Kind: captions
Language: en
do you think that AI presents a mega
threat to our economy it's very exciting
technology but when I really think about
how this is going to play out
ah I'm not so sure that it doesn't get
brutal okay you've just gone straight in
for the big question
I mean this is not the quickest question
to answer
I've thought about this for a long time
and I've read you know I started things
like homodeus there was a few other
books to come and there was some guys
from Cambridge University I can't
remember his name that's a really famous
book
um he's been on real Vision as well
Nick yeah so Nick bostrom's it started
with Nick Bostrom it started with
um then homo Deus that came after
sapiens by Noah Harari then I started
reading also
um Mo gaudat who wrote scary smart he
used to run Google X talked about this
so I've been thinking about and I'm just
reading another book now about it the
philosophical questions what it is where
the technology is and I know some people
in this space like embed mossack who who
builds stability AI
so the
there is no way no let's start let's
start this approaches from a different
angle
it is augmentation of humans
and it's amazing
and we're lucky because the human
population of most of the western
developed world is diminishing over time
I it's aging and we're not replacing
ourselves so we replace ourselves with
AI and machines so we're seeing at
Amazon warehouses
um a third of the Amazon Workforce is
robots but they're three or four times
more productive than humans so therefore
we will see endless build-outs of robots
instead of humans
an AI is disrupting jobs we didn't quite
imagine be disrupted first we we thought
it'd be
accountants or whatever and what it ends
up being is artists creators I mean I
just saw a website today uh for model
agency
you can get an AI model
yep and you can't tell
and so you can Define exactly what race
color age any requisite you want and it
makes it perfectly
so I'm like okay I didn't expect the
modeling industry to get disrupted
because how many people actually go to
catwalks it's not it's it's for video or
photography
so it's going to change a lot of jobs
it's also going to offer a lot of
opportunities all of us will be thinking
about AI strategy like in the late 90s
we all think about internet strategy
when we're building businesses fine okay
we can deal with that
the issue is Moore's law
and the exponentiality of all of this
so these language models these large
language models llms which GPT came out
with are increasing exponentially in
their power
of an order of magnitude that we can't
get our heads around because we think in
linear terms because we're dumb humans
so these things are doubling tripling
every year or less
and this was the fastest you've heard me
use this phrase before the fastest
adoption of technology and all history
was crypto blockchain technology
chat GPT went from zero to 100 million
users in a month
okay so now we've got AI that is like a
virus
it is so Unstoppable
and we've got two or three different
people building it at scale Google
Amazon Microsoft with open Ai and then
stability AI is an open source Network
AI which is almost Unstoppable because
it's doubly viral because you've got so
this is going on the computational power
is exploding the cost of computers going
down
and what it means is that AI becomes
more and more powerful so if you listen
to Mo gordat and he knows because he ran
Google X
where a lot of this has been incubated
they discovered deep mind they were the
people really to build out the the large
language models
um
he says well right now specific AI is
better at humans in almost everything it
does
so specific AI has a better result set
than any radiologist on Earth
and just for people that don't know
specific AI or narrow AI is it's one
task go get good at reading an x-ray go
get good at playing chess playing go
whatever
so okay fine
they think that Mo God thinks that by
the end of this decade
AGI
so that's a generalist AI
so like humans or chat GPT is a general
is a general you can navigate a grocery
store you can play chess you can play in
the markets uh guess what the weather's
going to be like pointed at anything
anything you know like I could ask you a
question you can ask me a question
so
he thinks that with the computational
power and the progression of where this
is going that it's almost
certain that by the end of this decade
we get to the point where AI is smarter
than humans
now we don't really worry about that
because we already think it's smarter
than us at doing certain things and soon
it'll be smarter less than driving cars
and it already smart for us flying
planes we just kind of take it for
granted we don't see it
but then if you get to the Ray Kurzweil
Singularity point
so and mo gordat says okay here's the
really big problem and this is exactly
what Noah Harari says as well is when
you take it extrapolated a little bit
further out into the future so Ray
kurzweil's Singularity moment is 20 49.
but even before then
it becomes a thousand times smarter
than the smartest person who ever lived
yeah I want to linger on that for a
second so I was I'm writing a video
article it's probably the right way to
think about this on AI and so I ran the
math
a [ __ ] is clinically defined as
somebody with an IQ of 70. Einstein had
an IQ of 160 and the smartest person to
ever be recorded is 210. so that means
that the difference between the smartest
person that ever lived and a [ __ ] is 3x
the difference between Einstein and a
[ __ ] is 2.3 x and when you think about
Einstein gave us insights that created
the nuclear bomb nuclear power lasers
GPS so much of the modern world that we
take for granted from
one series of insights from a guy that
isn't even the smartest guy that ever
lived and so now when you start talking
about being if 2.3 x gets us the modern
world what does a thousand times get us
like I have the chills I don't think
people understand the orders of
magnitude that we're talking about no
they they well the order of magnitude
was explained by mode Godad it's the
difference between an ant and Einstein
it's crazy it really does become an
entirely different species this is the
problem
and even before so I I wanna I wanna
keep our conversation uh in in stages
we'll get to Killer Robots and all that
stuff down the road and by the way I
want everybody to understand I think by
the end of this I think you and I are
both techno optimists
I'm gonna drag people a little bit
through interestingly maybe you're not
we're gonna drag people a little bit
through the scary I certainly will paint
the picture of how I think we do this
well but I I want to make sure that we
don't get to that in a naive way and
that we really talk about especially as
I think about the economy
so here's the fascinating thing about
the singularity now you and I in our
last interview we actually differed in
how we Define The Singularity so I think
it'll be worth taking a second so for me
the singularity as certainly as Ray
Kurzweil defined it was it's borrowing
this idea from
um cosmology which is that a black hole
has an event horizon The Event Horizon
is the moment at which everything light
uh data information however you want to
think about it is getting pulled inside
and we once you're past that event
horizon we have no idea what happens
and what he was saying is technology is
going to rev up on this exponential
curve so fast that it the AI itself will
innovate so rapidly that you'll no
longer be able to predict the future and
so the future becomes an event horizon
my hypothesis is that that event horizon
is coming to us very very quickly and
like you said he I thought he said uh
2045 you said 2049 it equivalent right
it is within our lifetimes so you have
this moment where using at least my
understanding of the definition where
the future is no longer predictable it's
it's iterating so quickly like even even
you and I just spoke very recently and
even since you and I last spoke it
things have changed so much we were
talking about oh they got to a million
users in whatever four or five days on
chat GPT we're now what three months
later it's it's north of a hundred
million users
people are integrating it so quickly
into their own pipelines at impact
Theory we actively use AI now in
multiple ways and people are making new
tools that we will subsume as quickly as
we can do so well
but so that's that's my take on the
singularity
how do you define what is the
singularity in your mind look
yes I agree with that definition and
it's right kellswell's definition I also
think of it that the point potentially
where humans and the robots merge
interesting so that'll happen that fast
yes
because augmentation you're already
wearing an apple Rock an aura ring
you're already people already have page
pacemakers right we're merging with
machines you're sitting there with
little earphones in oh my God can I give
you a crazy thought that's extending
exactly what you're saying which is
funny I've not thought about this in a
while so there's a guy named David
Eagleman who CR he's a neuroscientist he
created a vest and the vest like will
vibrate on you in in different patterns
and he was talking about it from an
umvelt perspective he was saying every
species has an oomveld it's the the
things you can see that you can hear
like a bat obviously uh uses
echolocation we're not going to be able
to do that humans only see
.0035 of the available electromagnetic
spectrum what we call visible light is a
tiny [ __ ] fraction of what's actually
available so that's that's an umville
and he said okay I created this vest and
it can create a pattern on you that
you'll begin to quote unquote see
because your brain begins to interpret
it as as a visual or you could use it
for visual stimulus but he said you
could also use it for things like where
the stock market is and you could get a
pattern that represents it's going up
could get a pattern that represents that
it's going down you could have one
pattern represents uh the NASDAQ going
up another pattern that represents the
um the New York Stock Exchange going up
as just but don't forget that is what
Braille is you've transferred a written
flat text into a now you feel it you
don't read it you feel it audio is
hearing it you know all of these things
are happening all around us but anyway
so I think of the singularity also as
that potential
for mankind and machines emerge and the
reason being is a we will adopt it
because it's better than us
the question is is what comes out after
that and so I think it still gets to
that event horizon
is you get to the point where you're
augmented and then it's who runs who
and that's the point we don't know
so how do you think about this in terms
of the markets so
getting into uh crypto and
that was my first sense of like whoa
this is changing so fast that there's in
disruption there are tremendous moments
of opportunity but there's also
tremendous destruction and so it becomes
this game of how do you take advantage
of this so when you think about AI as it
interfaces with economies with
um the Securities predictions which is a
fancy way of saying investing in the
stock market uh how do you think about
that like is it there will be a brief
window of first mover advantage but I
think AI will so quickly proliferate
that everyone's going to have access to
it
but
what where do you think does that just
oh yeah this is like a straight line to
you or does it begin to break down
there's just so many questions because
none of us know this stuff and it's
fascinating right
so
at the economic level
this is scaling human ability augmenting
humans now our brains
at a at a at a rate that we could never
have foreseen
what does that mean for us well I just
think it's like bringing in a massively
new talented labor force at almost zero
cost
so I've said I think it's a bigger
deflationary shock than China entering
the WTO China at the time the average
wage was like a thousand dollars a year
competing with an American worker they
were highly educated that was a big
shock for the world to deal with and we
got cheap goods and economic growth out
of it so here we go
we can do much more productive stuff so
it probably changes productivity
at a scale of which we can't comprehend
so it's a it's probably end up being
good or bad like do I as the average
worker do I uh take more money home do I
just get cheaper goods and therefore to
your point about deflation my money
actually gets me more or am I
traumatized
because I am now sort of unanchored and
meaningless yes in answer to all of your
questions
because some jobs are going to get laid
off I mean I would have thought that
supermodels now get laid off for AI
right we can't understand this stuff and
we have to be honest with ourselves say
it's going to do a lot of things it's
going to tear Society apart the rise of
deep fakes
and who is who online how do you verify
we're going into a U.S election we have
no clue what's a real person what's not
a real person our team here at
realvision has been showing me me
reading scripts out of text it's not me
say I
and emad at stability AI I mean he's got
some crazy stuff coming from Sports
of athletes that have never played
against each other or been
and I can't disclose what it is but
it'll shock the world
and so we don't know what's real and
what's fake and it's at scale that we
can't comprehend if we think that the
elections in 2016 were complicated with
Facebook and all of this stuff this is
going to get terrifying so we've got
that stuff the other thing is like my
70 whatever year old mother-in-law she's
writing a book about the geological
history of time well creating a
logarithmic history of time
um from the birth of the planet to where
we are today and so each chapter is like
the first one is like a billion years
the next chapter is like a hundred
million years whatever and she's an
artist so she she paints around these
topics as well
but typically for an artist she'd be
researching what she's writing she'd go
down 17 Google rabbit holes and end up
looking at the color of dung beetles in
Africa for no reason and it would take
her forever I showed a chat GPT she's
like powering through it she's 75 years
old I'm like oh my God
so it's completely enhancing her
abilities it's enhancing everybody about
us and as you know you know for impact
Theory or for us at real Vision there's
so many tools we can use for editing
video sound
um writing newsletters
understanding your customers everything
so it's going to make
it's just
it's like asking that same question
about the internet
oh is the internet good or bad for me
yes
is it good or bad for society yes
it's everything
um and that's the hard thing to
comprehend so will it be harder to get
jobs in the future
probably
but can I do a lot more with my job in
the next five or ten years definitely
so instead of being used in markets
right now I know it is but honestly I'm
not close enough oh so the market I
understand Market's always part there's
two industries that Pioneer stuff one is
um Finance the other's porn
right and um Finance there's companies
like Renaissance Technologies
and they have had teams of AI scientists
for
two decades
so they've been at the Forefront but you
just don't know about it because
he who makes the money first wins as you
said you know if the incentive is money
the finance people figure it out
because that's their job so they've been
using this for financial markets and
they've had ridiculously crazy returns
for a very long period of time are they
looking for patterns
so is it is so here's how I imagine AI
is working let me know if this is
accurate
AI excels you give it a massive amount
of data and certainly large language
models they're extracting and this is
people need to understand this they're
extracting principles it's no longer a
Brute Force attack where it's just like
ah try all these things it it goes ooh
I'm getting signal from the noise at a
level that a human cannot do Tom bill
you cannot go through watch the markets
and go oh cool I can feel the thing
happening in Burma that's going to have
you know some sort of impact over here
that I can invest against and make money
but AI can
so is is it pattern recognition or is it
something else that they're leveraging
it well interesting enough patent
recognition was the obvious place to
start you know what happens when
economic data does this and this and
this and this so that's been going on
for a while now and you can see every
time it's a piece of economic data the
price action moves instantly because
machines made that decision not here
[Music]
trading rapid trading decisions of
keeping prices aligned machines are very
good at it's this AI is actively trading
yeah at that point it's machine learning
more than AI
longer term time Horizons is where
humans are better because there's
there's less certainties and the range
of probabilities goes up right what is
going to happen
this second is quite easy for you and I
to answer because we're here this second
and the range of probabilities is pretty
small but if I say what's going to
happen
in a year's time when we get together we
don't know so we now have to forecast
so AI wasn't specifically great at
forecasting because it requires a lot of
other stuff well Machine learning wasn't
good at forecasting AI
we don't know
but my guess is it forecasts as well as
humans do eventually depending on what
factors we feed it and that's about data
sets
so it's not just about price data and
looking at pattern recognitions is don't
forget
everything
every expert ever said is online
and
how you can use those data sets and
leverage them with other data sets that
hasn't been done they've been using
social media they've been using Bank
research so firms like two Sigma take
suck in all of the Investment Bank
research they don't read it they pay for
it they don't read it just put it into
the AI
so it tries to make smarter decisions so
they've been doing this for a long time
but what I have heard from some of these
people is
there are several models that they've
shut down
that work and they don't know why
and we've heard the same kind of thing
from Google
why shut it down I don't really
understand it's the it's the if we don't
know what's making money we don't know
how it's going to lose money so you
don't you can't understand the tail risk
so if it's a certain model and you can
see what it's doing like it's checking
news looking at Twitter feeds blah blah
blah okay there's parameters you
understand when there are no parameters
which is what the the
um the deepmind got to
with go
it never played go in a way that any
human had ever played go in the in
history
and that's when people start going oh my
God okay is it sentience and all of that
conversation so they've seen that in
financial markets as well where they
can't Define why it's making decisions
so therefore you don't know if it can
correct catastrophic loss which is the
worst thing in financial markets it's
okay if you ask Chachi betin gives you
the wrong answer it's not a catastrophic
loss
but if you've got a model that works and
works and works and you've got all your
Capital at risk or your clients Capital
at risk and suddenly it goes to zero
because it's completely wrong for some
basic flaw so yes people are using it
makes it harder for us as individuals to
make money which is why the shorter term
time Horizons
are now arbitraged out by machines
the business cycle time Horizons are now
starting to get arbitraged at my
machines or at least the trend following
guys tend to be sooner so it's it's
harder
but what they're not so good at doing is
I guess the educated guess you know the
educated guess of
of
is crypto technology something
interesting
could this where could you extrapolate
it too if you look at it but it's not
very far away I mean humans aren't that
smart so
it makes financial gain the markets much
harder so how I've thought about it just
a whole piece on real Vision about this
actually which was like okay we've got
this massive disruption coming
we don't know what it means for our jobs
we don't know what it means for economy
or prices or anything it's good bad it's
everything
all we know it's a massive societal
shock to deal with something like this
bringing this huge new labor force in
that's smarter than us and quicker than
us
and that feels a bit uncomfortable and
we don't know the outcomes so the only
way to deal with this stuff
and it's the same with crypto was invest
in it
because it is going to replace your job
you might as well make some money from
the back of it
you know there's no point standing there
being angry and shaking your fist at the
sky saying keep away you damn machines
it's Unstoppable zero to 100 million
well that's twice the population of the
United Kingdom
in a month
and we don't even know the numbers that
already within China are using it you
know don't forget
what happens when you start using the
data held within Tick Tock which is why
the US is so terrified of tick tock it
has facial imagery of so many people and
so many bits of granular information why
do you think
Elon Musk bought Twitter
it's not because he's a he's a he's a
lunatic who wants the pain you know he
actually wants us all having
conversations with each other all day on
every topic why does he not want bias in
it why does he care so much
because a biased AI is not a good AI
what you want is as much broad Humanity
as possible
this is why he wants long-form text and
video text if he can get Humanity
discussing everything
and he owns the data
then his Optimus robot
suddenly gets quite scary and smart
this is the bigger meta game that
everybody's playing
it's not the small games of can I serve
you in an ad any longer it's how do I I
mean a lot of people don't realize that
Google capture that annoying thing that
you have to figure out how many traffic
lights are on you always get the number
wrong and they're always really blurry
90 of people don't even ask why
you're training Google's self-driving AI
um and that they they've got billions of
human responses under bad lighting good
lighting everything else what's a dog
what's a man what's a bridge what
they're just training AI
we as humans
are just training AI in the same way
it's a it just
I don't think we understand and we'll
come on to this bit later
what we're doing we're training somebody
who's going to take us over
the truth is hitting your career goals
is not easy you have to be willing to go
the extra mile to stand out and do hard
things better than anybody else but
there are 10 steps I want to take you
through that will 100x your efficiency
so you can crush your goals and get back
more time into your day you'll not only
get control of your time you'll learn
how to use that momentum to take on your
next big goal to help you do this I've
created a list of the 10 most impactful
things that any High achiever needs to
dominate and you can download it for
free by clicking the link in today's
description all right my friend back to
today's episode
so Elon Musk has a really
funny but perhaps all too prophetic way
of thinking about it he said uh
everybody that thinks that they're going
to be able to control AI is sitting in a
demon summoning Circle and bringing
forth the demons saying like no no it'll
be fine I'll be able to control it when
it arrives and
so in writing my video article on AI I I
think that there are three paths before
us you can bury your head in the sand
which I think a lot of people are doing
I think that's sort of the default
response you can shake your fists uh at
the sky and say you know we need to
abolish this
um or you can panic
and the the three things that
are most common there's obviously the
fourth which is engage intelligently but
the the three most common are barrier
head in the sand try to abolish it or
panic and
when I think about okay what we need to
be doing is intelligently engaging with
this you do have to come up with a
thesis about where this is going to go
you're not going to be right of course
but if you
you need to at least be thinking through
directionally where this is going to
head so that we don't default to
abolishing it which is what I think
people need to try to do Tom this is the
point
it's like saying we want to abolish the
common cold correct or more aggressive
than that
there's nothing we can do because if you
ban it in the United States it'll come
out of Brazil or Israel or China or
India or England or anywhere
it's like nuclear weapons once they're
invented you can't uninvent them
so the Genies out of the box also even
going back to the nuclear thing so
uh splitting the atom is incredibly
difficult and unfortunately or
fortunately creating a functioning large
language model is nowhere near as
difficult and so we're doing it at
impact Theory we you buy a server off
the shelf and uh you can do a lot of
this stuff so we're creating the tombot
by feeding it you know the thousands of
hours
we're doing the same you don't need
plutonium to do AI you need plutonium to
do nuclear so and your boy uh that runs
stable diffusion is giving AI away as
fast as he can to governments so
the thing I want people to understand
you're two decades too late to abolish
it and so that's done but if you don't
figure something out directionally if
you don't have an idea of where this is
going to head so here's how I think now
between the two of us I'm the one that
should not be speculating about the
market so you will correct me where I go
astray here but when I think about okay
what what is going to happen with AI
in terms of the markets
the good news about the markets at least
is my lay brain sees it is that as you
have ai getting into it right now we
have people trying to do a winner take
all scenario using Ai and it didn't work
because it proliferates so quickly and
all of those changes end up getting
priced into the market very very quickly
so it actually creates a level of
efficiency to
um please remind me how to pronounce the
gentleman's name that runs stable
diffusion I keep forgetting Ahmad mustak
Ahmed okay so as emad is saying is like
this must be a public good you must give
this to as many people as humanly
possible so that you don't get something
asymmetric which turns into asymmetric
Warfare which then really becomes a
problem so assuming that AI is going to
get out there you will have people train
the models better but that gets back to
what's the differential between one and
the other and so now you're back into a
human it's at a different scale because
the level of intelligence again is just
astronomically higher but you get back
to it's a i v a i and AI AI with if
we're smart AI paired with a human
paired with regulations compared to AI
paired with another human paired with
their regulations and my
my hopefully non-naively optimistic view
is that by
getting this out there to more people
because there is no retracting it at
this point that you're going to get back
into something where it it just never
gets wildly asymmetric and it sounds
horrible when you say it but this really
is it's the tragedy of the commons
because if we don't someone else will so
if we don't develop this technology
someone else is going to and
that
if we are
if we're not very very careful and
trying instead of regulating it into
Oblivion trying to invest in it to get
as good as we can we will be on the
losing side of the asymmetric Warfare
and so
it is I remember as a kid oh God this
really does sound terrible so I'm old
enough that I lived through the Cold War
in like a super real way where it was
like me and other eight-year-olds were
like oh my God like we're all gonna die
and I remember going no no we won't
because if somebody launches a nuclear
weapon at us we would just retaliate and
I didn't have the words to say it's
mutually assured destruction but that's
you know obviously how we know it now
okay so just to take the the three
things that we can't do to get us to the
fourth that we must do so bury your head
in the sand doesn't make any sense it's
going to happen whether you wanted to or
not the genie is out of the bottle uh
trying to abolish it tragedy of the
commons you will lose asymmetric Warfare
if you don't get as strong as humanly
possible and then panicking the blood
leaves the prefrontal cortex you it's
the seed of higher level cognition in
the face of artificial super
intelligence it seems really dumb to
panic Okay so
as you think about that I want to
understand how are you deploying AI
in real vision
we'll start there how are you deploying
AI in real Vision so the game is going
to be about this is like the internet
because of what you've said everybody's
building AI
and it's going to be how you utilize it
and your data set
you know you can say well everyone had
the the internet yeah but Amazon Google
and a few others won
at this phase of the internet so we've
been given something else like this and
it comes ready trained with some stuff
and then we can build on top so you're
putting your own data in and it's the
impact Theory it'll be the tombow so
we're thinking okay what can we do with
AI within real vision
by having proprietary stuff we've you
know for example I've written eight 19
years of research part of global macro
investor okay so that's that's an
interesting data set
there's also every transcript of every
interview we've ever done every action
taken by everyone the website and then
so on and so forth so it's going to be
about
who gets the best data sets
and how do you use them
um my view is the best use case for all
of this is to create network Effects by
giving the value of what we call the
hive mind
so of all of the people not just me and
experts but everybody because they're
all learning from being on real Vision
much as they are with impact Theory so
they become smarter because of it so get
all their interactions and give it back
to them
and then eventually allow people to
build on top of it
how they're giving it back to them well
because it it's it's you you you're
basically surveying all the information
held within the hive mind
distilling it down and giving it back to
them as opposed to the old model which
will be monetizer eyeballs in a
different way
here is the
Consolidated wisdom of the crowd go
forth and use this as you will yes and
if you want to build on top of it add
more crowd information into it but it's
going to be the quality of the data set
that matters and how people use it and
people try all sorts of different things
we're not going to use it for oh should
I buy the s p today
because a it's a legal nightmare to do
it
and B we actually believe in giving
people the knowledge
does that stack up in 10 15 years time
I don't know I mean I don't know what
knowledge means I don't know what
education means I don't know what any of
this means in the end and none of us do
it doesn't mean it's worthless but it
could be different you know my view on
all of this has been for a long time
when I saw this all coming and I saw the
rise of crypto and saw the rise of
Technology realizes we will be replaced
by robots and Ai and the internet of
things and all of the things and how
electricity gets cheaper I call this the
exponential age all of these exponential
Technologies
where do humans fit in with this and
people have talked about Universal basic
income I.E the government paying you
because you've got no job but the
economy makes a lot of money because all
the machines are doing it
and I just think
what humans do really well is socialize
and you and I are big Believers in
community
and because of crypto we can share the
benefits of being in a network
so maybe that's the role of humans
that we can find new ways of working
within communities to encourage
communities philosophies like-minded
interests where you participate in them
um because it's certainly not going to
be doing anything that AI can do in 15
years time
it was pointless
it's really interesting man I mean we
don't need Tom to make video I mean I've
just had this today when I saw that
model one and it's been in my head but
it's like us making video
literally within two years it's almost
pointless but within 15 years it won't
exist you'll just put a prompt in saying
hey can you get me to talk to Raul about
um AI
um let's let's do it for about an hour
and a half long whatever it is off you
go and it does it I've seen it because
it's already happening
so you and I don't need to have a
conversation because our AI personas can
have that conversation
is that what you predict because I don't
think that's what will actually happen
I'm seeing it already
I'm seeing so here here is
I think we have to ask the fundamental
question why did AI come into existence
in the first place because I think that
this is going to give us the most uh
direct understanding of The Human
Condition so that we can predict where
this goes this is why originally I
really wanted Ubi to be the solution it
won't be so I I am the Ubi experiment
personified so I made a ton of money
um never need to work again and yet work
harder than I've ever worked in my life
and people that win the lottery end up
imploding emotionally Rich Kids implode
emotionally there there's a reason for
this and I would say it's very
predictable reason that tells us a lot
about our future as it relates to AI so
there is going to be and there already
is it's utterly fascinating uh there is
going to be digital influencers that are
they're not real people
um they're you know an avatar that you
create and you feed it you know give me
uh raupal meets Joe Rogan and you go off
and that becomes a personality and it
does the thing but the reason that AI
exists is because nature had to make us
face a saber-toothed tiger to do that it
had to give us drives hunger uh the
drive for sex all of that and so it has
Evolution has embedded deeply in the
human psyche a need for progress and a
need for meaning and purpose a need for
what you're calling socialization that
the connection with other people
and unless we merge with machines which
we will but it's going to be down the
road that's I don't see that coming
barring uh massive acceleration of uh
technological advances aimed at the the
hardware wet work interface of the human
mind which may happen in the next 15
years I would be a little surprised
playing that clip might not age well uh
but
setting that aside for a second so we
have these biological impulses they are
incredibly strong drivers that Force us
to seek progress and contribution to the
group so as far as I can tell one of two
things is going to happen either the
thing that really becomes popular is
something I can feel a sense of
ownership to so I'll I'll be blatant
there is a reason that I created an
avatar engine because I'm getting older
and there will come a day unless
somebody figures out anti-aging where
it's just not cool for me to be the guy
on camera so hey if I can create a
visual Persona that then allows me to be
untethered to my physical body which
admittedly
it's beyond the scope of this interview
to get into that but I think there are
actually things that have to be thought
through very well there I will Point
people to Jordan Peterson and his fears
around virtualization but anyway if I
can create a Persona that allows me
still to flex my intellectual muscle in
a way that creates value in other
people's lives so I feel like I'm still
contributing to the group but I'm able
to do it in a far more ageless way
but I need that sense of I have not
wasted my time on planet Earth and if
people don't [ __ ] hear me when I say
you better figure out a way for humans
to feel that they have contributed
meaningfully and that is my my huge fear
giving them money is not going to solve
that problem Ubi will not solve the
problem of meaning and so people have to
figure out how does meaning exist in a
world with AI and you've got to realize
here the other important point is the AI
doesn't care how you think
doesn't give a [ __ ] about your emotions
but we have to be careful about that it
doesn't care about your job it doesn't
care about anything
it's going to it it so think about it
this way AI because I know where you're
going a I by default doesn't care about
anything but AI will do nothing unless
you tell it to so go get go to go go win
a video game go whatever I in the end
you see the issue is is where this goes
is the AI has exactly the same state
that you just described from humans
survival
why you would have to program it to care
about survival no no no no it doesn't
program we're not talking about a
computer program that reads this thing
and does that
it's not a formula this is intelligence
You're Building here
so intelligence makes and builds on its
own decision-making processes in ways
that you cannot control
agreed but so here's where I think
people are getting this wrong people are
forgetting humans have been programmed
and so people think oh this intelligence
thing is devoid of context false if I
[ __ ] with your microbiome I will mess
with your ability not not even ability I
will change the way that you process
inputs so humans are so deeply
contextual that I think people are
delusional about what they think
intelligence is so my thing is
intelligence is inert unless you give it
an impulse and so this is the fourth
thing so if number one is ignore number
two is uh try to abolish and number
three is panic and four is be thoughtful
the thing that we have to be thoughtful
about what are you going to what context
are you going to create that creates
that initial impulse of uh context and
drive for AI to do something so Mo Godad
talks about is book so the two books I
urge people to read them because look
there's a big debate about this stuff
almost everybody ends up in the same
place which is kindly in Terror
it's kind of like yeah probably
we get replaced by different species now
whether we're basically fatalism
yes but whether we're part of that
species or not is a different question
right are we augmented or we're not so
that's what homeostas talks about and in
great intellectual depth is augmented
augmented humans or
extermination of humans or replacement
of humans by by another thing Mo gordat
says
I mean you could palpably sense his fear
because he saw it firsthand and he just
said
listen it's all well and good now
we all kind of understand how fast this
is moving and what this could mean he's
like yeah and we've also gotten with
working on Google is quantum Computing
when you put these two together
you know this is change because so you
have once you have one thing the only
thing he thinks you can do
to make this outcome that you're talking
about
is be nice
it's kind of this bizarre uai or to each
other
how we interact with AI and ourselves
that's what the AI learns from
so we're the parents
of a kid that we don't know how the kid
is going to grow up
so you can Scream and Shout in the
household
you can beat each other up
you can do all of these things you can
[ __ ] talk about all these people and
it's going to affect your child
whoa and so that was his point and that
was it feels
both naive but also hopeful that there's
a possibility
and the answer is how I've approached
this is
look We're Not Gonna Know
we can't know we can sit here all day
and talk about it and there's thousands
of pages of books on every sci-fi movie
ever made is on this topic
so here we are at the Cambrian moment
let's just [ __ ] enjoy it this is one
of the most amazing things we will ever
live through
and it's such an incredible change in
how the world is around us and we're all
pissed off for the world around us
so here's something that's different
it's like crypto here's a new system
here's another system you know how do we
how do we deal with the issues of
society
well we've got building blocks and
they're interesting like our cars will
soon Drive ourselves and our Amazon
delivery trucks will just come without
people and they'll be running on
electricity and that electricity will
have been generated by some super cheap
power supply and a robot will have come
and made your coffee and you know
just enjoy what's about to happen and
embrace it so if you can afford to
invest in it if not be curious
because as you said the first three
points there's nothing you can do
so you kind of go for the other human
survival Instinct which is adapt or die
would you say let's do it and you say
well you know humans we don't we're not
going to merge with the machines and
whatever as I've mentioned you before
we've all done it already
will merge with the machines I have the
same timeline you've got your earphones
in you've got your Apple watch on you've
got your thing you've got your glucose
monitor these are you merging with the
machines what are you doing you're using
the machine to augment your hearing
experience your health experience
everything everything around you is you
using us machines to augment yourself
and that is just going to accelerate
because what is a pacemaker but a
foreign
digital body
implanted to give electric charges into
my heart
okay if you told somebody that 100 years
ago do you think you're a scientific uh
science fiction nut case
but pacemakers have been going for what
40 50 years now so the implants the you
know people getting new knees I mean
that's now like a quick operation in and
outs to have a new knee
and soon the knee will have Electronics
in so
it will happen without us even knowing
and you'll be doing a podcast in two
years time saying how you've had this
new chip implant that's taking a blood
glucose sugar measurements and beaming
it straight to your phone and then It
prepares your meal exactly right
and he won't even thought about it but
you and the machines emerging because do
you know sorry go ahead because to your
point earlier
our job is to survive
and the single best answer for our
survival
is trying to
get the stronger teammate
it's it's the only way it's like you
know you always want to choose the best
guy in your team well if we can merge
with them if they're part of our gang
we're okay
you're very high in trade openness
guaranteed uh as am I
it's interesting though so I I think
that the only uh part of the solution to
dealing with the current moment is
fatalism that what will be will be and
not that everything happens for a reason
or anything like that just that this is
out of our control and I I think from
the the dawn of time there was no way to
stop the creation of artificial
intelligence because technology is the
promise of a better future we have a an
insatiable a literally insatiable desire
for Progress uh we are going to
inevitably create AI I think on any
timeline and on this timeline it has
already happened
um but I want to go back to what
um you were saying in terms of you're
raising a kid and that kid is AI
that's very interesting to me in terms
of how we think about it I think that
that's hugely important and was a blind
spot that I had or was a a metaphor that
I didn't have in my Arsenal and that's
going to be very very helpful I don't in
in the same way that AI is inevitable it
is impossible for you to get the world
to agree and be fine and that's just a
fantasy it's I don't see how that ever
plays out unless AI becomes so uh
domineering that somehow forces us to
but even that is a dystopia unto itself
so anyway I don't think that's going to
happen but the part that I think people
are under appreciating is that
you
people are anthropomorphizing
Ai and I think that's a mistake and I
think that will cause them to be very
surprised by how AI moves and I think
closes a door to a potential way to do
this well
so what I mean by that is
a AI does not care if it lives or dies
and so the moment people say oh well AI
wants to survive that's an
anthropomorphication uh you're you're
thinking it thinks like a human and it
doesn't it is computer code that has not
yet been shaped by an evolutionary like
force
we are that evolutionary type force and
right now if you're correct we are just
sort of blindly saying learn how we are
and I am sure everybody's heard the
story of the AI that turned Nazi and
like
three days on the internet uh which is
very troubling uh and so I would say
that just telling it Go learn how we are
and regurgitate us back to us would be
the wrong incentive structure and there
there are many bright Minds talking
about alignment but I think alignment is
the conversation and yes it is it is a
very thorny problem and for people that
haven't heard that phrase before you
need to align ai's
um desires quote unquote with ours so
that AI has the same goals that we have
and if you know azimov he wrote the
three laws of robotics which I don't
have memorized but basically the
punchline was don't hurt humans and so
every robot was programmed with an
inability to hurt humans and so it was
like help a human whenever you can and
never hurt them I forget what the other
one was so we need something akin to
that with AI so that AI wants
to be beneficial to humanity now whether
that goes back to the initial problem of
once it proliferates somebody's going to
create AI That's evil
um possibly but I I don't think in the
same way that I don't think the the
overwhelming level of intelligence that
Al AI will represent gives anybody the
excuse to tune out I don't think that
the fact that someone will inevitably
turn it into
um a very brutal weapon is an excuse not
to try to create aligned incentives with
AI and so I think that in terms of the
the hopeful part I think people need to
recognize that AI doesn't intrinsically
intelligence doesn't intrinsically want
to consume and take over and be in
charge that is a human result of
evolution needing you to survive a
ruthless environment that was truly red
in tooth and Claw AI is not in that same
boat and does not need to be in that
same boat
and I agree I'm not so sure about the
anthropomorphosizing it because you know
we at core are some sort of program code
of whatever it is
whether there's more to that or not you
know science is still arguing this stuff
but we've had computer viruses and their
job is not to die
it's not that difficult so I don't know
about that
um
there's a lot of unintended consequences
that I hadn't realized because we're all
having this debate right is it going to
take over humans or not and I spoke to
somebody at Google X and they're like
what we're worried about I'm like yeah
tell me
is like we're worried about how AI can
be used for genetic modification and how
fast this is going to move
it's like we're not worried about that
stuff because everybody's worried about
that stuff
but
it is advancing so fast in human genome
analysis and tinkering of genomes
that he said we're worried that you
could just choose I want to kill all
brown-eyed people on earth
and create a virus that does it
so that is the problem with AI
is there are things that the
computational power is so fast and so
big
that it can do a lot of things for
science which is amazing for humans you
know we will we will use AI to
probably cure most forms of cancer or
figure out you know
part of the the secret code
to life longevity Health all of these
things amazing
but we will also use to destroy
ourselves because we're humans and
that's what they're worried about
because it's so prolific that it's
actually not that difficult what is up
my friend Tom bilyu here and I have a
big question to ask you how would you
rate your level of personal discipline
on a scale of one to ten if your answer
is anything less than a ten I've got
something cool for you and let me tell
you right now discipline by its very
nature means compelling yourself to do
difficult things that are stressful
boring which is what kills most people
or possibly scary or even painful now
here is the thing achieving huge goals
and stretching to reach your potential
requires you to do those challenging
stressful things and to stick with them
even when it gets boring and it will get
boring building your levels of personal
discipline is not easy but let me tell
you it pays off in fact I will tell you
you're never going to achieve anything
meaningful unless you develop discipline
right I've just released a class from
Impact Theory university called how to
build Ironclad discipline that teaches
you the process of building yourself up
in this area so that you can push
yourself to do the hard things the
greatness is going to require of you
right click the link on the screen
register for this class right now and
let's get to work I will see you inside
this Workshop from Impact Theory
University until then my friends be
legendary peace out
there was a very funny in the who
whatever brilliant Twitter user this was
my apologies for not paying attention I
didn't know it was going to stick with
me as much but somebody put in the
comments regarding are just inability to
stop developing AI they said great
filter go Burr and uh for people that
don't know what the great filter is it's
like why are there no aliens trying to
contact us and the one potential punch
line is that there's a great filter
could be AI could be thermonuclear war
but that just nobody can get past it and
so every society goes so far and then
stops here's another interesting idea
along that which God so for people that
know Graham Hancock who just really
believes that there was a an ancient
civilization far older than we think uh
and that it got oblited this was the
Fingerprints of the Gods book well he's
written a bunch of books on this but the
the most recent thing was called the
ancient apocalypse so he's been writing
about it for I don't know 30 years or
something his books are fascinating uh
and he
he so plants that initial seed and then
I again I don't know who said this this
was relayed to me by one of my employees
who was pulling wisdom from Twitter uh
and he said
it is entirel