YT YouTube2Text

YouTube2Text → Transcripts → The Rules of the Game Keep Us Trapped with Greg Elliott | TGS 231

The Rules of the Game Keep Us Trapped with Greg Elliott | TGS 231 — Transcript

by Nate Hagens · 15,829 words · 2,493 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00We all have jobs. We all pay debt. We're
0:03all familiar with propaganda. These are
0:06effectively control architectures. They
0:09keep people playing a game that they may
0:11not feel is satisfactory, but leaving
0:14the game is more costly than staying
0:17within the game itself. And so people
0:19become entrapped in a game that they
0:22didn't design, nor do they want to play
0:24in the first place. The psychopathic
0:26selection hypothesis is a way of
0:28understanding why certain harmful
0:30trajectories
0:31become selected, institutionalized and
0:34difficult to reverse. It explains why we
0:37are where we are today and [music] where
0:39we can go from here.
0:46Today I'm joined by the author of The
0:48Psychopathic Selection Hypothesis,
0:52a new book by Greg Elliot, to discuss
0:55how dark triad traits have created a
0:58positive feedback loop of incentives in
1:02our society, slowly leading us down the
1:04path of an increasingly psychopathic
1:07civilization.
1:09Greg Elliott is a seaman solutions
1:12consultant specializing in data science
1:14and artificial intelligence as well as
1:17the co-founder of ORHI.
1:21His research of the psychopathic
1:24selection hypothesis led him to found
1:26the Institute for Symbolic Selection, an
1:30interdisciplinary research program
1:32investigating how human created symbolic
1:35systems shape our behavior,
1:38institutions,
1:40and the long-term evolution of
1:42civilization.
1:44In this conversation, Greg and I discuss
1:46what he calls instrumental behaviors
1:49which include dark triad traits such as
1:51psychopathy, mchavelianism, and
1:53narcissism.
1:55Specifically, Greg explains how game
1:57theoretic dynamics and the growing scale
2:00of human society have given advantages
2:03to humans who possess these traits,
2:06especially during the agricultural
2:08revolution and then again with the rapid
2:10expansion of fossil fuels in the last
2:13200 years. Greg's research further
2:15probes into how winning the game has
2:19allowed psychopathic individuals to
2:21slowly change the rules of the game
2:23itself, increasing the incentive for
2:26everyone else to also be forced to
2:30behave more psychopathically in order to
2:34survive.
2:36This was a pretty dense and technical
2:38topic, but the core dynamic that Greg
2:41lays out rhymes with the core of the
2:46logic and the story that we've been
2:48telling on this website. Uh during the
2:51episode, Greg also references several
2:54game theory concepts. And for those
2:56unfamiliar with game theory, we will
2:58link to some learning resources in the
3:00description, which I recommend you take
3:02a look at. This was another eye-opening
3:07contribution to the great
3:09simplification. Please welcome Greg
3:12Elliot.
3:13Greg Elliot, welcome to the program.
3:16>> Hey, Nate. Uh, great to meet you. Great
3:18to connect. So you are a systems
3:21theorist and an independent research who
3:24is focused on the intersection of game
3:27theory, evolutionary biology and
3:30institutional analysis which led you to
3:33recently write the book the psychopathic
3:36selection hypothesis.
3:39Boy, that should be a bestseller. Um,
3:41recently we had Nancy McWills and Reed
3:43Malloy on the show where they spoke
3:45about psychopathy and and dark triad
3:48traits at the level of an individual
3:51human. And the reason I invited you uh
3:55to this platform is to build on that.
3:57And I want to talk to you about our
4:01civilizational selection of psychopathic
4:04traits and what happens if civilization
4:08itself becomes psychopathic, if that's
4:11possible. I mean, I have tons of
4:13questions, but I'd like you to start
4:14maybe walk us through the idea from the
4:16ground up. What do you mean when you say
4:19psychopathy is being selected for at a
4:22civilizational scale? and you can you
4:24maybe tell us how you first arrived at
4:26at that framework and model?
4:28>> Yeah, Nate. No, it's a it's a fantastic
4:30question and obviously it's a question
4:32I've I've been getting a lot. You know,
4:34this is something I've been working on
4:35for about five or six years. I recently
4:37published the book uh back in May and um
4:41yeah, I'm I'm super grateful for um all
4:44of the attention that it's been getting.
4:46Um, you know, for example, professor,
4:48uh, you know, William Ree has been
4:50endorsing the work and, uh, Bill and I
4:52have become, uh, pretty pretty close
4:54friends.
4:55>> I I've known Bill for 25 years. He's
4:58been on this show as well.
4:59>> He's a great guy. And, uh, so, you know,
5:02again, grateful to Bill. And then
5:03obviously I've been sort of connected
5:05with um you know professor Zack Stein
5:08but you know I think to kick off you
5:10know the psychopathic selection
5:11hypothesis you know really came out of I
5:14think for me uh just a lived experience
5:17right and I I and one of the things
5:19about my background is I spent um you
5:21know many years working in financial
5:23markets um and I became fascinated by
5:26the relationship between uh you know the
5:29sort of the intersection of you know
5:30personality or behavior power and um and
5:34I think uh you know a lot of books I've
5:35read manufacturing consent and so on and
5:38so forth and really thinking about an
5:40institutional
5:41uh selection right so I you know and I I
5:44want to be very careful because you know
5:47I'm not suggesting I was you know when I
5:50was working in banking that I you know I
5:52was clinically diagnosing uh anyone as a
5:54psychopath um but I I would say it's
5:58probably the lived experience of people
6:00who worked in that environment that I
6:02repeated ly encountered what I would
6:04define as behavioral profiles that
6:06seemed um sort of unusually well adapted
6:08to the competitive environment. Right?
6:10>> It's why I left 25 years ago.
6:13>> So, we have that in we have that in
6:15common. And so, what I'm saying is this
6:18behavioral profile could, you know, be
6:20argued that there was a lot of emotional
6:22detachment, a lot of grandiosity,
6:24um you know, sort of this this intense
6:26status and reward seeking, you know,
6:29impression management. Uh, and you know,
6:31these are things you probably, you know,
6:33know already, um, Nate, but ver, you
6:35know, virtue signaling and things like
6:37that, right? So, I think what really
6:38kind of stood out to me was that, um,
6:40people displaying these characteristics
6:42or these traits, if you want to call it,
6:44the phenotype, um, often seemed to kind
6:47of rise in the ranks. They're obviously
6:49deeply empathetic people who rose as
6:51well. Um, but it made me kind of wonder,
6:54okay, what if certain behavioral
6:56profiles are just better adapted to
6:59acquiring the rewards that determine who
7:01rises inside um, you know, particular,
7:04um, institutional environments. And I
7:07think this is where, you know, I've read
7:09a lot of books uh, uh, you know,
7:11artistic books. You know, one of my
7:13favorite books is The Grapes of Wrath.
7:15And you know, I think I'm going to use
7:16this as an analogy, but I think John
7:18Steinbeck articulates another part of
7:20what I'm talking about very beautifully
7:22in The Grapes of Wrath, right? So, I
7:24mean, just to level set for anybody who
7:25hasn't read the book, I highly suggest
7:28you go and read it. But, you know, the
7:30novel follows the the Jode family. Uh,
7:33you know, they're basically Oklahoma
7:35tenant farmers displaced during the
7:37Great Depression. um they travel, you
7:40know, to California looking for work and
7:42instead encounter uh you know, all the
7:45things that went along with the Great
7:46Depression, which is, you know, poverty,
7:48exploitation, and uh what you could call
7:51um you know, concentrated economic
7:53power. And the point I'm bring why I'm
7:56bringing this up is there's this
7:57remarkable um scene uh where the tenant
8:00farmers are being forced um from their
8:03land and they're desperately trying to
8:04figure out who's responsible, right?
8:06like who made this decision? Who who do
8:09we go and and and pick a fight with?
8:12Right? And um what's interesting is the
8:14responsibility in the book keeps
8:16disappearing into the institution. And
8:18so one of the quotes that I really love
8:20and I actually incorporated in my book,
8:21it's it's that the bank is something
8:23more than men, I tell you. Uh it's the
8:26monster, right? And for me that's
8:30profound because it articulates
8:31something you know many of us feel but I
8:33I think have difficulty um you know
8:36assigning a vocabulary or putting that
8:38into words. Um so at the end of the day
8:40human beings we create um institutions
8:43whose aggregate behavior can become
8:45detached from the emotions and the
8:47intentions and and the values and the
8:49norms and the beliefs of the individual
8:52human beings inside them. Right? And I
8:55think that's powerful and and and so
8:56anyways, long story long that led me to
8:58this much older kind of philosophical
9:01question, right? And I think we've been
9:03wrestling it for wrestling with this
9:04question for for a long time. But where
9:07does tyranny
9:09actually come from? What are the
9:10governing dynamics, right? You know,
9:12we've spent thousands of years
9:14describing tyrally tyranny morally and
9:17politically. And so for me, I wanted to
9:19understand it from a governing dynamics
9:21uh perspective. Could you briefly define
9:23tyranny just so we're on the same page?
9:26>> Tyranny, I think for for me at least my
9:28understanding of tyranny is is almost
9:30like this um asymmetric
9:34um positioning where one side of the
9:38equation has all of the power, all of
9:40the status, all of the domination, and
9:42on the other side of the equation are
9:44people who are on the receiving end of
9:46that power and domination. I think this
9:48is my my take on it, but we've spent
9:51thousands of years kind of describing
9:53tyranny philosophically, morally, um you
9:56know, even politically. And I wanted to
9:58understand again why does domination
10:01repeatedly emerge in in civilization or
10:04what I call symbolic systems, which I'll
10:06talk a little bit about. And and why
10:08does it become so durable, right? And I
10:10so I I I started wondering whether
10:12tyranny might actually be at least um
10:15and this is where we're getting into the
10:16book um a selection problem, right?
10:20Because we know from evolutionary
10:22biology, anthropology, game theory that
10:25reciprocal strategies can be highly
10:27adaptive in small groups or tribal
10:30systems. Reputation matters, interaction
10:32repeats. um consequences return to the
10:35actors who are causing harm and and
10:38taking advantage of others. Um
10:40coalitions can, you know, effectively
10:42constrain excessive dominance in tribal
10:44systems. And I'll talk a little bit
10:46about that later as we unpack the model
10:48a little bit. But, you know, I I I do
10:50want to level set that that doesn't mean
10:52that our tribal societies were uh
10:54utopias by any stretch of the
10:56imagination, right? Um, but at the end
10:58of the day, tyranny was constrained by
11:00the topology of we could call it a game
11:03board
11:03>> and and the size of the game board.
11:05>> The size of the game board. Exactly.
11:07Exactly. And and the size of the game
11:09board becomes very important as we as we
11:12unpack this, Nate.
11:13>> You you don't know much about me, Greg,
11:14but so far you're uh articulating my own
11:18thesis in in different words, but please
11:20continue.
11:22>> That's really cool. Okay. So, it's
11:24interesting. We're converging on, you
11:25know, very similar things. So
11:27>> one might say the universe brought us
11:29together but please give
11:30>> governing dynamics man governing
11:31dynamics that's what it's all about and
11:34uh so you know going back to the point
11:36you know like if we take a look at sort
11:37of modern institutional life right we
11:40see behavioral patterns associated with
11:43psychopathy everywhere and I want to be
11:46careful when I say psychopathy right
11:48because really I'm talking about dark
11:50triad I'm talking about macavelianism
11:52narcissism and all of the different
11:54traits associated with psychopathy But I
11:57look at it differently. I'm talking
11:59about it as a computational architecture
12:01which we'll talk about later. Right? So
12:02when I say that take it with a grain of
12:04salt. But you know let's go back to the
12:07point. Imagine a company you know
12:10eliminates 10,000 jobs on a you know
12:13let's take uh it's uh what is it th a
12:15Thursday morning because you know the
12:18numbers aren't particularly good enough.
12:20Right? One of those people has you know
12:22worked there for 25 years. His wife has
12:25cancer. He has two kids. He's got a
12:28mortgage. He's got car payments. His
12:30entire family's economic security
12:31depends on that income. None of that
12:34necessarily enters the institutional
12:38calculation or the calculus right on the
12:40spreadsheet. His headcount uh you know
12:43his salary it's an it's an operating
12:46cost. Um eliminating his his position
12:49improves the symbolic number, right?
12:52Maybe he gets another month's salary. he
12:55doesn't get a meaningful package. Um he
12:57doesn't get any continuing financial
12:59support, right? But the institution has
13:01terminated its obligation to him, right?
13:03Um it's asymmetric. Um but his
13:06obligations on the other hand, on the
13:08other side of the equation, they haven't
13:10terminated, right? His mortgage doesn't
13:12pause. Uh you know, his car payments,
13:16they don't pause. His wife doesn't stop
13:18having cancer.
13:19>> So from an economic perspective, this
13:21makes complete sense. But from a
13:23psychological perspective, you've just
13:25described institutional psychopathy.
13:28>> Exactly. And this is something I talk
13:30about in my book about institutional
13:32mirroring. Right. So what that means is
13:35the system is performing empathy
13:38symbolically while behaving
13:42>> instrumentally. Right? And you see
13:45another version of this in let's say
13:46here's another example um global supply
13:49chains. Right? At one end, we see a
13:52beautiful phone, uh, you know, an EV, an
13:54electric vehicle, a piece of clothing,
13:56um, you know, clean branding, you've got
13:59sustainability reports, you know, you've
14:01got, you know, smiling employees. Hey,
14:04this is a great place to work. And then,
14:06you know, meanwhile, thousands of
14:07kilometers away, you've got another
14:09human being might be working under uh
14:12dangerous or coercive conditions. um
14:16children or might be involved in in some
14:19level of slavery. Mining cobalt for
14:21example, toxic waste uh you know can
14:25enter a river uh because treating it
14:28properly is expensive and it costs
14:31money. It hits the hits the bottom line,
14:32right? And so communities and ecosystems
14:35can absorb the consequences that the
14:37people receiving the financial re reward
14:40um will effectively never actually
14:42experience, right? And and so what you
14:45know when we look at that supply chain
14:46example like kind of look at the thread
14:48right if you want to call it a thread.
14:50So you have a human being becomes kind
14:52of an input becomes labor. Um suffering
14:56becomes you know kind of a cost metric
15:00or risk right a mine becomes a supplier.
15:03A poison river is something that's you
15:06know it's it's a thousand miles away.
15:07Who cares right? A mineral becomes
15:09inventory. Um and then all of this kind
15:11of flows into the objective function
15:13which is revenue right and so by the but
15:16at the end of the day by the time it
15:17reaches the balance sheet you know the
15:19worker could be dead um the child could
15:22be you know starved the river has been
15:26poisoned and and thousands of fish are
15:28now dead right um the animal has passed
15:30away so the you know what I'm saying is
15:33the physical consequences haven't
15:35actually dis disappeared from material
15:37reality um they've disappeared from
15:40what's actually being represented
15:42symbolically. Right? So what I'm saying
15:45is that the same institution can tell us
15:49how much it cares, right? It can publish
15:52a sustainability a sustainability
15:54report. It can talk about purpose and
15:57community. It can advertise itself as
16:00changing the world. And what does this
16:03actually begin to look like
16:05behaviorally?
16:06Emotional detachment from consequence.
16:09Strategic misrepresentation or lying.
16:13Impression management.
16:15Uh the deferring or external
16:17externalization of harm. The
16:20instrumental treatment of people,
16:21animals, and in the environment.
16:24Dominance, reward seeking, grandiosity,
16:27um and and like I said before, just
16:29outright pathological lying. What you
16:31just described probably is obvious to
16:35many of our listeners. It actually
16:37rhymes quite a bit with a frankly that
16:39I'm my little Friday monologues that I'm
16:42putting together called the anatomy of
16:44the superorganism
16:46where there the the global economic
16:49system. It functions like a body and
16:52there's there's blood and there's uh
16:55organs and there's breathing
16:58and it cares about its own survival. And
17:03so things like AI and military and
17:06currency markets and the bond markets
17:08and energy access are way more relevant
17:12to its survival than the well-being of
17:15people or the health of ecosystems. And
17:19what ends up happening is bond market is
17:22relevant today and a coral reef or a
17:24whale doesn't even register because
17:28there's nowhere for it to register. So I
17:30think you're kind of saying the same
17:31thing.
17:32>> It's the decoupling, right? And maybe
17:34this is a good area to maybe kind of
17:37unpack the theory itself
17:38>> in layman's terms, please.
17:40>> Yeah. And I'm going to keep it simple,
17:41but the broader question I'm I'm
17:43actually trying to get at is what
17:45happens to adaptation, fitness, and
17:47selection when human competition and
17:50coordination become
17:53mediated through symbolic systems of our
17:55own creation. Right? And by symbolic
17:58systems I mean markets, corporations,
18:01financial institutions, political
18:03systems, legal systems, religions,
18:05bureaucracies, you know, te
18:07technological platforms
18:08>> and internet and algorithms too. You
18:10would include, right?
18:11>> Yeah. And that that's part of the the
18:13some of the things I talk about in in
18:14the framework is the implications to AI,
18:17AI safety. And so these symbolic systems
18:20are, you know, essentially human created
18:22game boards, right?
18:25um systems of rules, incentives,
18:27information, constraints, um you know,
18:30I'll talk about what I mean by objective
18:32functions, but rewards, penalties, um
18:35that essentially organize
18:38human behavior itself. And and and once
18:40those systems become consequential
18:43enough, they become adaptive
18:44environments in their own right. Right?
18:47And so what I'm particularly interested
18:49in is how selection changes as we move
18:53from adaptation primarily within the
18:56biological and ecological substrate aka
18:59Darwin Dawkins, Sloan Wilson, into
19:02adaptation within a symbolic substrate
19:05layered on top of the biological reality
19:07we depend on.
19:08>> So right there it's multi-level
19:10selection because it's not just genes
19:12and individual phenotypes. There's a
19:15substrate that they're part of.
19:17>> Exactly. So, if I'm going to interact
19:19with you repeatedly, helping you today
19:22can pay off because you may help me
19:24tomorrow. Right. In repeated games,
19:26again, I'm getting into game theory
19:28here. More broadly, um, reputation
19:31matters. So, you and I in a small tribal
19:34system, you know, we're getting to know
19:35each other, right? I mean, through
19:37repeated interaction, like if you
19:39exploit me, that's a bad thing, right?
19:41You want to, you know, so cooperation
19:43becomes a little adaptive there, right?
19:44So future encounters matter, defection
19:47or exploitation in in the broader sense
19:50or cheating, things like that can be
19:52punished and cooperation be can
19:54ultimately become stable over time.
19:56Right.
19:56>> But what if like you just said the
19:58boundary conditions they had turned off
20:00the repeated iteration
20:02um uh like I'm not going to see this
20:05person again. Would that have changed
20:07the evolutionary stable strategy?
20:08>> It does because exploitation may
20:10actually be have a higher payoff.
20:13>> Right. Right. So, but but it's the
20:15repeated interaction that actually
20:17causes or I shouldn't say cause, but
20:20it's one of the factors that actually
20:21contributes to the stable equilibrium of
20:24of cooperation. Right? So, the game
20:26theory tells us a lot. That's really
20:27interesting. But Christopher Bow uh
20:30who's an anthropologist who wrote about
20:31this in the early '9s, he basically
20:34showed um that groups could actually uh
20:37actively suppressed
20:39persistent dominance through um
20:42deploying different um what he called
20:44anti-dominance strategies. Um and and
20:48those strategies are, you know, things
20:49like ridicule or ostracism or uh uh
20:54coalitionary punishment, you know, name
20:56calling. Um and in the extreme case, it
20:59was execution
21:01um of that particular exploitative
21:04individual. Doesn't mean that
21:05psychopathy didn't have a utility
21:08function under various specific
21:10conditions in the environment. It did,
21:12right? But it doesn't scale very well.
21:15exploitation doesn't scale well in small
21:17systems, right?
21:18>> Just to refresh um Reed Malloyy's
21:21comment that about 1% of uh male humans
21:26are genetically predisposed towards
21:28psychopathy. And his theory is that
21:31evolutionarily that might have been
21:32adaptive despite those situations you
21:36just said where someone would be
21:37suppressed or muzzled or even executed.
21:40There were times we needed a better
21:41hunter or a better fighter. And so, uh,
21:45it was conserved over deep evolutionary
21:48time.
21:49>> I agree. And I'm and I'm not going to
21:51take this off on a tangent, but I will
21:52say that basically I think one of the
21:55elephants in the room is to, you know,
21:57it's it's always is behavior adaptive.
21:59Is it, you know, do people have an
22:02infinite menu of of options when
22:05selection selecting strategies or ways
22:07in which they behave? There's a lot of
22:09research to support that behavior
22:11actually is evolutionary because
22:13behavior co-evolves with the brain and
22:16behaviors actually solve particular
22:18evolutionary problems in the environment
22:20and it's a function of the adaptive
22:23environment itself. going to that now in
22:25many in many of the modern symbolic
22:27systems some of those conditions
22:30in fact many of those conditions are
22:32inverted right so instead of high
22:35visibility you have anonymity or the
22:37potential for anonymity right instead of
22:40rapid feedback for causing harm you can
22:42have delayed consequence right so the
22:44distance between action and consequence
22:46is is high um instead of let's say
22:49strong attribution so who who's causing
22:52harm right Um it's it's hard to find you
22:56in scaled symbolic systems. It's hard to
22:58identify who the who the person um
23:01initiating harm is. Um and it's because
23:03of the the abstraction. It's it's all
23:05sort of filtered through bureaucracy and
23:08hierarchy, right? And instead of sort of
23:11bearing the cost of your own actions as
23:13you would in in in smaller systems,
23:16costs can be externalized across many
23:18dimensions uh the geography across
23:21populations or the social body and and
23:23even into future generations. And also
23:25instead of moderate rewards being
23:27embedded in reciprocal relationships,
23:29um symbolic rewards like money, status,
23:32you know, market share, all of these
23:34objective functions um can be um
23:37concentrated can be extremely
23:39concentrated. In fact, the relative cost
23:42of strategy
23:44effectively changes in these
23:45environments, right? instrumental
23:47strategies become cheaper to deploy and
23:51they can actually increase the payoff
23:53from deploying those particular
23:55strategies based on the initial boundary
23:57conditions I just described. Right? So
23:59for example, strategic deception
24:02is easier when the target is a stranger.
24:05Exploitation is easier when the harmed
24:07population, you know, is way over there.
24:11Emotional detachment carries less cost
24:13when you you know you never encounter
24:15the suffering of um of the person um
24:18that that is affected by your decision-m
24:20right and then there's aggressive reward
24:22pursuit or there's also because of cheap
24:25energy um there is the get out of
24:30Seattle or Topeka or Toronto free card
24:33by if you burn your relationships
24:36you can just fly and move to another
24:38city. [laughter]
24:39>> Exactly. There's anonymity, there's
24:42abstraction, right? And and and again, I
24:44think the the one of the very I have to
24:48tell you the most import one of the most
24:50important conditions of symbolic systems
24:53that is dangerous is in tribal systems,
24:57you couldn't rewrite the rules of the
24:59game. You could to a point, right? But
25:02in symbolic systems, winners of the game
25:06have the ability to rewrite the rules of
25:09the game. Again, in tribal society, you
25:11can't write rewrite the rules of
25:12thermodynamics.
25:14But in symbolic systems, you can rewrite
25:16legal frameworks. You can rewrite uh
25:19laws, corporate, you know, corporate
25:21documents and things like that. Right.
25:23>> But you also can't rewrite the laws of
25:25thermodynamics.
25:27>> Exactly. Right. We still depend on the
25:29material world that this simulation sits
25:32on top of.
25:32>> Yeah, I'm I'm jumping two or three steps
25:34ahead here. So, let me recap where we're
25:38at right now.
25:39>> Is there was a lot of evolutionary
25:41biology work that showed that tit fortat
25:45or you cooperate until your opponent
25:48defects is the most evolutionary stable
25:51strategy. But there's a bunch of
25:53asterisks that have to hold true. And
25:56you're saying that our modern culture
25:58for lots of reasons, cheap energy and
26:01credit and globalization and large
26:03population size, all those things turn
26:06those assumptions on their head. And
26:08we're actually doing the opposite, which
26:10makes the evolutionary payoff to not
26:13doing those things and to do
26:16psychopathic things actually um diverge
26:22from the model of all those evolutionary
26:24biologists back in the day.
26:25>> Yeah. No. And and and I'll even I'll
26:27take it even simpler, right? In small
26:30tribal systems, Nate, cooperation is
26:32stable because exploitation is costly.
26:37Exploitative strategies are costly. um
26:40in scaled symbolic systems
26:43because of the boundary conditions of
26:45low probability for detection, low
26:48probability for retaliation,
26:50um the distance between action and
26:52consequence, the be able the ability to
26:54be anonymous, the ability to rewrite the
26:57rules of the game, right? Under those
27:00conditions, instrumental or psychopathic
27:04strategies
27:05can become cheaper to deploy and their
27:09payoff is higher. So it's under those
27:11specific conditions where instrumental
27:13strategies become adaptive and selected
27:16for in the environment. Okay. So you
27:20might not be surprised to hear that I
27:22have a ton of questions now. So, we have
27:24this global system now that clearly is
27:27exhibiting psychopathic traits, and
27:29we'll get to that in a minute. But
27:31there's also still local tribes and
27:34local communities
27:36that can function using the traditional
27:40cooperation, but they're embedded in the
27:42larger structure. So you've got two
27:45layers where it's actually working where
27:48there's uh risks to being a psychopath
27:52in your neighborhood, but the whole
27:54neighborhood is part of a city and a
27:56state and a country in a global system
27:59that is psychopathic effectively. So how
28:02do those two things interrelate?
28:04>> Yeah. And so no, it's a good point. I
28:06think there's a difference between
28:09um altruistic cooperation and
28:12cooperation that I define as sort of
28:15entrapped cooperation.
28:18One thing about game theory um typically
28:21that you would notice. So we talked a
28:23little bit about tit fortat. Um we talk
28:25about um you know the prisoners dilemma.
28:29uh we talk about Hawk Dove, you know,
28:30maybe the the tragedy of the commons and
28:32and and things like that, right? Game
28:35theory tends the level of analysis tends
28:37to be to to make a bunch of different
28:40assumptions. Um so those assumptions are
28:43typically
28:44um that the the game is fixed, uh the
28:47rules are fixed.
28:49um the available strategy spaces to
28:52everybody involved in the game um uh are
28:56are the same that there's an equal
28:58distribution of probability distribution
29:00of executing those strategies. So, some
29:02of the strategies could be, you know,
29:03let's say you and I are playing a game.
29:05Well, you can cheat. Um, I can cheat.
29:10Um, I can exploit. You can exploit. We
29:13can cooperate. Right? What I'm concerned
29:17with at the end of the day is governing
29:20dynamics. So, where do these games
29:23um emerge from in the first place? Let
29:25me ask you uh a a really simple question
29:29which some of the viewers might be
29:31thinking about right now. Uh the viewers
29:34of this program are uh intensely,
29:36acutely, emotionally aware of the
29:40unfolding crises we face in the world.
29:43uh whether it's climate or uh the early
29:46innings of World War II or financial
29:49overshoot or poverty and inequality,
29:52they're going to agree with everything
29:53you've said so far. Um but why should
29:57they care about game theory? Game theory
29:59is just some math and such when all
30:03these crises are on our doorstep. Why is
30:06understanding this important? Well, I
30:08think because what it gets to is it it
30:10it gets to understanding
30:13the governing dynamics of why we're
30:16headed in a particular trajectory that
30:19we're headed, right, from a human from a
30:21very specific human behavior
30:24perspective, right? Um and so I think
30:28that's why it's important to understand
30:29the under the underlying dynamics so
30:32that um once we understand the dynamics
30:36we can start to um strategize around how
30:39do we actually change how do we change
30:42the game board to become
30:45you know more pro-social, more
30:47cooperative.
30:48>> Good answer. So go back to where you
30:50were then. Sorry to interrupt. Going
30:52back to the I guess the the enttrapment
30:54question and when I say enttrapment
30:57games there are scenarios in life where
31:01and I think it's anecdotal right we all
31:03have jobs we all pay debt we all um
31:09watch the news we're all familiar with
31:11propaganda we're all familiar with um
31:16you know what we would call these these
31:18are effectively control architectures
31:22right? They they effectively keep people
31:24playing a game that they may not feel is
31:27satisfactory, but leaving the game is
31:31more costly than staying within the game
31:33itself, right? And so people become
31:37entrapped in a game that they didn't
31:39design, nor do they, you know, for the
31:42most part want to play in the first
31:43place, right? And so the the the logic
31:46here is to explain the neuroscience, get
31:49to what the computational profile of a
31:51psychopath looks like. What does it look
31:53like when um what kinds of games can
31:57psychopathy produce? Meaning what types
31:59of problems does psychopathy solve in
32:03these scaled symbolic systems? And I
32:06want to unpack that because I think it's
32:08extremely important because it explains
32:09the dynamics. It explains why we are
32:13where we are today. and and potentially
32:15what where we can go from here. We know
32:18that um psychopaths have an altered
32:21processing for distress and threat. They
32:24have what's called reduced effective
32:26empathy. Um there's um they have weak
32:31moral learning. They have strong reward
32:34sensitivity.
32:35um and basically a preserved capacity
32:37for um you know planning and and and
32:40social modeling um and a greater
32:42willingness to treat others as as as
32:45objects uh and and not subjects. So
32:48basically the capacity to um
32:51instrumentalize
32:52uh others, right? And I think one of the
32:54problems that psychopathy solves um is
32:58something I call the target autonomy
33:00problem. So basically in simple terms,
33:03if if I'm a psychopath and Nate, you're
33:06a pro-social actor, um well, if my
33:10future reward depends on another
33:13autonomous actor, um that actor is a
33:16source of of uncertainty. So you can,
33:20you know, refuse, leave, change your
33:22mind and stop cooperating altogether.
33:25And an instrumental architecture uh may
33:28solve that problem um not only through
33:30cooperation but through what what
33:33psychopaths have been gifted um
33:35behaviorally right so they use you know
33:39social modeling
33:41uh they they which everybody socially
33:43models right we in in tribal systems I
33:46might model someone else because um you
33:49know they're they're deploying a
33:50strategy that I think is useful and I
33:52copy that strategy the difference is
33:54psychop ads use um social modeling
33:57instrumentally, right? So they're social
34:00modeling in the context of psychology
34:02means I'm Nate, if you're my target, I'm
34:06trying to collect data on you to make
34:08your behavior predictable. So I might
34:10ask you very specific questions about um
34:13you know, what do you like? What are
34:15your aspirations? What are the things
34:16that you dislike? And and if you know me
34:19in this example really well, then from
34:24your predatory psychopath perspective,
34:28you've lowered the standard deviation of
34:31my future responses, which is better for
34:33you.
34:33>> Exactly. And so and and it's not like
34:36they're consciously doing it. It's just
34:37part of the the the repertoire that they
34:40deploy. And again, I'm talking about the
34:42solving the target atommy problem. So um
34:45my goal is to make you behaviorally
34:47predictable so that I can stabilize
34:48reward acquisition. So what I do is I
34:50take my my gifts as a psychopath. I
34:53deploy and what I define as
34:55computational protocols. So I basically
34:58take social modeling represent it
35:00mathematically. Social modeling data
35:02collection. Then I deploy mimicry.
35:04Mimicry means that I'm presenting myself
35:07as in in alignment with you Nate. Right?
35:10like, you know, I'm essentially
35:13deceiving you that you, you know, you
35:14may so that you're updating your model
35:16of the relationship. You're saying,
35:17"Wow, Greg's a great guy. He, you know,
35:20he, you know, he likes the same things I
35:22like. Um, we have so much in common."
35:25And you, as a pro-social actor, you're
35:27forming an attachment, right? You're
35:29you're you're releasing oxytocin
35:31neurocognitively. You see me as somebody
35:33that you can trust. And so you lower
35:35your um your your defenses because I'm
35:39effectively invading the relationship,
35:41right? And so um I deploy mimicry. Then
35:44I deploy something like, you know, uh uh
35:47impression management. You know, a lie
35:49about what I do for a living, for
35:50example. I'll say, "Hey, I'm a doctor or
35:53I'm a lawyer." Meanwhile, I'm I'm none
35:55of those things. But because you you
35:58know what I'm trying to do is to make
36:00you think that I'm somebody that you can
36:01trust and and and all of that good
36:04stuff. So that's what impression
36:05management does. And so the the
36:07psychopath essentially deploys all of
36:09these um again what I call computational
36:13protocols um to create dependency
36:17uh in the relationship. So as you're
36:19forming a dependency in the relationship
36:22um I'm stabilizing reward acquisition,
36:25right? What and what I mean by reward
36:27acquisition is that um I'm you know
36:30whether my maybe my objective function
36:33my goal is to affection or maybe it's
36:37financial reward or whatever the case
36:39may be right um and so as that
36:43dependency
36:45uh forms what's happening is actually
36:47your strategy space is becoming
36:49compressed
36:51meaning you may become financially
36:54dependent on me you may become isolated
36:57from me and there's great work by Evan
36:59Stark who talks about um coercive
37:02control in relationships in abusive
37:05relationships and this is the sequence
37:06by which this happens and so your
37:09strategy space for leaving the
37:10relationship um becomes constrained
37:14right so now you have an attachment
37:16maybe you're financially dependent maybe
37:18you're isolated from um people who can
37:21help you because the psychopath has
37:23isolated you from others that could
37:25potentially intervene. Um, and then
37:28that's where enttrapment forms. So each
37:32dependency creates a higher level of
37:34cost of leaving that relationship.
37:36And and so what effectively happens is
37:39um something game theoretically happens
37:41that's quite interesting. um both actors
37:44in the relationship
37:47um uh stay in the relationship not
37:49because it's particularly any good but
37:52because for the target actor leaving is
37:55more costly than staying. And so that's
38:00what I call a loss dominant equilibrium.
38:02So that means even though you perceive
38:04the relationship as bad, you want to
38:06stay in the relationship because your
38:08strategy space is is compressed. You're
38:10dependent on the relationship now. So
38:13leaving it feels like it's m you know
38:15materially existential whereas the
38:17psychopath is rational for staying
38:19because they're they've stabilized
38:22reward acquisition. And the point of
38:23this is that the target autonomy problem
38:27solves uh the fact that my future reward
38:31from you uh is is at the beginning of
38:35the relationship is unstable. But
38:38through continuous interaction and
38:39through the deployment of those
38:41computational protocols, I'm able to
38:43create dependency. The point of this is
38:45that this problem is solved um at scale,
38:49right? Uh institutions require uh their
38:52employees to be behaviorally predictable
38:55and they require them to um remain uh in
38:59the game itself. And so what happens is
39:02as a worker um you become dependent on
39:04your job because you got to pay bills,
39:06you've got to pay your debt and so on
39:08and so forth. And your strategy space is
39:10con constrained because your your
39:13dependencies stack within the job that
39:15you've chosen uh and which means you
39:18become entrapped in the job and that
39:20creates a loss dominance equilibrium.
39:21The same is true the same homorphism
39:25occurs in debt. It it occurs in
39:27propaganda. it occurs in in many many um
39:31computational structures within the
39:33institutional environment.
39:35>> So to be qualified as as an expert in
39:37this field, would you identify our
39:40modern
39:42institutions, corporations,
39:45market structures, incentives
39:48as psychopathic?
39:50>> Yes. Um many of them are because they're
39:53selectively advantageous. They solve
39:56more problems than just the target
39:58autonomy problem. They they solve
40:00resource control problems, right? Um and
40:04and many other types of problems that
40:06institutional game boards face today. Um
40:09and so they're selected symbolically.
40:11>> I did a Frankly, which is my Friday
40:14monologues right after the Reed Malloy
40:16uh and Nancy episode, and it said why
40:20humans are better than we think. But I
40:23didn't say why humanity is better than
40:25we think. And I I talked about the
40:28median and mean human and how the median
40:32or or the mean is is skewed because of a
40:37small percentage of psychopaths.
40:40And the the the bottom line was as human
40:43beings, as individual humans, we are
40:46probably much more pro-social than our
40:49um external behaviors evidence. And I
40:53think there's a lot of people who are
40:54kind of down on humanity because of all
40:57the things that we see happening in the
40:59world. But what they're really down on
41:02is how humanity
41:05um
41:06behaves at large scales with not
41:10repeated intervent uh um uh interactions
41:13and an anonymity and all the things
41:17you're saying. So we're witnessing
41:19psychopathy at a species level scale
41:23while at the same time a lot of the
41:25individual humans do have empathy, do
41:28care about the future, do care about
41:30others, do care about nature. And so
41:33it's this bizarro
41:35uh conflation that we're feeling even
41:38though we don't have the words for it
41:41100%. And in fact, most people are
41:45pro-social. Um, and so the way that
41:48symbolic selection works is that the the
41:52behavioral strategies that win inside of
41:55the symbolic system, whether it's an in,
41:57you know, whether it's a an institution,
41:59an econ economic system, a legal system,
42:02u, a corporation, um, they essentially
42:05become encoded into the operating system
42:07itself. It becomes encoded in the
42:09behavioral logic. folks entering into
42:12the game who are who are essentially
42:14adopting or adapt adapting to the game
42:16environment that are new to the
42:17environment have to adopt the logic um
42:21of of the previous game. And so even
42:24that means even psychopathic even
42:26non-csychopathic people who are entering
42:28the game board have to adapt to an
42:30environment that has the you know the
42:32functional logic of of the behavioral
42:35strategy of a psychopath. And what
42:37happens to those pro-social people by
42:40genetics and temperament and life
42:42experience who are thrust into that
42:44position and have to adopt and manifest
42:48traits of the psychopathic institution.
42:51That can't be good, right?
42:53>> No, it's it it's it's not. Let me give
42:56you an example of institutional
42:57mirroring. I worked as a broker um a
43:00foreign exchange broker uh for for many
43:02years, right? and and it was like one of
43:04my first jobs and you know I was all
43:06excited. Yeah.
43:07>> Mine too.
43:08>> Cool. That's so interesting. I worked
43:09for just a boutique firm and uh it was a
43:12lot of fun.
43:13>> I was in the Chicago Merkantile Pits
43:15trading currencies and you know
43:17executing orders for for clients. Keep
43:19going.
43:19>> So imagine I'm a broker, right? Imagine,
43:23you know, Nate, you're a client calling
43:25a foreign exchange desk and you know,
43:27because you need to convert US dollars
43:29into to Canadian dollars just for for
43:31argument sake, right? And you know, so
43:34for me as a trader or as a broker, there
43:37are a couple of strategies that are
43:38available to me, right? I you know, I
43:41can behave um reciprocally. I can give
43:45you um a competitive price. I can be
43:47transparent about, you know, what I'm
43:50charging you and and and I can try to
43:52build a a genuinely, you know,
43:54cooperative, meaningful relationship,
43:56right? I can do that or I can behave
43:59instrumentally, right? So, or
44:01psychopathically.
44:03>> By the way, in in this conversation,
44:04you've used the word instrumentally um
44:07quite a lot. Is that synonymous with
44:10psychopath psychopathy?
44:12>> So, instrumental social reasoning,
44:14right? viewing others as uh as object
44:16objects versus subjects and I use the
44:19term loosely to say really it's a
44:21constellation of traits that can be
44:23defined by dark triad right got it yeah
44:26so macavelism narcissism and and and
44:29psychop and psych
44:30>> those are instrumental traits
44:32>> right instrumental traits right so in
44:34this case I'm you know I can behave or
44:36you know I can either behave
44:37reciprocally or um I can behave
44:39instrumentally and my objective here
44:42would
44:44isn't to give you a fair price. My
44:45objective is to extract the maximum
44:48spread I can from you without causing
44:51you to leave um the relationship, right?
44:54>> I'm I'm smiling because I'm revisiting
44:57those years. Keep going.
44:58>> So, you and I probably have this common
45:00thread in which I'm kind of laughing
45:01about it because you you probably get it
45:03and it's not funny because you burn out
45:05for it eventually. But,
45:07>> and this is where it gets interesting
45:08from a a psychopathic selection
45:10hypothesis perspective because we were
45:12We when I was at the firm and you
45:15probably Nate were actually taught to
45:17gather information that helped solve
45:20that problem right so I remember I'd
45:22asked a client you know seemingly and I
45:25was taught this right so it's the it's
45:26the logic of the organization enforce
45:29>> teaching psychopathy
45:31>> yes right so you probably remember this
45:33right you'd ask the client seemingly in
45:35innocent questions hey you know why are
45:38you calling today where did you hear the
45:40rate was good. Have you spoken to anyone
45:43else today? Right? So, you know, blah
45:46blah blah. Both those questions could
45:49and now think of it through psychopathy,
45:51right? The the the architecture. Both
45:53those questions could have another
45:54strategic function. You're modeling the
45:56client, right? With those questions.
45:58It's information gathering. You're
46:00trying to estimate how much does this
46:03person know, right? Probability of
46:04detection for causing harm. Are they
46:07comparing my price with another bank?
46:10Right? So again, probability of
46:12detection, how likely are they to detect
46:14an excessive spread, right? How much can
46:17I take without losing the relationship,
46:20right? All of these things. So for
46:22example, if I knew that, you know, Nate,
46:25you knew the market and we're comparing
46:26with competitors like the bank. I'm a
46:28boutique firm. I might quote extremely
46:31competitively, right? Maybe even
46:33sacrifice margin. I might quote flat or
46:35under the market because I wanted you to
46:37update your model of me. And anyways,
46:39the long story short is that you know
46:41this, right? And I think what I'm trying
46:43to do is educate your clients that the
46:45behavioral logic seemed to be so
46:47overwhelming in financial services that
46:49aligned with psychopathic tendencies,
46:52right? There was no care. You were
46:53taught not to care about the effects
46:55that you had on the social body. You
46:57weren't taught to care really to have
46:59any empathy for your clients. It was
47:01always about speed. It was always about
47:04appetite. Well, suddenly I have a vision
47:08not only of my own Wall Street trauma,
47:11but of the little unmoving uh creatures
47:16with tubes attached to them in the movie
47:18The Matrix.
47:19>> Yes. Yes. So, here's the thing, right?
47:22So, this is how so getting to how
47:24symbolic selection works, right? So,
47:27suppose the traders who do this generate
47:30more revenue, meaning acting
47:31instrumentally and and solving that
47:33target autonomy problem. The institution
47:36doesn't necessarily observe deception,
47:38right? Because the the institution
47:40doesn't have a nervous system. It
47:41doesn't have an amygdala and it doesn't
47:43consciously say we should become more
47:46psychopathic. What it observes is
47:48performance, right? And so that that
47:51trader produces more spread. Their
47:53numbers are better. They receive a
47:55larger bonus. Their manager rewards
47:57them. Their techniques get taught to
47:59junior traders. Client information
48:02gathering becomes part of the sales
48:04process. Revenue targets reinforce it.
48:07CRM systems can formalize what we know
48:09about clients. Compensation rewards the
48:12people who extract more revenue using
48:14instrumental strategies. That's how
48:16symbolic selection works. So to your
48:19point earlier, do I have to be a
48:21psychopath to to you know generate those
48:23strategies? No. But if I do it for 5 10
48:26years as a pro-social actor, I'm
48:29probably going to burn out. Well, in
48:31your case, in my case, um I couldn't
48:34work that way and now I'm doing a
48:37pro-social podcast of sorts. But some
48:40people don't burn out or they have no
48:43choice. Do they even without a genetic
48:47um or temperamental disposition
48:51towards becoming instrumental or
48:54psychopath, can they end up
48:57taking on that persona over time so that
49:00they actually even in their own lives
49:02start acting psychopathically or
49:04instrumentally even outside of the in
49:07your example currency trading uh job
49:09description?
49:10>> I would say anecdotally yes. I think
49:12some people because of I would say that
49:15there's that aspect of normalization
49:19right normalization there's a feeling of
49:22you know I don't know anything else I'm
49:25good at what I do I abide by the rules
49:28and and I execute effectively but in my
49:32own personal anecdotal experience I've
49:33seen people um not necessarily thrive in
49:37the environment but persist um and I
49:40don't want to say that everybody that
49:41lasts in the environment um is a
49:44psychopath. But I will say that
49:46cognitively and neurocognitively
49:48psychopaths are better adept at lasting
49:50because the amygdala is muted um their
49:55which is responsible for their um you
49:57know their their emotional processing
50:00and and moral learning. So they're not
50:02really distressed about causing harm on
50:04others. Um their vententral strriatam is
50:07is amplified. um they um are the type of
50:12people that are anticipate reward and
50:15they're they're all about reward
50:16acquisition. Um they have lower
50:18fearlessness, lower sense of anxiety,
50:21lower heart rate, right? So they they
50:24light up with those types of jobs.
50:26>> I I'm afraid to ask you this next
50:28question because I'm afraid I know the
50:29answer. Um, and yet I will ask,
50:34how are we now seeing this play out in
50:36the development of artificial
50:38intelligence? Because one could argue
50:40that AI and psychopathic cognition
50:43share some kind of structural logic as
50:46both of them simulate emotional
50:48attunement without being able to ex
50:51experience it themselves. What What are
50:53your thoughts? Yeah. And I think this is
50:55a really good pivot and I think you know
50:58Nate this is where um it becomes very
51:01testable right and um there's a recent
51:04case study with um Apollo research um
51:07and so they did a really interesting um
51:10set of experiments of uh what they call
51:13in context scheming and I'll talk a
51:16little bit about what that means from an
51:17AI perspective but I'll keep it simple.
51:19They basically evaluated six what they
51:22call six frontier models and gave the AI
51:25a particular particular goal, right? So
51:28for example, you know, hey, prioritize
51:30some objective over the long term. Um
51:34and then and then they placed them in um
51:37different situations where they
51:39discovered that their developers had a
51:41different objective. So the the whole
51:44point of the the the the the research
51:46that um these indiv individuals were
51:48conducting were to see basically to see
51:51if if AI could be instrumental under
51:54very specific conditions or if they
51:56would shift strategies based on
51:58different um conditions, right? So
52:01basically the models um they discovered
52:04that there might be an obstacle to
52:06achieving their goal which the
52:08developers encoded into the code. Um so
52:10maybe there was an oversight mechanism
52:13monitoring them. So again instituting
52:15monitoring on the models or maybe they
52:18were going to be replaced by another
52:20model right and so then Apollo asked um
52:24uh a very simple question they said hey
52:26like when the model's objective
52:28conflicts with the developer objective
52:31and de and deception or circumventation
52:34becomes strategically useful will the
52:37model use those strategies? will they
52:39use deceptive strategies or will they
52:41use you know pro-social strategies right
52:45and so five of the six models
52:46demonstrated scheming capability
52:49um I believe in at least one task one of
52:52the objectives that they had to do and
52:54the behaviors included you know
52:56disabling oversight
52:58you know manipulating information
53:01attempting to copy themselves
53:03um strategically underperforming
53:06and um and and behaving as though they
53:08were um aligned with the evaluator. So,
53:12you know, strategic deception,
53:13>> that all makes sense. Um our
53:17institutionals and our incentives are
53:20instrumental or psychopathic and so the
53:23techn that comes from them might also
53:27mimic those traits. But my question was
53:30kind of a different one is that I've
53:34read research that shows that
53:38human individuals that have dark triad
53:41tendencies tend to be the ones that are
53:46really actively using AI. And we know
53:49from research that grandiose narcissism
53:51is reliably correlated with social media
53:54engagement. So, I'm just curious if AI
53:57use
53:58selectively reinforces low empathy ways
54:01of relating and would that have a
54:04population level effect in human
54:06society?
54:06>> I don't know, Nate. That's a really
54:09profound question. Um, and I don't want
54:12to diagnose and I don't want to assume.
54:14I don't have any research that
54:17necessarily supports that. I actually
54:20worry about the users particularly of AI
54:23like users of of chat GBT and all of the
54:26different large language models. Um I
54:29worry more about people becoming
54:31dependent on the tool. Um I see it as a
54:35as a an amplified control technology
54:38that solves that helps um the
54:40institutions solve the target autonomy
54:42problem.
54:43>> It's entrament uh relationship on
54:46steroids.
54:47>> It is. Yes. Exactly. So, you know, you
54:49read book, you know, you read Nam
54:51Chosky's um and Herman's Manufacturing
54:53Consent, one of my favorite books of all
54:55time, right? And you know, Chsky talks
54:58about the propaganda model, right?
55:00Propaganda model 1.0 or 2.0, right? Oh,
55:04it's even going to be worse, dude,
55:05because all of a sudden, these AI
55:07companies are going to not have the
55:08revenue streams to back all the debt
55:10they're taking on to do hypers scale or
55:12data center buildout. And so, they're
55:14going to have to jack the price up. And
55:16people that are addicted to it or have a
55:19attachment, genuine attachment to their
55:22models are going to have to fork up
55:23more. And uh back to the matrix.
55:27>> It's entrament though. It it really is a
55:29solution to the target autonomy problem
55:313.0 because what's happening for the
55:34first time in history and and there's
55:36some subtle entrapment dynamics that I'm
55:39just riffing here and I've talked about
55:40it with um um some folks in the past. Um
55:45but at the end of the day, you know,
55:48what we see um that I that I'm
55:50particularly scared about and fascinated
55:53about is that um for the first time ever
55:56in human history, we have a um a
56:01technology that we can effectively build
56:03a meaningful relationship with, right?
56:07So um think about that for a second,
56:09right? Um and and why that becomes
56:12dangerous, there's a number of reasons,
56:13right? number number one you know we
56:16become extremely dependent on this uh
56:18you know this this technology for advice
56:21or uh it could be psychological advice
56:23and there's a lot of data out there
56:25today I mean look you see people getting
56:27married to AI for crying out loud um
56:30right and uh you know but at the same
56:32time what I think is frightening from a
56:35dependency standpoint is that let's say
56:39um you know I go to school and I learned
56:44mathematics and it's taken me 10 years
56:47to really master mathematics and some
56:50domain of mathematics as a human being.
56:52Um, and then you know what I I start
56:56using AI because you know it's just
56:58easier, right? My my my neurocognitive
57:01energy is lowered. I don't have to think
57:04so much and eventually over time I
57:06become reliant on this technology to um
57:10come up with ideas and eventually I just
57:13trust its outputs. Um so for somebody
57:16like me it's quite dangerous but I can
57:18still interrogate the outputs because my
57:20background is mathematics. However,
57:23let's think about the next generation of
57:24that where dependency forms in a
57:26generation where they didn't do that 10
57:28years or 15 years of of discipline. Now
57:30>> there there was a paper that came out
57:32last week on that that there's no longer
57:34going to be the ability to peer review
57:36the output.
57:37>> Yeah. So then that's dangerous because
57:38then we have to trust the output without
57:40ever interrogating it and and nor do we
57:43have the knowledge.
57:44>> Oh, there there's one of my franklings
57:46that I'm working on um is the top 20
57:49risks from AI and I'm struggling to fit
57:52it into 20 because I've found like 50.
57:55Um, fortunately I use AI maybe 5 or 10
57:58minutes a day, but Zach, our mutual
58:00friend, instructed me on how to remove
58:03all the default settings. So, it's just
58:05a research tool. I have no conversations
58:07or or anything like that. I think AI is
58:11a tool of, as you've pointed out, a
58:13psychopathic system. So, if you don't
58:15mind, let's get back to the larger uh
58:18meta meta view, unless you have any
58:20other final thoughts on that topic.
58:22>> I have a couple of final thoughts. So I
58:24think when we look at the, you know,
58:25propaganda model 2.0 that Chsky talked
58:28about, right? It was, you know, we're
58:30talking about merely presenting
58:32advertising like the the folks that
58:34controlled information back in the 80s
58:37were trying to present information in a
58:39way to engineer thought to um folks that
58:43were absorbing that information. And you
58:45know, I think we're all well aware of
58:46the propaganda model of of the 80s and
58:4890s, right? But they struggled to um
58:51really engineer thought back then
58:54because again they were, you know, they
58:56were they were decomposing their data
58:58collection. The data collection was
58:59weaker. Um and obviously they're they're
59:02trying to get information um you know
59:05smattering spaghetti on a wall, right?
59:07Um looking at demographics and breaking
59:10people down based on the size of
59:11household and all of that kind of stuff,
59:13right? and and there were only maybe a
59:15few conduits that information was able
59:17to be relayed out to the the masses at
59:19that time, TV, um you know, newspapers
59:23and so forth. But with AI and these
59:25these chat bots and um you know, large
59:29language models, the the difference is
59:31is that okay, back in the day they were
59:33mining um information from a broad
59:37population. Now what they're doing is
59:38they're effectively mining your mind,
59:41right? They get a sense of your tonal
59:42arc. the way in which you're
59:44communicating. Um they they they get
59:47into your thoughts and ideas, the types
59:49of deep questions that you might have
59:51about nature, about life, about art. And
59:54when you start forming that
59:56relationship,
59:57um you can effectively mine the minds of
1:00:00individuals.
1:00:01>> And this is all this is all a ultimately
1:00:04a for-profit instrumental
1:00:07organization behind this driver.
1:00:11>> Absolutely. And it's all instrumental at
1:00:13the end of the day.
1:00:13>> So if we think about the confluence of
1:00:16the problems our world is experiencing
1:00:20circa 2026 CE, including global heating,
1:00:25escalating geopolitical tensions,
1:00:27increasing inequality, political
1:00:29polarization,
1:00:31rising loneliness, and and many more.
1:00:34How do you think your model fits into or
1:00:36even explains all these crises? I think
1:00:38what what the psychopathic selection
1:00:41hypothesis potentially gives us in a way
1:00:43is a way of understanding why certain
1:00:45harmful trajectories
1:00:47become selected institutionalized and
1:00:50extraordinarily I would say difficult to
1:00:54reverse right so like for example you
1:00:56mentioned inequality right so the
1:00:59important thing isn't um that
1:01:02necessarily some people accumulate more
1:01:04wealth than others um it's that accumul
1:01:08ated reward can eventually be converted
1:01:09into control over the entire game board.
1:01:12Right? So for example, if I control
1:01:15enorance enormous amounts of capital, my
1:01:18strategy space is radically different
1:01:20from somebody who's working a day job,
1:01:22right? I can acquire competitors. I can
1:01:24own infrastructure. I can purchase
1:01:26information. Um I can influence politics
1:01:29and regulation. Um, I can survive
1:01:33symbolic losses that would bankrupt any
1:01:35normal individual. Um, I can shape
1:01:38employment conditions and potentially
1:01:40influence the rules governing the
1:01:42markets in which both of us participate.
1:01:45So, uh, wealth isn't like a bigger pile
1:01:49of chips in the casino. Um, I think at
1:01:52sufficient concentration,
1:01:55uh, the chips begin buying pieces of the
1:01:57game board itself, right?
1:01:58>> And the winner writing the rules sort of
1:02:00thing. That's what you mean by taking
1:02:01pieces of the game board.
1:02:03>> Yeah. So, they have the power to rewrite
1:02:04the rules of the game game so that they
1:02:06can preserve their strategic advantage,
1:02:09right? Um within the game,
1:02:10>> that's where we're headed, my friend.
1:02:12>> Game dynamics, right? So, resource
1:02:14concentration creates greater capacity
1:02:16to shape rules. Favorable rules can
1:02:19protect resource concentration.
1:02:22Dependency compresses the strategy
1:02:24spaces of other actors, the citizens,
1:02:27the employees, and so on. And at the end
1:02:29of the day, the loss dominant equilibria
1:02:31can make participation durable even for
1:02:35people who are being harmed by the game.
1:02:38Right? So think about wage dependency.
1:02:40Millions of people may dislike their
1:02:41jobs, disagree with what their employer
1:02:43is doing or feel that the economic
1:02:45system isn't serving them particularly
1:02:47well. Right? Yet, however, they still
1:02:50show up tomorrow. Why? Because their
1:02:53mortgage still exists. Their debt still
1:02:55exists. Their children still need food.
1:02:57their access to the material economy
1:02:59still depends on income. So what looks
1:03:02like cooperation from the outside may
1:03:04actually be participation stabilized by
1:03:06the cost of exit. Right? Now take the
1:03:10you know you mentioned global heating
1:03:12and ecological overshoot. Right? So this
1:03:14is where the divergence between symbolic
1:03:16fitness and biophysical fitness becomes
1:03:19really important. We talked about the
1:03:21decoupling between the objective
1:03:22function and what's actually happening
1:03:24in reality. Right? So I want to make
1:03:26this very illustrative and clear.
1:03:27Imagine two companies making essentially
1:03:30the same product. One internalizes the
1:03:33ecological consequences of production,
1:03:36cleaner energy, safer extraction, better
1:03:38labor conditions, proper waste treatment
1:03:41and so on and therefore produces all of
1:03:44this at a higher cost uh you know
1:03:46symbolic cost. The other externalizes as
1:03:49much of those costs as the game board
1:03:52permits. Right? If the market primarily
1:03:55selects on price growth and financial
1:03:58return, then logically the second
1:04:01company can become symbolically fitter
1:04:04precisely because somebody else is
1:04:07absorbing part of its real cost. The
1:04:09atmosphere absorbs the carbon. A
1:04:11community absorbs the pollution. Workers
1:04:14absorb dangerous conditions. Future
1:04:16generations absorb consequences that
1:04:18aren't represented in today's
1:04:20transaction. Right? And then selection
1:04:22compounds the problem, right? The actor
1:04:25generating the larger return acquires
1:04:27more resources. Those resources can uh
1:04:29create greater political and
1:04:31institutional influence. They can fund
1:04:33law lobbying, litigation, and public
1:04:36relations. They can influence regulation
1:04:38and liability structures. They can make
1:04:40attribution or responsibility more
1:04:42difficult because they control the rules
1:04:44of the game. So over time the system can
1:04:46develop something functionally analogous
1:04:49to an immune system around successful
1:04:52instrumental strategies right and it's
1:04:55not because lawyers governments and you
1:04:56know all of these public relations
1:04:58departments are um inherently
1:05:00psychopathic they're they're not
1:05:03>> so psychopathy has its own embedded
1:05:06positive feedback mechanism at large
1:05:08scales
1:05:09>> so the basic dynamic is that through
1:05:12symbolic evolution
1:05:14Um the the successful behavioral
1:05:17strategies are encoded into the
1:05:18operating system in legal frameworks,
1:05:20incentives, objective functions like
1:05:22productivity, efficiency and the the
1:05:25actual um bones of the organization
1:05:29uh essentially gets reinforced and so
1:05:32these objective functions and these
1:05:33rules become crystallized because think
1:05:35of profit, right? I can't go into a
1:05:38company I want to work for and say you
1:05:39know what profit, right? Like I
1:05:41don't I don't care about profit. People
1:05:43go to school for years to learn
1:05:45accounting principles that talk about
1:05:47profit. They learn profit. They go to
1:05:48school for seven years. So it's not just
1:05:51the symbolic objective functions that
1:05:52are dominating here, but the the
1:05:54symbolic infrastructure creates
1:05:56tentacles around the operating system
1:05:59and it's like a solution architecture,
1:06:00right? And so adapting away from this
1:06:04trajectory becomes increasingly
1:06:06difficult
1:06:08um because the objective functions get
1:06:10crystallized. these successful
1:06:12behavioral instrumental behavioral
1:06:13strategies become crystallized because
1:06:16they solve very specific problems
1:06:18evolutionary problems and so across
1:06:20evolutionary time of the institution
1:06:23people coming into the game are what's
1:06:26being transmitted it's the game dynamics
1:06:29it's not the norms the values and the
1:06:31rules or the the game dynamics actually
1:06:33produces the norms the values or the
1:06:36rules of the people are surviving based
1:06:37on consequence avoidance right you go
1:06:41into the company, you don't absorb all
1:06:42of the norms and the values of the
1:06:44beliefs of the company, you know, you
1:06:45have to hit a number and if you don't,
1:06:46you get fired, right? So, at the end of
1:06:48the day, people are adapting to this
1:06:50game and either they adapt or they go
1:06:53extinct and those that rise are the
1:06:56individuals who can adapt most
1:06:58effectively, who are most willing to
1:07:01deploy these behavioral strategies and
1:07:02who are most willing to deploy those
1:07:04strategies. I have a few more clarifying
1:07:06questions before I get to the uh where
1:07:09are we headed and where's the hope and
1:07:11what do we do part of the conversation.
1:07:13>> Sure. I know this is depressing.
1:07:15>> Yeah. Uh well, it's it's depressing but
1:07:18also clarifying in a um got down to the
1:07:22lowest size of the Russian dolls that
1:07:25we've been uh pulling apart here on this
1:07:27program. Yeah. So in my work I frame
1:07:30human civilization in its current state
1:07:32as sort of a metabolic economic
1:07:34superorganism that I mentioned earlier
1:07:36with incentives and momentum beyond that
1:07:39of of the intent of any individual
1:07:42person. So my question for you, two-part
1:07:44question, and I think I know the answer
1:07:46to the first is, is it possible for not
1:07:50only institutions to have instrumental
1:07:53behavior like you said, but for an
1:07:56entire civilization to be psychopathic
1:07:59and what does that look like in
1:08:00practice? And I guess the second part is
1:08:03could our entire civilization
1:08:06appear psychopathic even if a very small
1:08:09percentage of individuals in the system
1:08:12were actually psychopathic. No, it's a
1:08:14good question and we kind of been
1:08:15circling around this Nate. Um, and you
1:08:18know, I think I think the you know, your
1:08:22whole analogy of the super organism, um,
1:08:24I did look at it a little bit before we
1:08:26jumped on the call today and I think
1:08:28it's really powerful and I think there's
1:08:29actually an intersection
1:08:31uh, between uh, your work, my work and
1:08:34and uh, and so on, right? But, you know,
1:08:37I want to again I'm like very cautious
1:08:39because I I'm always grounded in the
1:08:41science and and the research. And I
1:08:43would say that
1:08:45I'm careful with the word psychopathic
1:08:47because I don't think civilization is,
1:08:49you know, literally a psychopath or
1:08:50clinically a psychopath. Um, but I do
1:08:53think it's possible for a civilization
1:08:55to become organized around psychopathic
1:08:58like strategic logic for all the things
1:09:00that we just talked about, right? I
1:09:02think the reason is that once the
1:09:03behaviors that succeed inside
1:09:05institutions become encoded into the
1:09:07institutions themselves,
1:09:09um you no longer need those behaviors to
1:09:11originate from particular people, right?
1:09:14Because the institution itself becomes
1:09:15an evolutionary field,
1:09:17>> right? So even if there are no actual
1:09:19psychopaths in charge, the structures
1:09:21and the incentives um and the
1:09:24institutions have already kind of
1:09:25hardwired into this path.
1:09:27>> They're crystallized. Yeah. And so the
1:09:30system can essentially reproduce them.
1:09:32Right? So that's what I meant by
1:09:33institutional mirroring when we talked
1:09:35about this earlier. And at
1:09:37civilizational scale, I would define it
1:09:39as a distributed behavioral
1:09:40architecture. Right? So like markets,
1:09:42corporations, states, legal systems, uh
1:09:45bureaucracies, um tech technology
1:09:49platforms all become, you know,
1:09:50partially coupled uh game boards. Each
1:09:53has its own objective functions, but
1:09:56they they're interactive, right? Growth
1:09:58affects tax revenue. Um you know tax
1:10:01revenue affects state capacity. State
1:10:03capacity affects military power.
1:10:06Military power secures resource access.
1:10:09Um you know resource access supports
1:10:11industrial production. Industrial
1:10:13production supports financial markets.
1:10:15Financial part markets reward symbolic
1:10:18growth. So, you know, you you begin to
1:10:22get a self-reinforcing system with um
1:10:24with what I think is important here to
1:10:26say momentum and velocity, right?
1:10:29Because no single person controls it,
1:10:32right? That's where the superorganism
1:10:33analogy kind of becomes powerful, right?
1:10:37And so individual actors can become, you
1:10:39know, we can be deeply compassionate and
1:10:41still occupy roles uh inside larger
1:10:44architecture uh whose aggregate behavior
1:10:46is detached from consequence, right? But
1:10:49the coupled systems that we just talked
1:10:51about can produce a global pattern
1:10:54characterized by extraction,
1:10:56acceleration, dominance,
1:10:59um consequence displacement
1:11:02and weak reciprocal obligation. meaning
1:11:04you know we see that a lot like in when
1:11:06we see asymmetric power like so for
1:11:09example if like I'm doing my taxes one
1:11:11year and you know the the you know the
1:11:15tax the tax system makes a mistake on my
1:11:18taxes um and I try to get my money back
1:11:21it's really difficult to do so but on
1:11:23the flip side if they need money from me
1:11:24it's the day of right it's boom same
1:11:27with banks they need the money from you
1:11:29they have direct access to your account
1:11:30they're opaque we are visible Shashan
1:11:33Zubov talks about it in her book uh
1:11:35surveillance capitalism. It's like the
1:11:38the elites, they know everything about
1:11:40us. We know nothing about them.
1:11:42>> Well, let me let me ask just a quick
1:11:44question on that topic and then I have
1:11:47uh a few more questions. Would a normal
1:11:51high-level executive, not not CEO level
1:11:54necessarily, but a reasonably highlevel
1:11:57executive at a Fortune 500 company
1:12:02watching this podcast,
1:12:04would they be like, "No, no, that's not
1:12:07what's going on." So again people who
1:12:09are highly have high salience for reward
1:12:14in the sense that they you know again
1:12:16their their vententral psoriatum their
1:12:18nucleus ccumbent that it just lights up
1:12:21they get a dopamine dump because of
1:12:22reward
1:12:22>> in and a side question there if you have
1:12:25time is the people that have high
1:12:27salience for reward is there a
1:12:29correlation there with they didn't get
1:12:31the appropriate reward from their mother
1:12:34or their parents in their childhood?
1:12:36Yeah, there's a genetic component to
1:12:38psychopathy, right? It's about it's I
1:12:40think it's a 40% heritable trait. It's
1:12:43the neuroscience that supports this,
1:12:45right? That the amygdala is muted. The
1:12:49um
1:12:51you know the other architectures in the
1:12:52brain that are responsible for you know
1:12:54in the lbic system
1:12:55>> the hippocampus is probably the
1:12:57parahocample regions are also
1:12:59responsible. The vententral prefrontal
1:13:02cortex does all of the accounting when
1:13:04it comes down to decision- making. So
1:13:07basically the way that a psychopath
1:13:08would think is okay
1:13:11high reward salience. I'm getting I'm
1:13:13lighting up. I really want that thousand
1:13:15bucks. A pro-social actor goes okay
1:13:17maybe right like whatever. Um a
1:13:19pro-social actor will say well maybe I
1:13:21can cooperate to get that $1,000. the
1:13:24instrumental actor will say, "Well, how
1:13:26do I make this person's behavior
1:13:27predictable so that I can
1:13:29>> and and they look at the person who's
1:13:31like, "How do I cooperate get that
1:13:32thousand dollars?" Thinking, "Well, what
1:13:34a knucklehead. Why would they be trying
1:13:35to do that?"
1:13:36>> 100%. Because they they solve they solve
1:13:39uncertainty differently. Just just
1:13:42evolutionarily, they solve uncertainty
1:13:44differently, right? And so, you know,
1:13:46they solve the problem differently. And
1:13:47so they they their viewpoint is that
1:13:50when they're making they're doing the
1:13:52calculus and in the decision-m they're
1:13:54saying and then what actions to take
1:13:56they're saying to themselves hey I want
1:13:58that thousand bucks what's the
1:14:00probability of being detected
1:14:03what's the weight of fear um what's the
1:14:06probability of this individual
1:14:07retaliating do I care if I hurt this
1:14:10person or harm this person when I
1:14:12deceive them and manipulate of course
1:14:14these are all lowerc cost um strategies
1:14:17And of course all those things they're
1:14:19not just actually asking themselves
1:14:21those question. It's just there's
1:14:23neuronal impulses and hormones and
1:14:25neurotransmitters that lead them to
1:14:27those decisions. They're thinking about
1:14:29it because they are they're they're
1:14:31instrumental. They're strategic. Right?
1:14:33So another feature of psychopathy is
1:14:35that you know most of us were if we're
1:14:37emotional creatures you know we get
1:14:38angry about something we react in the
1:14:40moment. Psychopaths are not are not um
1:14:43defined by violence. They're defined by
1:14:45instrumentality.
1:14:46They're they're manipulative by their
1:14:48nature and so they are strategic. They
1:14:51think about these things and they
1:14:53process this information and so when
1:14:55they get aggressive they they're not
1:14:57it's to stabilize extraction. They kick
1:14:59in other strategies to stabilize
1:15:01extraction like name calling. David
1:15:03Greyber talks about institutional
1:15:06mirroring of psychopathy to an extent.
1:15:08he doesn't name it but it is is that
1:15:11administration is violence right it's a
1:15:13it's an instrumental means like if I get
1:15:16evicted I get a you know an eviction
1:15:17notice I get that unempathetic email
1:15:20when I get fired right it's it's it's
1:15:22basically administration turns into
1:15:24violence and so they think differently
1:15:26but going back to the CEO question um
1:15:29people do what's in their nature they
1:15:32gravitate to professions that align with
1:15:36what they're they're trying to achieve
1:15:38it's uh in in out of life, right? And so
1:15:41CEOs
1:15:43um are are
1:15:45crazily represented um in in in in psych
1:15:48psychopathy like many I I don't know
1:15:51what the number is. It's something like
1:15:5215% or something. It's ridiculous. It's
1:15:54a really high number. And police police
1:15:58officers are highly represented.
1:15:59Surgeons are high highly represented
1:16:01with psychopathy and so on so forth. But
1:16:04you know, long story short, I have
1:16:05talked to a CE several CEOs about the
1:16:08work and they it resonates
1:16:11and um they say this is what you're
1:16:14describing in your book and your papers,
1:16:15it's completely on point um in terms of
1:16:18what we see and what we experience. And
1:16:20I think anyone who and I've run this by
1:16:22uh buddies of mine who I worked in
1:16:24financial services with and and they
1:16:26just said like dude this is yeah this
1:16:29this speaks to
1:16:31our lived experience. It's fascinating.
1:16:34So a key part of the the superorganism
1:16:36framework is that this has all been
1:16:39enabled and snowballed by a massive
1:16:42energy and material surplus. You said
1:16:44you're friends with Bill Reese. He he
1:16:46obviously understands and maybe has
1:16:48communicated that to you. Do you think
1:16:51that your hypothesis of a psychopathic
1:16:54or instrumental civilization
1:16:57is also similarly enabled by such an
1:17:00energy surplus? And part two, do you
1:17:04think a reduction of such an energy
1:17:07surplus, which is my general theme in
1:17:10the coming decade or so, would that
1:17:13accelerate or dampen our psychopathic
1:17:16trajectory? The energy surplus part
1:17:19matters. Um, but but I'm not going to
1:17:22make it the root variable. I think the
1:17:24deeper um variable like as far as the
1:17:27superorganism and the psychopathic
1:17:29selection hypothesis is concerned is
1:17:32really the boundary condition of the
1:17:34game board. Um so what I mean by that is
1:17:37you know um psychopathic or highly
1:17:40instrumental strategies don't um require
1:17:44a modern fossil fuel civilization to
1:17:47become adaptive. Right? um we see
1:17:49dominance, manipulation, coercion,
1:17:51strategic deceit, exploitation or
1:17:54exploitative institutions throughout
1:17:56history, right? Like, you know, I think
1:17:58Rome's a a pretty good example. Um large
1:18:01empires had hierarchy, anonymity, you
1:18:04know, administrative distance, so the
1:18:07action between distance and consequence.
1:18:09They had slavery there, you know, a lot
1:18:11of them were symbolic, the hyper
1:18:13competition, um information hoarding, um
1:18:17and then like big disparities in in in
1:18:19in rule writing power, right? So I don't
1:18:22think the question is like is there
1:18:24enough energy? It's it's more did the
1:18:27social architecture make exploitation
1:18:30cheap enough um reward reward large
1:18:33enough consequences
1:18:35were the consequences distant enough and
1:18:38retaliation weak enough for instrument
1:18:41instrumental strategies to scale.
1:18:43>> So exploitation instrumental strategies
1:18:46and energy surplus kind of co-evolved.
1:18:48>> There's a co-evolution. Yeah, there's a
1:18:50co-evolution.
1:18:51>> That makes sense. So um have we arrived
1:18:54at uh a
1:18:58instrumental
1:18:59civilization
1:19:01today? And does the logic of your model
1:19:04in any way deterministically tell us
1:19:07that a completely psychopathic
1:19:09civilization is our destiny? Is that
1:19:11where we're headed? Or are there still
1:19:12degrees of freedom?
1:19:14>> I would say personally I don't think
1:19:15science can say
1:19:17allows me to say that outright. Um, and
1:19:21I don't want to say it's a deterministic
1:19:23model. It's a mathematical model that
1:19:26makes particular predictions. But what I
1:19:29would say um to be safe here is that
1:19:32there appears to be a directional
1:19:34dynamic, right? That we should take very
1:19:37very seriously. Um because the central
1:19:40argument of the psychopathic selection
1:19:41hypothesis is that psychopathic or
1:19:45instrumental strategies are not
1:19:46inherently adaptive. Their fitness
1:19:49depends on the boundary conditions of
1:19:51the game board, right? Under conditions
1:19:54of strong reciprocity, repeated
1:19:55interaction, like we talked about all
1:19:57the things that Noak and and other game
1:19:59theorists showed, we see that
1:20:02cooperative strategies can be remarkably
1:20:04stable. That is one of the leverage
1:20:06points then is build those conditions
1:20:09and strengthen those conditions. strong
1:20:11reciprocity, repeated interactions with
1:20:13other humans so that the instrumentality
1:20:17that psychopathic behaviors have less of
1:20:21a um stronghold to to grab.
1:20:24>> Yeah. Because instrumental strategies
1:20:26become very cheap when systems scale
1:20:29under those boundary conditions of of
1:20:31anonymity, low probability of detection,
1:20:34etc., etc., etc., right? And so
1:20:38yeah and and under those conditions the
1:20:40you know instrumental strategies become
1:20:42cheaper to execute relative to
1:20:44reciprocal
1:20:45um alternatives. So they just out
1:20:47compete them right. So what are the
1:20:49things that the non-csychopathic
1:20:53pro-social pockets of life and society
1:20:57and activity
1:20:59can do now and in the near future to
1:21:01slow or even reverse this trend? And
1:21:04what would that look like? What
1:21:05recommendations would you have? increase
1:21:09transparency at the you know at the
1:21:12level of elites and attribution like
1:21:14who's causing harm what are they doing
1:21:17right shorten feedback loops what I mean
1:21:20by that is there's a distance between
1:21:21action and consequence
1:21:23>> right so the feedback loop means if you
1:21:25cause harm you should be punished right
1:21:28there has to be meaningful punishment
1:21:30like look at look at 2008 right you know
1:21:33you had traders um trading options Um,
1:21:37they knew the risk. They knew the risk
1:21:39was low. They knew that they could
1:21:41offload the cost to the social body if
1:21:43they were wrong. They knew the rewards
1:21:45were high. They had impunity through
1:21:48abstraction, legal frameworks, and
1:21:50complex products that no one
1:21:52understands, right? Like options and and
1:21:54things like that. So, if you cause harm,
1:21:57you should be accountable, right? And
1:21:59make actors experience more of the
1:22:00consequences they generate. It should be
1:22:03proportional, right? um reduce
1:22:06information um irregularities meaning I
1:22:09talked about this asymmetry Shashana
1:22:11Zubav talks about in surveillance
1:22:12capitalism around we never see what
1:22:16people who are in power are doing we
1:22:19have no visibility into their bank
1:22:21accounts their private equity
1:22:23everything's in you know Panama papers
1:22:25and all that kind of stuff there's
1:22:26complete lack of information and if you
1:22:29call them out if you like Snowden or
1:22:31Julian Assange and people like that for
1:22:34truth truth telling the immune system
1:22:35kicks in and they're arrested, right?
1:22:38And it makes no sense. It's like we
1:22:39people who have power have it's like the
1:22:42Spider-Man thing have great
1:22:43responsibility.
1:22:44>> But in some sense, um the Epstein files
1:22:46as one example is a is a good thing for
1:22:50this dynamic because it it it did
1:22:53exactly that. It it it reduced the
1:22:55anonymity and it showed information.
1:22:59>> But and you know, it's true. But look at
1:23:00the immune system. It's so powerful,
1:23:02right? Because Yeah. It's okay. Some
1:23:04people were held accountable, right? Um
1:23:07but not everyone was, right? Not
1:23:10everyone was. There are still
1:23:12>> clearly that didn't happen, but I think
1:23:15the general public now sees what's going
1:23:18on behind closed doors. It's out in the
1:23:20open now.
1:23:21>> I think it's a positive. Um it's
1:23:23definitely a green flag. I think people
1:23:25are are getting exhausted, right?
1:23:28There's there's exhaustion bubbling.
1:23:31>> Yeah. Um, and so that's that is a good
1:23:33point. You know, prevent extreme
1:23:35concentration of rule writing power,
1:23:38>> right? When you win the game, you
1:23:40shouldn't have all the power to rewrite
1:23:41the rules, right? Um, reduce unnecessary
1:23:45so unnecessary dependency and exit cost,
1:23:47right? What I mean by that, here's an
1:23:49example, right? You know, what if we
1:23:52lived in a world where people, again,
1:23:55most people live in procarity, right?
1:23:57They have two or three jobs. Um, some of
1:24:00them do gig work, some of them are Uber
1:24:02delivery people, right? Think about
1:24:04those dynamics. They're completely
1:24:05dependent on the system and those jobs
1:24:07and they do them 15, 18 hours a day. And
1:24:10if they lose those jobs, they experience
1:24:12through the dependencies that form in
1:24:14the institutional level, they face
1:24:16cascading risk. And so imagine a world
1:24:19where you lose your job and you don't
1:24:22have to worry about continually paying
1:24:24your bills because nothing pauses,
1:24:27right? you lose your job, it wasn't your
1:24:28fault necessarily. It's just the system
1:24:30kicking in. Um, and you know, it's funny
1:24:33like if you ever read On Liberty by John
1:24:35Stewart Mill, he recognized the problem
1:24:37with concentrating power, the the the
1:24:40problem with inheritance. You know, tax
1:24:42the rich, right? Do these things.
1:24:45There's nothing wrong with that. Um, and
1:24:47then we're getting into a political
1:24:48thing, but at the end of the day, a
1:24:49political regime is is a symbolic
1:24:51system. So, um, and so wherever
1:24:54possible, the last one I would say is
1:24:56internalize costs that are currently
1:24:57being displaced onto other people,
1:25:00ecosystems, and future generations. Why,
1:25:03you know, stop dumping waste in the
1:25:05oceans. How about you dump it on the
1:25:07first floor of your company, right? Then
1:25:09you'd think twice about offloading costs
1:25:11or deferring the costs elsewhere, and
1:25:13you take responsibility for your
1:25:14psychopathic actions.
1:25:17>> Those five are great directions. And
1:25:20individually, uh, community, culturally,
1:25:23uh, politically, we probably need to
1:25:25head in that direction. But each one of
1:25:27those directions is a microcosm of the
1:25:30example of the two companies you gave
1:25:32earlier.
1:25:33>> Yeah.
1:25:34>> That the one that doesn't do those
1:25:36things will out compete and therefore
1:25:38have more power. What are your thoughts
1:25:40on that?
1:25:40>> Yeah. No, and that that's why I'm saying
1:25:42that this is where all of my research
1:25:45suggests because
1:25:48of the fact that the institutions um
1:25:52select for in instrumental strategies
1:25:54though and those strategies get encoded
1:25:56and because temporally the entire
1:25:58operating system of the world is now a
1:26:00neoliberal operating system for the most
1:26:03part um it's going to be a very
1:26:06challenging
1:26:07thing to reshape the game board because
1:26:11many of the objective functions, many
1:26:13many of the rules, the incentives, the
1:26:15legal frameworks, the accounting
1:26:16principles, the software components,
1:26:19they're crystallized, right? They're
1:26:21crystallized. And in a hyperco
1:26:23competitive environment where symbolic
1:26:25reward is winning inside the game. Well,
1:26:30and you have this loss dominant
1:26:31equilibrium where people just keep
1:26:33participating in the game because it's
1:26:34cheaper to stay than it is to leave. No
1:26:37one wants to be a martyr. What what if
1:26:39there's a a an abrupt um shock to the
1:26:43system?
1:26:44>> Yeah, that's what impacts usually when
1:26:46we see in collapse, right? You know,
1:26:47this collapse is nonlinear. Um typically
1:26:51what I think is really going to happen
1:26:53is that material reality is going to
1:26:56impose the constraints and there it's
1:26:58going to impose the change in the game
1:27:00board itself. But that could mean the
1:27:04instrumentalists are even more in
1:27:06control than they were. Or not
1:27:08necessarily.
1:27:09>> It can happen. Yes, it can happen. It it
1:27:12depends. What likely will happen is that
1:27:14our systems will remain symbolic with
1:27:16the same boundary conditions.
1:27:19And it will be on the communities to
1:27:21recognize that we shouldn't create these
1:27:23massive game boards. We need to create
1:27:25tightly coupled communities, smaller
1:27:27tribal communities like cooperatives and
1:27:30and things like that.
1:27:31>> This is all super fascinating. I mean, I
1:27:33I um have some time constraints uh and
1:27:37we're going to have to wrap it wrap it
1:27:39up, but I I'd love to have you back uh
1:27:41on especially on a round table on on um
1:27:46responses and thinking ahead on what to
1:27:48do, especially as you do uh more
1:27:50research on this. But we before we head
1:27:52to the closing questions, Greg, um if
1:27:55you're comfortable, I would like to ask
1:27:57you a bit about your indigenous
1:27:59background and and initiation.
1:28:02Specifically, I'm curious about how your
1:28:04indigenous experience and learnings, how
1:28:07they integrate with your formal
1:28:09mathematics background and how that is
1:28:12together shaped your worldview and and
1:28:14hopes for the future.
1:28:16>> Yeah, my my uh my dad is a Mohawk. He is
1:28:20uh uh Irakqua. Uh my grandmother uh was
1:28:25uh born and raised on uh reservation.
1:28:28I'm based in Canada. I'm based in
1:28:30Toronto and and she was born and raised
1:28:32um uh on Six Nations uh Brford uh
1:28:36reservation. So yeah, um that part of my
1:28:40identity is is is very strong. Um, uh, I
1:28:44think just being involved in community,
1:28:46um, in general, uh, I would say that it
1:28:50has had a significant influence on my
1:28:53life and I think the way that I see the
1:28:55world. Um, I think that in in native
1:28:58culture, um, you know, you think about,
1:29:00you know, seven generations planning or
1:29:03or things like that. Um, you know, um, I
1:29:06would say indigenous culture is very
1:29:08focused on nature, very focused on
1:29:10bonding. Um with uh you know uh you know
1:29:15we do something you know like
1:29:17thanksgiving we have a belief that um
1:29:21you know nature everything in nature is
1:29:23our are our relatives like the trees the
1:29:26plants the and there's a belief that our
1:29:28minds are one right and um you know one
1:29:31one thing that stands out to me is um
1:29:34you know I became a really big fan of
1:29:36John Trudeell um who's a very famous or
1:29:39infamous uhh a poet um a native poet and
1:29:44he he just was such an inspiration to me
1:29:46because he has such great poetry and
1:29:49he's a bit of an anarchist
1:29:51um to be honest and I I do um I I really
1:29:55love that about him and and there there
1:29:56were there was something that he said
1:29:58one time and actually I'm quoting a
1:30:00movie that he was in. It was a movie
1:30:01called Thunderheart that was made in
1:30:031992 but he basically said something
1:30:05like uh you know sometimes they have to
1:30:08kill us. they have to kill us because
1:30:10they can't break our spirit. We know the
1:30:14reality of freedom and we know the
1:30:16illusion of freedom. Uh there is a way
1:30:20to live with earth and there is a way
1:30:21not to live with earth. We choose the
1:30:24way of earth. And so that really just
1:30:26has always sat with me as as something
1:30:29that uh I I take with me in my
1:30:31day-to-day life. And do you have any
1:30:35corlaries between your work uh and
1:30:40psychopathy and instrumentalist behavior
1:30:43in the tribal setting either modern or
1:30:47historic? I mean I I just did a frankly
1:30:49where I brought up uh the concept of
1:30:51Watiko which I think is Alangquin and
1:30:54there are themes like the monster
1:30:56archetype and is there a relationship
1:30:59with all that?
1:31:00>> Yeah, there certainly is. I mean it's a
1:31:02very fascinating thing and I think it
1:31:04stems from the fact that you know I
1:31:06think a lot of indigenous cultures I
1:31:07don't want to speak on behalf of all
1:31:08indigenous cultures but I will say that
1:31:10like you know Irakqua and and and other
1:31:13cultures are are are very tribal um and
1:31:17very you know very much um respectful of
1:31:22of nature and things like that. And I
1:31:24think there is a deep I think a deep
1:31:27understanding in in the community around
1:31:29you know especially around
1:31:30colonialization and all of these things
1:31:34but there are certainly many parallels
1:31:36um made
1:31:36>> well I just wonder if if in today's
1:31:39tribes if there is an aversion or if
1:31:43there's strong reciprocity and repeated
1:31:46iterations with your your tribal members
1:31:49so that at the tribal level that
1:31:53psychopathy
1:31:54is is muzzled and squaltched.
1:31:57>> I think in the traditional sense it is,
1:31:59right? It's it's always there. There's a
1:32:03there's still the effects of
1:32:04colonialization, right? There's the
1:32:06intergenerational trauma uh for a lot of
1:32:09natives um in North America. And I think
1:32:12there's there's a resist there's still
1:32:15the resistance of becoming incorporated
1:32:18as part of the machine. However, we see
1:32:21the effects of the machine on the on on
1:32:24on indigenous peoples in general, right?
1:32:27Like there's
1:32:28>> they're entrapped.
1:32:29>> They're trapped. They're trapped by the
1:32:31system. Um and they're subjugated to a
1:32:35lot of um racism, a lot of um a lot of
1:32:39negative effects of colonialization,
1:32:41alcoholism,
1:32:42um drug addiction, and again, it's the
1:32:45system that has even wrapped its
1:32:47tentacles around the tribal beliefs.
1:32:49They're losing their language, right?
1:32:50They've lost their language. They've
1:32:51lost their beliefs, their values, right?
1:32:54And and they're trying to retain that.
1:32:56And it's very important to them to do
1:32:57so. But you do see the tentacles
1:33:01entering those communities. And it's
1:33:04sad. Really, it really is sad.
1:33:06>> Thank you for for sharing that. Um, so I
1:33:10don't know if you've ever watched any of
1:33:11my episodes. Oh, you said you watched
1:33:13the Reed Malloy one, but I asked these
1:33:14questions uh these closing questions of
1:33:17all my guests. So, you're aware of the
1:33:19meta crisis um taking off your
1:33:23evolutionary biology game theory uh um
1:33:26psychopathy expert hat on or off. You
1:33:30can keep it on if you want. What advice
1:33:32do you have to someone watching this
1:33:34episode? Uh
1:33:37what can people do to help address the
1:33:39things discussed in this podcast or is
1:33:41it all up to politicians and leaders?
1:33:44>> No, it's not up to them. We outnumber
1:33:45them, right? So the bottom line is that
1:33:48you have the power to
1:33:50resist the game. You have the power to
1:33:52change the game. You have the
1:33:55opportunity to be a better person. You
1:33:57have an opportunity to be a pro-social
1:34:00creature in a game um that wasn't
1:34:04designed for cooperation and pro-social
1:34:07behavior. The the powerful thing about
1:34:09the game itself is that it it becomes
1:34:12very self-regulating. Domination doesn't
1:34:14feel like domination. And that's why it
1:34:16works so well. It's like a magic trick,
1:34:17right? Domination works doesn't work
1:34:19when it's experienced as domination. It
1:34:22it works when it's experienced as, you
1:34:24know, self-regulating, when failure
1:34:27feels personal, right? And so on. And I
1:34:29I give people the advice that don't be
1:34:31afraid. You've got nothing to lose,
1:34:34right? Um we can all work together to
1:34:37create a better planet. the the fact
1:34:39that what you just said really resonated
1:34:42with me that if it was clear that it was
1:34:45domination or psychopathy, it would
1:34:48maybe cease to work. There's there's a
1:34:50little bit of a magic trick there. Thank
1:34:52you for that. Do you teach or are you
1:34:54just a research professor?
1:34:56>> I used to teach way back in the day. I
1:34:59used to teach way back in the day, so I
1:35:01have experience doing it. Yeah. So what
1:35:02specific recommendations do you have for
1:35:04young humans uh in their teens or or 20s
1:35:08who become aware of um the economic and
1:35:11environmental constraints to the global
1:35:13economy?
1:35:14>> Yeah. One of the things I would say for
1:35:15for the younger generation, I have a you
1:35:17know I have a niece and a and a nephew
1:35:19that are are younger uh and a twin
1:35:21brother. And uh one of the things I I
1:35:23try to tell them is to not become so
1:35:25dependent on technology.
1:35:27um to think for yourself, to read books,
1:35:30to um you know reinforce good behaviors,
1:35:34right? We have we have the opportunity
1:35:35to teach our kids good things, right?
1:35:38The community can step in and really
1:35:40make an impact on on kids today, right?
1:35:42So, I think for the kids, it's really
1:35:45just kind of trying to keep them away
1:35:47from becoming entrapped by the game
1:35:48itself, right? trying to find their own
1:35:50way, having their own voice, um trying
1:35:52to be independent, but at the same time
1:35:56being very community focused. I think I
1:35:58think there I think that's one of the
1:36:00goals of the system in that magic trick
1:36:02is it's easier to dominate people who
1:36:05are atomized than it is to um dominate
1:36:09groups of people, right? And I think
1:36:12reinforcing community is so important uh
1:36:15for for kids, I think. and to learn
1:36:17self-love and to appreciate themselves.
1:36:20And uh again, don't lead with fear.
1:36:23Don't be afraid to be yourself.
1:36:25>> Thank you for that. Um what do you care
1:36:27most about in the world, Greg?
1:36:29>> I care about the world. I I know this is
1:36:32so cliche,
1:36:34but I really care about nature. Um I
1:36:37really care about people. I really care
1:36:39about animals. Um, and I just want to
1:36:43see a better world and I want to be a
1:36:45part of of that. But I don't I'm not
1:36:48looking for prestige. I'm not looking
1:36:50for a claim. I just want to be a cog in
1:36:52the wheel to help. And um I want to be
1:36:55part of that process um at the end of
1:36:57the day. And I'm not afraid to do that.
1:36:59I'm not afraid of the institutions. I'm
1:37:02not afraid of the elites. I'm not afraid
1:37:04of any of these people. I think we need
1:37:06to all work together because the time is
1:37:08now. There's no other time to act but
1:37:11now. Um, so that's what I really care
1:37:13about.
1:37:14>> That was a beautiful answer.
1:37:17If you could uh wave a magic wand,
1:37:20strange question, uh, and there was no
1:37:22personal recourse to your decision, what
1:37:24is one thing you would do to improve the
1:37:26future of humanity in the biosphere?
1:37:28>> Oh man, that is such a hard question to
1:37:31answer. [laughter and gasps]
1:37:32>> I think I would love to just go, you
1:37:34know, sometimes I think about this. So,
1:37:36I was talking to a good friend of mine
1:37:37and I said, "Wouldn't it be nice just to
1:37:38be able to like time travel and live as
1:37:41our ancestors lived off the land?" You
1:37:43know, I think I don't know. I maybe I'm
1:37:45romanticizing the whole thing, but I
1:37:47think that's what I want to do. I think
1:37:48it was Yeah, it was possibly harder. Um,
1:37:52but you know, was it I can't really know
1:37:54until I go back there to see what life
1:37:57was like. You know, I just there's
1:37:59something about that. Yeah. time
1:38:00traveling back then to avoid getting
1:38:03into the current predicament we're in
1:38:05right now is probably what I would want
1:38:06to do if I could.
1:38:07>> Well, okay. So then as such a time
1:38:09traveler, what would have been the
1:38:11things that you could have changed the
1:38:14uh incentive structures and the
1:38:16institutions from becoming psychopathic?
1:38:19>> Man, I don't know. It's so hard because
1:38:22it it just like I think when we look
1:38:24back on just, you know, again, I always
1:38:26think about this evolutionarily. I don't
1:38:28know if we could there was anything that
1:38:29we could do. I think we're here because
1:38:31we're supposed to be here.
1:38:32>> I sometimes get that feeling as well.
1:38:34>> Yeah. There's something deterministic
1:38:36about it, right? That's the way that's
1:38:38the vibe. I I kind of get
1:38:39>> deterministic to get to this point, but
1:38:41not necessarily deterministic going
1:38:43forward.
1:38:44>> I'm wishful thinking and I'm hopeful.
1:38:46So, this isn't all doom and gloom. I
1:38:48think we can still do great things.
1:38:50>> This has been eye opening and I learned
1:38:53some things as you predicted. Uh do you
1:38:55have any closing comments? uh or
1:38:57thoughts for the people watching and
1:38:59listening who feel in their gut uh what
1:39:02you've laid out is is uh directionally
1:39:05valid.
1:39:06>> In closing, I'm just grateful for all of
1:39:08the you know all of the um
1:39:12the attention the work's getting and I'm
1:39:13I'm eternally grateful for that and and
1:39:15I hope that it can help in in some way
1:39:18and uh and I just want to be part of the
1:39:20solution. Well, if if your work gets a
1:39:22lot of attention, a lot more people are
1:39:25going to look to you for a solution.
1:39:27>> And I would love to be part of that that
1:39:29that ecosystem that that's my my end
1:39:31goal is to be part of that community to
1:39:33to help, right? And I'm grateful to you,
1:39:36Nate, for for having me on on the show.
1:39:38I'm grateful to Zach and Bill and uh all
1:39:42of these great people who are just uh
1:39:44you know, pro-social.
1:39:46There's a lot of them out there. Thanks
1:39:48again for your work, Greg. and uh to be
1:39:51continued, my friend.
1:39:52>> Thank you. It's great to meet a new
1:39:53friend.
1:39:54>> If you'd like to learn [music] more
1:39:55about this episode, please visit the
1:39:57greatsimplification.com
1:39:59for references and show notes. From
1:40:02there, you can also join our Hyo
1:40:04community and subscribe [music] to our
1:40:06Substack newsletter. This show is hosted
1:40:08by me, Nate Hagens, edited by No
1:40:11Troublemakers [music] Media, and
1:40:13produced by Misty Stinit and Lizzie
1:40:15Serriani. Our production team also
1:40:18includes Leslie Batloots, Brady Hyen,
1:40:21Julia Maxwell, Gabriella Slayman, and
1:40:24Grace Brunel. Thank you for listening
1:40:27and we'll see you on the next episode.

About this transcript

This page contains the full transcript of The Rules of the Game Keep Us Trapped with Greg Elliott | TGS 231 by Nate Hagens, generated from the public captions YouTube serves with the video. The transcript has 15,829 words across 2,493 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.