think you've you've lost the plot on accelerationism and kind of the other side I'll just briefly touch on that Jim which is to say that I think it gets a lot right I think I'm uh uh with you that if I have to pick um which I don't and I then I would I would sort of like lean in that direction for sure the optimism part is right I love
the deuts you know idea that we we the future is not predetermined we must make it so it can be better than the past and it's our responsibility to do it I'm all all for that I also think markets are This brilliant Discovery machine this driver of adaptation this wonderful way of using knowledge in society um I think this is one of the biggest takeaways
from my reading list is just understanding what markets do uh to be able to harness this dispersed local knowledge that we have and to kind of allow us to have you know an adaptation a mutual co-adaptation of our plans um in this like dynamical process that I mean it's just I I that's one of the most important ideas uh ever uh ever
conceived so I think that is actually right and I think of accelerationism as kind of like Classical liberalism without the moral architecture when I say that what I mean by this is really three things one is that it uh accelerationism um treats uh human life like a variable in a thermodynamic equation or like a link in an evolutionary chain it is
fundamentally non-h humanistic to say more about that ex effective accelerationists who I think are principled and not just Vibes they get accused of that I don't think that's right I think they we have to take the idea seriously they think of biology as a universal mechanical thermodynamic computer um they they think about a thermodynamic will I don't think they
say that but that's what they think about it's a thermodynamic God that animates all matter and it's this singular force that drives growth it drives expansion capitalism is the metaconsciousness and our only imperative as a species is abstract entropy production or greater consumption of energy so it's a thermodynamic will that kind of receives
metaphysical characteristics spiritual impulse and it does so at the expense of humans nothing human makes it out of the near future so that's one that's a big problem uh that we view human life in that way two is that um we mistake the end of human life uh with the means and what I mean by this is um Tech progress in the econ and economic progress is not
the sole end you know that's not the objective the purpose to turn the crank on technology should we do that hell yes we should do that you know it's amazing for uh it's like empirically amazing and it's just this wonderful Triumph as I said of human agency but the the goal of human life is a kind of flourishing of the of the individual that um technology
is causally efficacious for but not the same as so this is the second is a is a fundamental confusion um and what this leads to in in say Andre's piece is humans become this kind of like passive agent of this aoral technology trajectory we just sort of Stand By we do things we contribute to it but like it doesn't have any moral property other
than you know we should make it go um whereas if you think Tech is just a tool it opens up this like really complex space of moral deliberation political philosophy things that are not very fun for most people but are kind of essential third thing is the eak and techno Optimist Community I think are very polemical they're very cultural War
focus and they kind of um have a hatred of argument in other words if I go and I say hey I love technology I'm like incredibly Pro but I'm worried about um a loss of intellectual autonomy among people I I worry a lot about this okay if I'm worried about that I'm Blacklist I'm a I'm a Doomer for saying that basically I'm like I'm I'm against
technology and literally I have friends who have been like blacklisted on this uh because they they say that we should be responsible or whatever and so when you do that and you hate reason it descends into a form of uh myology like a hatred of argument where uh again you've lost you lost the plot the one thing as I was watching the eak group um virgin I I one of my
worries was this is beginning to look a lot more like a religion to me and um I I I'm not anti- religion right I'm I I'm agnostic because I think atheism requires as much Faith as the truest believer in a God uh for the most art um so I'm I'm just willing to say hey I don't know I I think it would be super cool if there were other forms of lives
and and and gods and and whatnot but I have no evidence of that and and so my challenge with religion at least historically is that I just looked at it as yet another and a very clever by the way command and control system to manage your Society right like the ancient Babylonians wanted to rule Samaria but they realized
they had a really puny God right and so they're like hm we got to fix this so they came up with a new story their God actually defeated this monster that was going to kill Humanity ripped it in half the half that it threw in the air became the sky the half that it threw down what became the the ground and Presto Cheno you've got like the coolest religion and
therefore you have the right to rule Rule and I just think that the reducio absurdum is at play here and I like you you know definitely lean toward the effective accelerationism but I really think that the the the the weak chain in the in the links here is as you say it is not an All or Nothing question you've got to have again we come back to humil
you've got to have intellectual humility to understand that these are very complex topics that really you're going to probably get to a better place by interacting with people who disagree with you like one of the central things I use large language models for is I always say I put in uh something I've written about a certain belief or thesis
and I say please generate a deil's Advocate position for all of the areas where I'm wrong and boy I'll tell you sometimes some of I've actually changed my opinions based on on doing that sort of thing and I think that again getting back to the cosmos Institute that's kind of what you're doing with the discussion groups no it is yeah and I want to so I
agree with you about the religious character I think it's kind of fatalistic and determined and surrenders human agency to this like impersonal Force which is is thermodynamics I think I'll just put a plugin for Tok phille here um and for people that don't know Alexis toille was this French Aristocrat basically who visits America in 1830 so
a little bit after you know the 1776 uh Fanfare andu fuff and writes about American democracy the American social State tries to understand how do we live what are our opinions how do the opinions influence how we live writes the best book on Democracy ever I mean I think I can say that and um he points out these very subtle challenges that
are not about technology it's 1830 but they merge with technology and they're very precient and I I I think there's there's a lot of kind of excavation that should happen I'll just I'll just kind of give you one because um you mentioned these impersonal forces one of the things that toil says is that when you have democracy take hold when it's thoroughly entrenched in the
social State then people stop listening to a favored class of people right in an aristocracy that's what you do you listen to The Aristocrats they instead become slavish to impersonal forces and those could be theories they could be um majority opinion right like what does everybody else think that's you can think of that as like an abstract personal force in
our case they could be algorithms but these become the dominant things and people kind of retreat within themselves um and uh and trust these these impersonal forces so I think the eaks are doing that on the thermodynamic side I think the effective acceler effective altruists are doing that on the um utilitarian side and um basically
you could have this possibility that we use AI it appears to us to be a neutral Arbiter like it kind of has a neutral homogeneous sort of standpoint and we trust it to the to an extent that it causes innervation and weakness and kind of undermines the strength of our soul that's required to form independent judgment so again lots more could be
said about this um but I think I worry about the kind of like intellectual Conformity and the loss of the energetic use of I think toille really nails that so just to kind to speak to Cosmos yeah we we thought that we thought that you know the world basically creates these moments where philosophy matters like the American founding or the French Revolution right
doesn't always work out well but but philosophy matters this is one of those types AI is one of those Technologies it is plausibly the end result of the modern tech project it's a technology that can create more technology it's a technology that can manage human life manage the economy and um I do not take that to mean that it should and I
don't I I do not take that to mean that it should in a kind of um robust way certainly but it but it can and we must you know think very carefully about about that it's also an era where we have like more possibilities opening for human agency this like space is growing and growing and growing what do we do with it it you know I think uh we we
have no IDE what we doing um and we need to you know kind of give people philosophical resources to think for themselves about how to how to live uh flourishing lives so you know I can talk you just about a couple of the things that we're doing one one of the things is that we believe that you know since we've been talking about the past um in the 18th century
there was a lot of like philosophy to law pipelines where you'd philos philosophic ideas that get translated into law policy stuff like that what we're really interested now is philosophy to code so we're interested in philosophical ideas ideas around let's say individual human agency um free speech ideas around the nature of
reason and rather than simply producing papers and having arguments about those which I think is important as well we want to translate those into math into algorithms and we want to translate those into open source software this we think is a very important vision for the future for allowing philosophy to guide the kinds
of technologies that we create as well as for making sure that as we do this you know thing we've had to do throughout history which is restate important ideas like about Freedom um you know I think about like Pericles giving his funeral oration to the mothers of Athenian dead talking about why they died for Athens well it was for the freedom of Athens
we have to restate it every generation and today it's not just about restating it in gen Z terms it's about putting it into code so that's one important project I'm in a I'm a visiting fellow at Oxford as you mentioned um I collaborate there with a guy named Philip corus who has done this in his own work he read Plato's dialogue the
Theus about the nature of knowledge it talked about what reasoning entails he formalized that in math he's putting it into open- Source python package which he thinks will be the the best psychologically realistic model of reasoning for for things like you know llms so he's done that and I work closely with him I can't talk specifically exactly about what we're
doing but it's something very big there that's one thing the second is we're trying to build a community of technologists who come from anthropic open AI de mine places like this who really want to work on these these questions around um you know classical liberal ideas reason um uh philosophic uh questions more broadly and who have the kind of like aptitude and attitude
to ship open source software basically so we we want to give them space to do it so you know people want to work on Technologies to counter digital authoritarian right I know you know Zoe Weinberg and xan and uh like wonderful like let's let's create space for creating more agentic technology people who want um make sure that individual human
agency shows up in the recommender systems great let's give them space to prototype that create an option for the world we have L we have a set of fellows that are roads and teal uh quality people who are uh uh you know as good as they get when when it comes to AI research we want to give them space to build those things and then um and then
we're doing a host of other small things like reading groups I I um you know uh put together there's something where we talk about what are the origins of the modern tech project like we bacon is an example he he looms largest he he in 1630 kind of WR wrote down some of the core ideas that led to that but we also hit you know the those that preceded him
in the ancient world of Aristotle um and some of the critics Tok phille being one haiger being another CS Lewis being another so we do that I'm teaching a class at Oxford on philosophy of AI where we're going to read things that uncover a philosophical issue and then we're going to hear from a tech entrepreneur who's Translating that into
the real world so like for example human creativity we can read n the gay science what what what do we learn from artists what do we owe to artists and then we can talk about how Tech enhances creativity right it's this kind of beautiful combination of old and new so we're doing a lot of those cool things last one I'll say and then I'll I'll
shut up is um there a a thing called the Lum project where we'll go to Aristotle's school at Athens um which is kind of a plot of land now but uh in the Athens conservatory and we'll talk about how would Aristotle's ideas influence uh technology ethics because it turns out there's more to ethics than just narrow utilitarianism and I I got involved it's
an Oxford plus Stanford thing um Oxford Institute for ethics and Ai and then Stanford High Lab and I got involved to kind of you know help practitioners go there because I think it's not just about putting in place regulation we actually have to you know be ethical and think through these issues in complex ways as Builders yeah I I'm very excited about
all those uh projects that you've got going and and what I think I'm most excited about is getting the practitioners involved I think that um as you say a lot of people trained and and or thinking through stem or or what have you you you you end up with cpos the two cultures I'm sure you've read that book and and we need to put the put
them back together again in my opinion and one way to do that is is all of the things that you are supporting and so I applaud that I I also wonder sometimes about you know getting back to human nature I I agree that high agency is I like one of the things that I admire the most because uh a very highly uh a person with high agency just literally
if you just I I love to read biographies and you don't find any low agency people in those biographies right and and I wonder uh as I was listening to all of the various uh things you've got going and the philosophers that you're citing I I I recently did a salon for inter intellect and and it was uh thrown at me by my friend and a gat were an investor
in that company too so P disclosure um and she's like hey Jim how about doing a salon on LSU and dosm and uh Ai and I'm like okay I'm game I'll do that and and one of the things that I led with which I thought that I literally just saw as I was getting ready to go on in about an hour at was this idea of wolves that were released into Yellowstone Park I
think there were 15 of them released and all of the completely unexpected things that happened as a result of that right so because the Wolves were there the deer population went down because the Wolves were hunting and eating them that led to the deer choosing different locations the locations that they vacated were allowed to thrive with the
Flora and Fauna which reintroduced new species of plants new species of trees and on and on and on and it's a there's a wonderful little film I actually put it up on Twitter and and I wonder is is there room within the series for grocking that right which is oi who would have been able to come up with all of the things that would have happened to yellow stuff and most
of them for the better by the way some for the worse uh by by just that simple sort of initial thing and the reason I'm bringing that up is it can translate immediately to the practitioners putting it into code right you can you can have an idea of what that might achieve but there are a lot of variables there that like like our wolves in
Yellowstone that are going to develop that you didn't anticipate and that's why I'm so in favor of the iterate o mistake let's change that let's pivot let's make sure that you know I fire which you know go I I know I I mean there's so much I could say about this but one one funny one is um cars helped ignite the sexual Revolution and I I I
mentioned this because how do you predict that I mean maybe somebody did I I wouldn't put it past someone to predict but like these are not not just they're not just like complicated effects but they're like orthogonal effects you know um and so I think this happens all the time and then there there is a lot of difficulty with imposing top down order on complex
systems like another one I think about is um the biodome experiment where we spent millions of dollars with the best possible science to try to create a self-contained habitat that I guess we could you know take to Mars I mean like Elon Musk should pay a lot of attention to this because we don't know and we did it and we put together this amazing um
you know kind of um experiment and then what happens well I think like insects run a muck cockroaches take over a lot of species die out plants do as well and again I can't stress enough this is the best science possible with no real limits to funding and so it's uh it's a great difficulty that we have you know the the possibilities in the future tend
to outstrip the uh the the human mind's ability to to predict them I would say in in AI one one thing that's encouraging is there are some people who like Dario amade from open or from uh was it open Ai and then founded anthropic um wrote a pretty incredible paper about scaling laws and this was Andrew had worked on this prior at BYU
but he wrote it for kind of like the modern you know uh large language model era and what it you know basically says is that if you scale a model up uh by you know scale I mean the amount of data the amount of parameters it gets better and better and it does so with smooth curves over orders of magnitude and you know maybe we have three four years left
of this maybe we have more but it's the reason why today it's $100 million to train or last year was $100 million to train this year a billion dollars to train 2027 hundred billion right it's the scaling laws that tell us this great um this is something I think prediction is good for right in other words I I'm fully supportive of saying you know next
year it's going to be that and maybe we'll get surprised like that's not 100% confident um prediction but within this narrow self-contained technical domain I think there there can be a prediction what happens in the economy what what are the effects when you release that bigger model gpt7 or whatever no one knows impossible to know and the best
system is going to be the most adaptive the one that can can undertake lots of iterations can do so with a kind of a degree of openness and preserves all the kind of adaptive feedback loop type things that the you know that the the free market and that the cosmos type meaning bottom up decentralized order sort of gives you I I I think that's um
a key a key takeaway from history philosophy and the the the failed experiments that uh that ignored this this uh this sort of fallibility yeah the uh I I love the uh example of the uh of the uh attempt to create the perfect um liable conditions uh and I I followed that very closely and given my uh kind of views on it I I anticipated that a lot of things would
happen that they didn't think about and and I'm always trying to be aware of like the unanticipated consequences of things you know you'll often hear like oh my God I hate spiders I wish there were no spiders in the world and then you realize if there were no spiders in the world we would be overrun by insects in a insignificant amount of time
because they do a lot of cleansing of the SE population but that also again like everywhere we look is a complex adaptive system I think about markets that way for example markets are complex adaptive systems that you know generally are clear because you have heterogeneous beliefs within the market system right I could sell you Nvidia uh and you could
buy it from me and we could both be right right I'm 63 and have grandchildren and maybe I want to pay for a grand child's education and and you're young and have young children and maybe you want to save for a child's education but as part of that there are failures in complex adaptive systems and just to use the metaphor of the market
when opinions in markets go from heterogenity to homogeneity Bo watch out there you get bubbles there you get crashes because information Cascades result and when that information Cascade gets so overwhelming to the human like I remember the last time I saw it in real time was the do bubble where literally everyone was thinking exactly the same
way and everyone completely ignored all of the other permutations that could possibly happen but I like the idea that that if the more we get people understanding you've really got to look at this at AI at all of the permutations of what can come from AI as a complex adaptive system and again we get back to uh intellectual humility cognitive
diversity all of the things that are going to go in to try to maximize our chances to get most of the decisions we make right we're never going to get all decisions right ever yeah I mean I love that Jim it with interesting is both the surprise and the homogeneity happen at many levels of the system so to give two examples like you're uh you made you
made a career in Quad Finance you know that um cross asset correlations hold in like kind of one regime but then when gsam golden Sach Asset Management liquidates its book or whatever there's a phase locking effect where everything kind of you know hooks to one another and it kind of goes down and that's when thek Market um functions at its uh most
uh uh terrifying um the other is that the surprises that we have are not just about oh the car unlocks the sexual Revolution it's also like at very low levels like gpus we all know were for graphics they're for gaming now they're for AI um stochastic gradient descent a key technique it was for scalability now it's for avoiding overfitting um you
know attention was for machine translation now it's for everything basically and so these are all surprises and so the question is like how do you create um or uh cultivate a system in which we can um you know accept that we must take some good some bad with the good like there will be surprises that are negative um but basically maximally harness those
those surprises and I think that's that relates to heterogeneity it certainly you know it certainly is very consistent with the market what I mean by that is that the open-source software model kind of beautifully allows for that because you put something out and then millions of potential interactions millions of experiments millions of
developers tinkering and building and each interaction represents an opportunity to kind of use something in a in a delightful way or make a new insight that um it's harder to do within a company even if the company's operating very efficiently but also I think the Su Point here is that like we would like a world that is neither fully
closed nor fully open in in in the software sense because that is itself a form of heterogenity you know like you you it allows for you to have you know incredibly planned you know organizations like whatever Amazon or something alongside these like organismic you know kind of open- source movements that don't really have any shape um and that kind kind of operate
around them like that's a beautiful form of of heterogeneity for a for a resilient Society yeah and one of the reasons why I am passionately in favor of Open Source but not absolutely opposed to closed Source being coexisting with it is what is for the reason you just cited I I passionately believe that open source gets us to better outcomes and it
gets us to better outcomes because of cognitive diversity because it's a million Minds working on something that are very diverse in the way they look at things and you know there's a reason why Linx runs the web right you know it it routes around problems and you know a top down system would in my opinion be much more brittle for a lot of the uh
challenges and problems that of the web whereas the the bottom up the complex adaptive system approach which is open source right allows for solutions that many people like even the top designers would have never thought about and EXA so I I I have no problem with coexistence I I think there must be right like if if I come up with
something super clever and I want to commercialize it and and keep the magic sauce uh to myself I I think I should be able to do that but I think I should also do as much as I can can to contribute to the open source Community because again to me the the very American idea of pragmatism right uh that seems to be a very pragmatic solution to dealing with these uh with
these very very complex problems and also just study nature study our immune system our immune system works it's a complex adaptive system and right when when when the novel Corona virus comes our immune system is like what the is going on here and then we start going over there and the cells start like trying to figure out a way in and what's
really interesting to me at least is that once some of the cells of the human immune system start making progress you know what happens all of the other cells start going towards that point which I find really interesting and because it has an analog in the idea of swarms I'm sure you've thought a lot about how uh swarms work and I think in our new
system swarms are going to be a really big thing because they are basically duplicating what what's happening with the immune system and other complex adaptive systems I like the idea of thinking of um modifications to um you know this technology landscape as being those of doctor kind of thinking about a body I I like Elanor ostrom's kind of
framing on this that it isn't about like trying to solve the top- down problem it's about like partial solutions to like specifically Define problems that we can kind of implement um and gain experience on as we go another really good thinker on this is James V Canon um I had not read James V Canon until this you know Mill Fellowship but what he
kind of says is that like look you gota like you got to lay down a set of rules let's say regarding the market right you have to lay down a set of rules he's a he's a what's called a contractarian constitutional contractarian so he lays down abstract rules as a form of social contract and um then what emerges from that is going to be spontaneous it's
going to have higher level system properties that aren't sort of like predictable from the individual Parts in a causal sense right um and uh but you don't have to just let it go you don't have to just accept wherever it goes what you can do is you can look using reason for Paro optimal you know Bargains better Bargains like you can
find little areas where you can make tweaks and when you do that you can feed those back into kind of the Constitutional rules so he has this phrase in um one of his texts that says that like um rules these rules can be used in a quote metaphorically evaluative and empirically corroborative sense and this is a mouthful of a phrase so I realize if you're hearing for the
first time it might be a lot but what he's basically saying is you lay down a set of rules you let empirics do what empirics do and then you loop back and you say the metaphorically EV valued pieces like would I have chosen those rules had I knowing what I know now would I have chosen those and then the second is empirically corroborative like
did it kind of work or did it kind of suck and so he's got this cool little like loop back that I think computation enables us to do in like super interesting ways the other thing Jim I was going to touch on is um on open source so venod kosa last week talked about how we should basically not do open source uh I mean he he did sun and
so they had some a lot of formative you know key open source stuff but he's basically like open source is too risky now I'm I'm paraphrasing because of China because of us China Dynamic and then Mark andreon responds and says well if you really think that why don't you have like actual security at open AI right like uh um because you know they
don't um I my my view on this and again I wrote I wrote the very first DOD strategy and what what was what the debate was like in the building in the Pentagon at that time was we know we're kind of we know AI is an important area for for defense we know China is making like pretty credible steps forward what do we do and whenever I would talk to de
people in general but particularly intelligence Community people they would always say a version of we need better defenses we need to focus on halting Tech transfer we need to get you know Chinese Nationals with a potential tie to the Chinese Communist party out of HED and um it was all about kind of the hunkering doubt right and if I think
about a football analogy it was like we need to boost our defense contrast that with a friend of mine was the co at at Google X um back in the day for like a while and they would have people come in and uh ransack them like try to seal their technology it's kind of wild but they would do this and rather than like making that
organization a safety a security optimized organization they were just like screw it like we're going to go faster that's how we're going to win is we're basically going to go faster now of course you want to have some security like I I get that you want to screen like I don't want a uh a pla you know people's Liberation Army Navy
person who's like you know uh embedded at Berkeley just shoveling you know research out out the door I don't want that but uh do I want uh Chinese Nationals working on important AI projects hell yes when I when I was doing the Google you know Maven program like everybody was Chinese on that program and I uh like um that's totally you know uh I I mean that's totally
wonderful so all I say here is that you know I I disagree with kosla because primarily because what you said that I think open source is critical to Innovation I think the even if 99% of Open Source software is not strategically important the 1% that is developed in an open way might be crucial Banning it prevents that we do not know what that 1% is it's a very
complex system as you said so you can't even really Trace what that is the other thing that I think never really gets talked about is that the spread of knowledge and Technology globally which open source helps quite a bit even if some of it is militarily relevant potentially does more to lift people out of poverty and thus promote
peace than it does to still conflict it's not Zero Sum you started at the beginning by saying that is not Zero Sum so if open source can help lift more people out of poverty around the world such that they're um they're not kind of hellbent on um uh you know extreme ideologies are taking over using military resorting to military means
that is much better for Humanity than uh than the alternative again preaching to the choir here um and the idea that is that is difficult it seems for some even really really smart people is what you just said right like when you're thinking when if you if you if you put everything into a box or a category and and you think just from that okay who we've got
to have really great security around our AI development because you know our adversaries are going to be able to come in and that totally blinds you I think it blinds you to the idea that well wait a minute like if open source spreads globally and if it ameliorates poverty what kind of New Attitudes are those countries going to grow into right I
think of South Korea uh as an example as know uh South Korea for a long time was a very authoritarian country and and the the influx of a middle class when they opened up the economy uh created the demise for the absolute totalitarian nature of the previous regime it started happening in China too right we all remember tanaman Square in
1989 that was because Deng had opened up Chinese markets and guess what Iden thought about the fact that he was going to create a middle class that really didn't like being told that they all of their thoughts were owned by the state right and so there's these secondary and tertiary knock on effects that it seems to me that people just seem to not even
think about and one of the things I'm excited about AI with is and by the way I'm a big believer in the human in the loop type model the senar model but I'm also a big believer in would it be great to just for example in scientific discovery um if you look at uh the loop or the Hoops that scientist has to jump through to get a thesis or hypothesis
fund it right to test well one of the things is they're probably not going to seek funding for a hypothesis where they suspect they're going to get a null hypothesis right and yet so so the entire incentiv syst is oh I'm going to discover something new and great and Brilliant and that ignores all of the information that we can learn via negativa and like why don't we just have
an AI churn out hypothesis after hypothesis expecting a null a null set publish that to a open uh uh data set where all scientists can have access to it and they can look you know it's they it's the Sherlock Holmes the dog that didn't bark that's how he knew that the that the uh Intruder was known to the family Etc and and I do think that there
are just so many Incredible use cases for AI where we let the AI go on these speculations right because guess what in my opinion there's probably going to be several in there where we were expecting a null hypothesis and we actually come up with something yeah love this Jim I think um you mentioned the human in the loop stuff so it's often the case that
people make a distinction between Automation and augmentation one of the interesting ideas that I've I've thought about lately is that um the airplane didn't automate fly for humans um the electron microscope didn't automate the ability to see these you know microscopic objects we just couldn't do them you know like we just simply
couldn't do them and then to take it a step further um Technologies like tumor detection those are incredible things for AI to help us with but as you said AI can also help generate hypotheses and so that we grow our knowledge of cancer generally right like it isn't just about kind of tasks that we've we've realized are important kind of components it's
this sort of like open-ended science that that can happen um that can be aided by by AI so I think that's a really exciting new way I I think one of the areas that we're kind of working on at at Oxford that Philip is a is a real expert on Philip corus this professor there is the nature of reason and I do think there are enhancements to reason
that would help with this scientific project I think um this is very pithy but like the way llms work today is kind of like um more like an artificial politician than an artificial scientist if that makes sense and getting them to be more like the artificial scientist I think requires a little bit of upgrade of the the reasoning capacities um reasoning is
everyone knows it's important for hypothesis generation for science generally the way LMS kind of work today is you have these like comprehensive generative uh repositories of Common Sense judgments like things that you know you and I would see and think oh that's reasonable that's right like but I don't know that that adds up yet to
the reason of the kind that is needed for open-ended sites I also don't know that it's the kind that adds up to the reason that you would want for a trustworthy tutor for your four-year-old uh or for robustly respecting moral constraints and I'll just give you one example of where like Common Sense versus this other thing might fall down
so if I'm if we're asking questions if we're doing this kind of like um dialectical process together this Socratic process at some point you're going to ask me a question where I realize that the answer that I was holding in my head can no longer be an answer you like you know and can I grasp that by my own lights that that can happen across two
people but it actually happens I think internally like I think we have these questioning processes in our own head when we're really thinking through something and I don't know that llms can do that yet um so I think there's there's a little bit more work to be done at the Edge of Reason to be able to you know move Beyond just search which
will help us immensely generating this kind of you know hypothesis but also to kind of like more maybe uh deliberate um probing of the natural world through this wonderful tool we've created yeah and it it as I Was preparing uh for a chat today I was thinking you know your list again I love your list I I haven't I and you gave me a lot of great ideas there's several
books on there that I haven't read that I'm going to read now um but I also thought you know probably wouldn't be bad to put some science fiction books in there as well have you thought about that definitely and I need to read more I just read consider flebus uh the culture series first one um because that's one of the very rare cases where
um there's a positive vision for Humanity utopian Vision um and uh I think there are there's a ton to be gained from uh from that I've sprinkled this in in my own but haven't kept it up but I think that would be good the other one that I would say is like kind of slept on is Russian realism I took a class on cze off this past uh semester for a few months I I go to
classes at St John's College um which is a great great books program what I love about the Russian realists is that they're they're writing by by which I mean like um check out Doki Tolstoy Nabokov um or tanv rather um they they're they're all different and they have different degrees of like moralizing in different kind of places
but one common theme is that they're they're showing that humans are complex that there's a great kind of check off quote that says I start wherever any category you could Define me you could put me in ends they focus on the particularity of human life they focus on the non- neonian characteristic of human life how we you know have our own volition as I
as I said earlier and it's just this like wonderful antidote to the radical idea that you could just flatten human life with theories that you could just kind of run a topown state now did was that listened to was that heated in Russia absolutely not um but but for it's for the rest of us to uh to recover and to to take heart so yeah I think
Russian realism and science fiction are two missing pieces that I'd love to kind of flesh out May that I mean inter intellect is like I think they're doing incredible work and I'm sure that these are like them that would be right up their alley to explore so yeah I would love to collaborate on that yeah and the I mean if you look at our vertical at
oosi ventures like that's one of the things that we think we can help add value to discuss the discussion on like in infinite films uh we're going to uh make some movies that that does not view all Futures as dystopias um in infinite books same idea right we're going to uh feature a lot of writers who've been largely ignored um
in terms of hey you know what there's a there's a lot of different scenarios where the future works pretty well but my my challenge is I I don't want them to be Utopias because as you know Utopia literally means nowhere right yeah and and and so and and it gets me back to the Plato right which begin he's one of the first of the philosophers that that you're uh
reading list covers and you know if we look at Plato the man what we see is a guy who got really freaked out by Athens losing the war to Sparta and wrote the Republic right that was his that was very reactionary the Republic is a very reactionary text to play the man going holy Sparta just beat us and i' I've got to write something about that
right and and and the other part of Plato the man that's really interesting is that guy was one hell of a marketer right if you look at the his various pieces you know the fadis and and all of them and then you kind of go back and look interestingly enough that they were all named after very rich Athenians and other Greeks who he would then go to and
say would you like to fund my Academy so I'm so he wasn't as good as KL marks on that point but um but the uh it's funny I'm reading the fedis right now and people ask me like what am I reading I'm like oh I'm reading about erotic love and it's kind of a weird answer but U it covers that and rhetoric it's really great but um what I will say on Plato is
you know he gives this kind of in laying out in the Republic this city and speech it's kind of utopian but there's a subtle message there I think and not everyone would interpret it this way but some would say that that this city is designed to represent the demands of perfect Justice and that Socrates shows that it in fact cannot that it in fact
cannot exist um that you know the philosopher would have to come down and become you know less than and he makes clear in kind of the antical logic that this is impossible um that this uh doesn't fit the the role of perfect justice so it's self-contradictory I know it's hard to kind of um get this Nuance discussion but essentially when
people you know think that the Republic is this um utopian thing which poer you know picks up on in a big way I think there's a subtle possibility that he's actually laying out attention and then the rest of history of philosophy explores this tension between you know can we build the utopian future and can we not I don't think we can at least not
in that way that we're you know that that's being described um there's one other point that I was going to make about so you mentioned uh trying to create these films that or not trying to you are that are uh painting a picture of a vision for Humanity that is not polanish but it's not um a jeremi I guess or I don't know whatever the
opposite of that is um it's a it's kind of like a positive vision of the of the world that we might uh create I think this is neat I heard it said once that like during the medieval period there was a uh a number of there's a rise of kind of these ubis poems that were um these like Lamentations about civilizations pasted you know it's like the medieval period
was kind of like a period of stagnation and they were looking back to the Romans and seeing the colise or seeing these like statues and columns and saying what race of what Titans built this you know what Titans came before us and it was like this um kind of sad uh nost and so not a lot was built in the medieval period I would say and what are
the ephemera that we are using cultural ephemera to kind of assemble into the next thing unfortunately we're surrounded by dystopian I was at a Santa Fe Institute thing and I heard um John Nolan funny story I tweeted about him as Chris Nolan it was my only tweet that's ever gone viral and it was just straight up wrong um and I talked to him at the
bar that night and he never corrected me so I felt like an but but um uh I'll never be written into a a Nolan film I'm sure but but he said that Utopias or dystopia sell twice as much as Utopias and Nolan is a very thoughtful guy I didn't take that he was saying that he's like a slave to that reality I did not he's like I I was really impressed by him he's very smart
um but uh but that is an important fact about our culture so I think it's up to people to lead there are films like you know the people who created A Beautiful Mind I think there there's some really interesting films that I think are uh are uh are kind of showing what's possible but this is a big artistic project that I think is going to be go
hand inand with the philosophy development and with the tech creation to kind of lead us where we want to go yeah um you bring up A Beautiful Mind Ron Howard is a friend of mine uh we had kids at the same school um and actually they were all in the chess club and if you know anything about that back in that day uh you l the parents were
coralled and put in a room where they couldn't interact with their children at all and these things would go on for hours and hours and hours so you just end up uh like having chats with people and I I'd love that you bring up a beautiful mind because I I thought that he did such a wonderful job with that in terms of like just all of the things
that made you think about right like objectively John Nash's life pretty much sucked if you're John Nash right and and yet I love the end you know that that where he's giving the speech at after winning the Nobel and you know there are there are the ghosts that uh have accompanying him the whole way along and and the other thing I loved about it I Ron does a really great
job with this in my opinion is he he gets at the incredible complexities of the story of Nash and and what he achieved but he's always really good at bringing out the humanity of the sit the situation and and so uh I mean God I could only dream to be able to make movies as good as rock but but you know the point you bring up about the uh the
dystopias outselling Utopias 2: one well absolutely they do and again I think that gets us back to you know human nature and Frank Frankly they're more exciting right I mean like who who who really wants to watch into our uh movie of everyone just like having wonderful conversations with the AI floating around and saying Jim did you consider
this and that uh might be a little boring so we're trying to figure out a way to make it exciting uh and yet also hopeful I you got any idea you got any ideas for it I mean I'd say you nailed it with the Beautiful Mind I think it's um that's why I thought to bring it up it's you neither want the kind of um milk toast or like weak um kind of
positive Vision nor do you want the kind of just um uh uh fearful like I don't know dog whistle stuff it's the complexity of human life it's the particular it's like this this is uh unavoidably and essentially human and I think the Russian realists bring it out like um Ron Howard does a great job I think you know those those Russian authors like no one would accuse them of
painting a overly Rosy picture in anak kinina you know but it's um but it's human and it and it points it points the way uh past ourselves so I wish you look it's a very difficult project but um then again so is a cosmos Institute so maybe that's what we're that's what we're both drawn to yeah yeah well this has been amazingly fun for me uh I want
to get one question in because you're also uh an expert in defense Tech and uh you did U the project Maven and uh or contributed to um and you also mentioned Mark andresen's tweet where he talks about if AI in 2024 is approximately equal to Atomic Tech in 1943 then right obviously he's setting he's setting us up here with this tweet um
but what what are your what are your and you've touched on them already but like one of one of the things that I worry about was outlined in the book AI 2041 10 scenarios uh for the uh AI future and and the one that really got me kind of like scratching my head and saying damn was they had a scenario in which autonomous drones with a drop of your
DNA or my DNA basically are Hunter killers and they only kill one person the person whose DNA is in that Hunter drone like I don't want to end out a really bad note but like what what do you what do you think about those kind of scenarios and and how how can we address them not try to hide them away and say oh don't worry we're
working on that we're we're gonna have countermeasures I fully believe we will have countermeasures but how do you look at and maybe we don't have to address that exact scenario but how do you look at our uh to be open enough to address a lot of these terrifying like I mean we're talking about movies right that' be a great movie because it's so scary what do you
what do you think about that well so one thing that came to mind when you when you brought up inre tweet that if AI was like nuclear weapons in the 40s or 50s or whatever um what would we what do you think that and I think about this phrase which like sounds like it's very offensive but I don't I don't mean it that way which is that our Germans need
to be better than their Germans and this is obviously from an old an old movie um Doctor Strange Love and um I love that movie yeah and what I mean by that is that I think we need um wonderful Talent from all over the world and um brilliant people who are in Silicon Valley and other places um thinking about this problem set um and helping modernize the
uh the military which is in in desperate I spent a lot of time you know advising uh you know Special Special Operations Command and other areas uh chair the chairman of the Joint Chiefs I did a project with on uh North Korea and variety of things and we just were're in desperate need of uh of more progress and optim um but I think what
what's troubling what's tricky about this one is one of the threads you see in political philosophy and history is that there are a lot of really difficult debates about you know what is Justice what is um you know the role of Reason in in a in politics or whatever and then there's this like um Tasmanian devil that comes through which is National
Security and it ruins everything it's like the one it's the CH the deepest challenge to the coherence of any political philosophy is the fact that hey if you're getting destroyed like if you're actually going to be you know extinguished in war then you end up doing almost anything to survive and um and that topples kind of
the the plans of the the political theorists I'm being a little bit extreme about it but I think that's a real challenge so then where I go to is if I know that that possibility exists well one maybe as an aside I reject a number of the common solutions to these challenges which are things like treaties I think we should have treaties I think every war is
precipitated by a broken treaty and I can give you like a a graveyard of treaties that you know um there was a pretty good treaty that uh Putin ignored when he attacked Ukraine for example um and so I think treaties alone do not solve it I think where I finally get back to in like a first principle sort of way is like if you have a democratically controlled lawful
military that whose job is primarily defensive and who actually operates in that way um that military needs modern technology to be able to function to do its job it needs um checks in in the court system and so forth but if you have that kind of a military and if there are other militaries that are authoritarian in nature that don't
respect human rights um then then we need to be better than those militaries in um wielding military force and at the end of the day like it's sort of like a there's a there's a potential Clash of military systems but as I said like don't view it as the in the kind of like um uh fatalistic you know determined way I don't think it's the case that we will
go to war with China um but I think we will uh the next few years in the Department of Defense are going to be enormously important to say whether we can like credibly deter it um because right now we're not doing a very good job at um Fielding AI Solutions and operational capacities in the military yeah and I think that's the the
right tenor to take in terms of looking at these problems and I love the uh metaphor of the Tasmanian devil because uh you know it's it's part of the we've been talking for our entire conversation about you know the the sort of dogmatic people who just want to like throw everything they can uh to you know destabilize a coherent and hopefully
creative conversation about all of these things at the end of the day I think that that's also a very very important aspect of moving forward right they if if if we if we make things off limits it's it's like kids right if if if you say you cannot do this guess what the kids want to do right it's like uh Christopher bugley in his thank you smoking satire book uh he he he uh in it
there's a protagonist who is a lobbyist with the tobacco industry and who secretly tries to get Congress to make them uh put a skull and crossbones on pack some cigarettes because he goes do you know how much that'll increase our sales so so so trying to make things uh by and large taboo and no we cannot discuss that I think really retards
movement forward and and therefore like I will bring up ideas that like I don't personally uh subscribe to but I think look they're out there and people are going to talk about them and use them and you know who knows maybe maybe it'll be uh incorporated into the uh Cosmos uh Institute yeah here's the bring up ideas even if they're not right I love chapter 2 of on
Liberty where he says that the idea is either true in which case it's good to bring up it's partially true in which case it can combine with something else or it's false in which case it can give a clearer sense and a livelier impression to the truth and so it's sort of like this ultimate argument for bringing forth ideas and in fact he says
that as fallible beings the greatest you know confidence we can have the greatest conviction we can ever attain is through inviting everyone everywhere to challenge our ideas you know all commers amen I I think you that that is a wonderful summation of the way I look at the world and I think you look at the world as well well um if you if you're
familiar with the podcast you know that uh our final question uh makes you the emperor of the world uh we do have a few ground rules you can't kill anyone you can't put anyone in a re-education camp but what you can do is speaking of Nolan you can incept people we're gonna hand you a magical microphone and you get speak two things into it and the two
things that you speak into it the entire population of the world is going to wake up whenever their next morning is and they're going to say gosh I've just had two of the best ideas ever and unlike all the other times I'm actually going to act on these ideas what what two things are you going to accept yeah so I think one is um a love
of liberal education uh by which I mean the the inquiry into the ends of what is a good life um which will surely not arrive at the answer to that question and we definitely won't arrive at it um from the emperor you know in other words there won't be a sort of unitary answer to this question but for individuals um a love of the that kind of like
open-ended uh query inquiry that takes a life and is never solved and enhances reason all the while and then the second is this like you know sense of epistemic humility this sense of reason as a superpower as a something that you know sets us apart from from from you know the animals but that um can only take us so far if we can activate our reason and
thinking through how do we live in a world like that where we acknowledge this limitation and say all of the reformer instincts all of the best plans that we have the greatest intentions must be subjected must be put through this acid wash of epistemic humility um we emerge with systems like the kind that Jim you you're uh advocating for in
the world and that we've discussed today love both of those brandan and uh this has been so much fun for me we'll have to do a round two because we didn't get to half the things I wanted to get to but before we go uh tell our listeners and viewers where they can find you yes so so um we've got a substack on Cosmos Institute that we're going to start
putting out not just um in interesting intellectual pieces but also we're gonna we're going to do a formal kind of launch post here soon where I can talk and we'll go back and you can here the we previewed it here but we'll talk a little bit more about the kind of exciting things we're launching so check that out and then the reading list is on
uh Brendan mord uh.com so please check that out and send me inputs I can uh you know update that as I get additional uh uh wisdom from from listeners absolutely great thank you so much for your time I think what you're doing is absolutely fantastic and hope to be h a collaborator with you on it looking forward to it all right thanks Jim thank you