hormetic effects when you're, you know, weightlifting.
weightlifting.
But there are a lot of cases where I
don't find a practical application of
anti-fragility. But I get why Taleb does
it like his entire investing strategy is
anti fragile. You know he as the system
collapses he makes more or it's just
volatility in the system he makes more
but I don't find as many examples. So
the concept of anti fragility doesn't
stick with me and even though I know
that's the book that most people think
has had the biggest impact on them and
he values it the most uh it doesn't work
for me. Uh so I don't read that book but
I'm always rereading a skin of the game
uh because that one has a lot of nuggets
that I can apply that that that I can
say oh that's the generalization of
something that I had noticed in my
specific life. So the minority rule is a
good example I keep seeing the minority
rule show everywhere. Um so you you do
you do need the practical applications
in your life you you have to do so even
to the extent today that I want to
become smarter and wiser that means I
have to work. Yeah, I was going to say
you're continuously injecting experience
and principle.
Correct. Yeah. You start with
experience. You start with specific
knowledge. Then the specific knowledge
turns into more generalized knowledge.
Then you can take that and turn it into
values which are you tie it into your
existing values. It improves your
judgment. And I think ultimately in this
age of infinite leverage, judgment is
the most important thing. like if you
knew um where to navigate, you know,
people will pay you for that. The
captain of a ship is chosen based on his
ability to get the team together and
tell them where we're going, inspire
them to go there. Um so, but where to go
is the most important part. If you have
two candidates for CEO of Apple, right,
most valuable company in the world
currently, maybe it's Nvidia, but
whatever. You have two candidates for
CEO of the company, and one is right 80%
of the time and one is right 85% of the
time, and the company's worth trillions
of dollars. Who are you going to put in
charge? You pay that guy right 85% of
the time a lot more like you'll pay him
billions of dollars more per year
because he's steering a multi- trillion
dollar ship and the direction matters
more than any other single thing and
then finally I would say that judgment
judgment is really important judgment
comes through experience and reflection
so you you experience things you have
honest reflection about those things and
you build your judgment and then at some
point your judgment becomes so good that
you cannot even explain it or articulate
it anymore. So, there's a point early on
where you don't know what to do and you
have to rationally think through what to
do and you take feedback. Eventually,
you get so good at making the decisions
that you don't go to other people. Yes,
you can get a little bit of feedback to
inform your judgment, but you're you
know that you're in the best position to
make up your mind. So, you exercise the
judgment and you can articulate to other
people why did I do that? But there
comes a point where your judgment is so
good that you can't even articulate it.
And that's when it's called taste,
right? When you're just like, I it
doesn't feel right to me, right? This is
just the way we should do it. This is
the way I want to do it, right? But at
that point, it's taste. And people who
have gotten to that point like the Rick
Rubin's of the world where they have
really good taste from the Steve Jobs of
the world. I think those are the people
who are the most creative and create the
greatest works of art and business
because they have really good taste.
Like I think I would guess that Musk for
example on SpaceX, he has taste and
there are a few key engineers around him
who have taste. Um in if you look at the
leading AI labs and you kind of see the
tweet threads from some of the
researchers there when they care to talk
about how they're choosing which
experiments to run, which ones not to
run, it boils down to taste. They have
taste about like I'm going to throw
thousands of GPUs at this for hundreds
of hours and spend lots of money and
probably wind up at nothing. But my
intuition, my taste tells me, my
developed intuition, which is my taste,
uh, tells me whether to try it out or not.
not.
Yeah. You had a tweet. It takes time to
develop your gut, but once it's
developed, don't listen to anything else.
else.
Exactly. Yeah. And this is just another
way of saying taste. It's your gut feel.
And it can be around people. Um, you
know, older people have very good
judgment about other people because the
one thing that we are always all gaining
experiences in human interaction. No
matter what we're doing, we're
interacting with other people. So, we're
building up experience about who to work
with and who to trust and who not to
trust and who not to work with. Um, and
that ends up your gut feel. And so, as
you get older, you got to trust your gut feel.
feel.
Does your investing now primarily feel
like taste?
Yes. Yeah. Almost entirely. I hate
articulating it. And and a lot of times,
you know, I I'll pass on things now.
It's not just taste on the company. It's
an understanding of my own tastes in the
sense of what I like and what I don't.
So there's a lot of companies now that I
don't invest in where I will I will I
will pass up on the investment because
I don't want to take a walk with the
founder. I didn't learn anything
or uh I am just genuinely not interested
in the category. I'm not curious about
it. I'm not going to stay up late
reading about or thinking about it. So
it's going to feel like a chore. Or it
could just be like, you know, I've got
one short life on this earth. Do I want
to be associated with this way of making
money? Probably not. There are other
ways to make money. So obviously that's
a post-wealth problem or post-wealth
taste. So it's not just taste about the
business. It's also an understanding of
what it is that I value, how I want to
spend my time, what I want my legacy to
look like, what am I going to be proud
of, what am I going to learn from. And
again, this is back to an effort to
being lazy. And lazy means not working.
And not working means not doing things
you don't want to do. If you want to do
it, it's not work. So an ideal life
would be where everything I am doing and
everything that I am associated with,
everyone that I'm associated with is
something that I would do anyway, you
know, almost for free and being with
people that I really enjoy being around
and learning from them. So for me,
that's people who are very capable. They
do things. Uh they're very low ego, so
you have to deal with all the nonsense
that comes with that. They're very
intelligent and uh so you're always
learning from them.
Is it actually ideal to be paid purely
for your judgment? That that's a thing
you've talked about, but as we're like
talking through this, it's like the the
it's very hard to extricate your labor
from your judgment at least to at some
ratio, right?
Yeah, that's right. You know, this is
why another tweet I had, they do
contradict each other if you want to
talk math where it's like, you know,
what you do and who you do it with is
more important than how hard you work,
right? But sorry, it started out work as
hard as you can, but you know what you
do and who you do are more important
than how hard you work. But you still
have to work as hard as you can. So it
shouldn't feel like work. You know, this
goes back to like it should feel like
play to you, but look like work to
others. That's really what you want to
be doing. That's idealized. You're not
going to be able to do that, you know,
the full time. You know, it's obviously
not just hard work. Like how many
companies is Elon running, right? The
corner, the guy running the corner
grocery store puts more hours in the
grocery store than Elon puts into any
one of his companies. So it's not just
purely about hard work. That said, you
don't get to be someone with great
judgment and great connections and great
capabilities without working hard and
you will not build up your knowledge
fast enough. It's when you're working
super hard, when you're really intensely
into something, that's when you actually
make the biggest breakthroughs. So I
remember back in college, I had this
computer science project where we had to
write a compiler. And a compiler is a
difficult thing to write. You know,
compiler takes a high level programming
language and breaks it down to lower and
lower level programming languages until
finally goes into a set of instructions
that the computer can understand and
writing compilers up there in CS. It's
like you know uh other things in that
level might be like writing an operating
system right a simple operating system.
So it's no joke. And I just remember
when I would go in the computer lab, it
would take me hours and hours and hours
just to load the problem into my head to
go through my own code from the previous
few days to kind of come back to speed
like, okay, this is what I meant over
here and this is the part where I'm
stuck on and this part over here is
janky and this is a patch and this thing
works over here and it's beautiful and
but it would just literally take me
hours just to get back into it if I've
been out of it for a little while. And
so then I found that the most productive
sessions that I had uh in front of that
computer writing that compiler were 24
to 36-hour sessions where I wouldn't
sleep. So I would just be up for 36
hours drinking Coca-Cas and just kind of
coding or just thinking and walking and
coding and thinking and walking and then
when I finally fell asleep the problem
would fall out of my head and it would
take me a few days to recover and then I
would have to reload the whole problem
back into my head. Now, that's an
exaggerated kind of example, but I think
in general, a lot of times a lot of
breakthroughs happen when you're deeply
into a problem. It just takes hard work
and time. Like even with my co-founder
of my new company, who is just an
amazing guy. He's brilliant, just one of
the smartest human beings I' ever met.
Super low ego, very pleasant, um, and
very hardworking. If he's awake, he's
working. He's probably working right now
while I'm goofing off recording this
with you. You know, I've even noticed
with him like we talk every day. We have
deep conversations. We catch up a lot.
And it'll be an hour into the
conversation when I'm already ready to
hang up and give up and we've already
covered all the topics and we're just
circling and shooting the that the
big breakthrough will come. The big
insight will come. We're like, "Ah, but
also this, oh yes, and now it all
connects together." It just takes time
for your brain to percolate through a
problem. And so the more time you devote
to a given problem, uh, you know, the
more likely you're to have a
breakthrough. It's like if you're trying
to solve a really hard problem, one
technique that I have is I want to load
it into my background subconscious. I
want my intuition working on it so I can
have that shower breakthrough 48 hours
later. And you can't put that on a
schedule. You can't time that. So what
you have to do is you have to so
intensely focus on the problem that your
whole mind absorbs it and even your
subconscious figures out okay this is
important. You're dreaming about it and
then on some unknown time scale you'll
have some insight or some creativity.
But you're not going to have that unless
you did the hard work of working through
all of it manually with your foreground
with your neoortex early on. So I do
think hard work is important. You can't
give it up. But um if it feels like
work, if you're forcing yourself to go
through the motions, then you're going
to lose because somebody who doesn't
feel like it's work is competing with
you and for them it's entertaining or
fun or at least fulfilling.
So then good judgment comes from a
combination of like spending so much
time on the explicit knowledge that it
becomes sort of implicit.
That's right. Yeah. You spend time on
the explicit and then it becomes
implicit and then there's a phase where
it's still it's implicit but you can
still articulate it. And you know this
is an analogy this is not strictly
correct but I would say there's a point
where you have to reason through it
every time using logic and that's basic
decision-m and there's a part where your
subconscious can kind of enter into it
and then I would say like now judgment
and developing taste and then I think
there's a point where your whole body
reacts to it right cuz there is wisdom
in the body there's knowledge in the
body and at that point it's just taste
and then you're done with that thing I
mean you're not done obviously it's
always going to improve but uh you're
you're top of the craft.
John Clees, the guy from Ming Python,
had a great quote about this that stuck
with me, which is um you simply have to
let your mind rest against the problem
in a friendly, persistent way.
That's right. That's actually really
well put. Yeah, he's right. Your mind
has to be involved and you need to focus
on it, but it has to rest against the
problem because that you're pushing on
it, but in in a way you're not going to
give up. And yeah, in a was it in a friendly
friendly
friendly persistent way.
Friendly and persistent. Yeah. You can't
give up and friendly so you don't give
up. Exactly. He's right.
Yeah. I Yeah. Him and uh Oglev have like
these really interesting ways of sort of
activating their subconscious around um
around problem solving.
Yeah. Everyone who I think is deeply
creative, both works really really hard
um but yet knows that it has to be
enjoyable otherwise you'll never get
there. Um there was a line by a painter
I think Pablo de Soros Sarasoto
something like that. Um I don't remember
the name but the the line was good and
he said you know for 10 years uh I've
painted seven days a week 16 hours a day
Does AI have judgment?
Uh no I don't think so. I wouldn't count
on AI to like exercise judgment. What AI
has is AI has incredible information
retrieval capability.
So it can crossorrelate all human
knowledge and if humans have figured out
before it can give you the conventional
correct answer which is going to be
correct in most cases for solved
problems. So you know if you're trying
to figure out like you know what is the
law on a specific point it can like read
all the legal texts and tell you what
the current interpretation of the law is
according to all the conventional
wisdom. But if you're looking to do
something creative or something brand
new that no one has yet figured out or
you're dealing with a highly complex
problem with a very specific situation,
um you don't have the AI make that
judgment call for you. So I don't think
of the AI necessarily as having
judgment, but I think of it as like the
ultimate leverage information retrieval
tool. Um also, humans have values. They
have points of view. They have binding
principles. They have things that
undergur the foundation of what they
believe to be true and what they believe
to not be true. And this is
idiosyncratic to the person. It's
adapted to the environment they're in.
The AI is taking as a one-sizefits-all
approach according to what's either in
textbooks or conventional literature or
what the data labelers labelers who are
labeling the data into the AI thought of
it. I think AI just looks like magic to
people because it just it's it's very
hard to wrap into your mind uh you know
what these incredibly huge data sets
operating in all these highly
mathematical dimensions are capable of.
I mean the amount of information they
can retrieve and crossorrelate is just
astounding. It's like if you ran a
Google search every time and you read
all top thousand results very carefully
and then you crossorrelated them and you
were able to stitch them together and
you would make a few mistakes along the
way just like the AI does. But you know
that level of research combined with I
do think it has some tremendous
calculation abilities because it can
write programs and it figures out um
some algorithms you know below the level
of human creativity but above a mere
calculator. So I think the AIS are
incredible shortcuts, but there's
shortcuts to giving you answers to solve
problems where you don't need a perfect
answer. So the one quick I heard was,
you know, AI is great when wrong answers
are okay. Like you're not going to die
because the answer is wrong. You're not
going to lose a lot of money because the
answer is wrong. Um, but for anything
creative or requiring judgment at the
edge, which is what you really get paid
for, you get paid for creativity. So the
AI lifts the boat for everybody. You
have an AI, everybody's an AI. you get
the AI answer, everybody gets the AI
answer. There's no there's no alpha.
There's no edge anymore. So, in that
sense, the the AI raises raises the uh
the tide. Uh and right now, you're going
to get an edge because most people
aren't using AI or aren't using it for,
you know, um a lot or they don't know
how to use it well enough for bleeding
edge problems. So, again, like any piece
of technology, um if you're early
adopter, if you leverage it well, you're
going to do extremely well. So this was
my quip on Twitter that um it's not that
AI is going to replace software
engineers that AI is going to let
software engineers replace everybody
else. And I stand by that because
software engineers not cuz they're
writing software because software
engineers are always using the latest
tools. They're just structured logical
systems thinkers who are trying to build
a system to solve a specific problem.
And now thanks to AI they can solve more
and more problems. Um they can make cars
drive themselves. They can make robots
walk around. they can make expert
systems or AI systems that will make
certain levels of decisions without
humans having to be in the loop. Um, and
that just gives software engineers
amazing leverage. So all the people
saying that programming is dead, go
learn art or the trades, they're idiots.
They're just completely wrong. And
software engineers are getting richer
and more powerful than ever. And the
most recent proof that just popped up
with Mark Zuckerberg paying hundred
billion dollar packages to recruit
individual machine learning engineers
because the most leveraged engineers are
the ones who are building these AI
systems and then the ones below them are
the ones who are using these AI systems
and then even below them is everybody's
affected by these engineers using these
AI systems.
Yeah. You had talked before about like
the the robot revolution is already
here. It's just packed into data centers
and that was 10 years ago.
That's right. Yeah. The robot revolution
has been here for a long time. There's
trillions of robots in the planet U, but
they're just packed into data centers.
They don't need legs and arms. They're
just computing.
But now there's AI and AI agents. And I
imagine, you know, the leverage that
somebody who's fluent with those tools
can get is orders of magnitude.
It's massive. I mean, now you can
program them speaking natural languages.
Now they can uh access natural language
databases. Um, now you can program them
just by pouring in huge amounts of data,
not even having to program them. They
program themselves. Then they program
for you. Uh but it still helps to be a
structured thinker. Like my most useful
classes when I was doing computer
science and physics were actually things
on you know like I love my courses on
computer hardware on computer networking
on statistics on a lot of the deeper
physics and not because I got to use
those directly but because it just
helped me with concepts. It just helped
me think in certain ways. So even if you
aren't writing networking code, knowing
how the computer is doing networking
underneath is incredibly useful because
then you know what's possible. U Steve
Jobs and his team, their genius that
allowed them to assemble things like the
iPhone was because they understood at a
very deep level what was possible, what
was technologically possible, what
wasn't the bleeding edge, what was
barely possible, what was almost
impossible, um and what was actually
impossible. So they knew the lines
between those different things and they
knew it in terms of the material
science. They knew it in terms of
manufacturing. They knew it in supply
chains and deliverability. They knew it
in costing. They knew it in computer
programming. They knew it in hardware.
They knew the electrical engineering.
They knew it in battery requirements.
They had space requirements and
bandwidth requirements. They had people
who understood all these different
pieces and could make the trade-offs to
assemble this perfect jewel like little
smartphone that all of a sudden was a
true personal computer in your pocket.
And that's what made the iPhone great. I
would argue the greatest product of the
modern age is still the iPhone. I mean,
I'm I'm in awe what they mesh doing in
AI. But if you just look at like the one
device that every human craves and would
not give up right now, at least this
moment in time in 2025, if you if if you
went to people and say you can have one
of anything you want. Uh but if you
don't take that thing, you can't have it
at all. They would all take the iPhone.
Um so, and the last product before that
that I think inspired that same level of
desire was probably the MacBook and
before that it was the Macintosh. So
kudos to Apple, but they were the best
product builders on the planet because
they understood all the hard trade-offs
deep down at the detail technical
levels. Even if you'd given Jobs and his
team in AI to do all the vibe coding,
you know, and all the vibe design, they
could not have built that good of a
product without knowing deep down what
is the abstraction that the AI was
hiding. Where was it making mistakes?
What did it not understand? Uh, and how
do all the pieces fit together? Because
they were truly creating something new.
Uh, and if you're creating something
new, you still need to understand
everything down to the deep deep detail
levels. It's just the AI at least as
currently structured in 2025. These
modern AIs, they will take the drudgery
out of it for you. If it's been done
before, then you know with a small
hallucination rate, they can probably do
it for you today. And they're getting
better and better. So eventually all the
stuff that's been done before, you can
have them do. But that's okay. You can't
get paid in any meaningful way for
things that have already been done
before. Nor should you want to be. It's
incredibly boring. One of the things
that you know a good entrepreneur will
tell you is they hate it when they're
doing the same thing day in day out.
That shows that there's a failure of
automation, a failure of imagination. Uh
so good entrepreneurs automate thing.
And that's why, by the way, there's no
entrepreneur I've met who says, "Oh, AI
is a bad thing." To an entrepreneur,
it's a tool. It's an opportunity.
Entrepreneurs are not scared of AI is
replacing them, right? No entrepreneur
is going to be replaced by an AI. They
might be replaced by another
entrepreneur who uses AI better. uh and
they need to get good with AI just like
they need to get good with any tool. Uh
entrepreneurs are not afraid of AI
replacing them any more than they were
afraid of uh MacBooks replacing them or
AirPods replacing them uh or
self-driving cars replacing them. It's
just an another opportunity. Um so I I
mean I my my vision and belief and hope
is that every human being wants to be
creative. Every human being wants to
control their own destiny. Every human
being wants to make new things. Every
human being wants to be engaged and
challenged. Um, and you know, maybe it's
too late for some people because they've
given up, right, cynically, but I think
certainly every child does. You know,
every child starts out super creative.
Every child starts out just wanting to
alter their reality around them and to
do new and delightful things. No child
wants drudgery. No child has given up
hope uh on day one. So, I I do think
that newer generations will take
advantage of these things. Yeah, I think
you know it's tempting for some people
to see entrepreneur as a fixed class of
people rather than something that
everyone could aspire to be. And you've
said before, you know, there are 7
billion people on the planet. I hope one
day there are 7 billion companies. And I
feel like a lot of the tools and
structure were moving in that direction.
Click on any text or timestamp to jump to that moment in the video
Share:
Most transcripts ready in under 5 seconds
One-Click Copy125+ LanguagesSearch ContentJump to Timestamps
Paste YouTube URL
Enter any YouTube video link to get the full transcript
Transcript Extraction Form
Most transcripts ready in under 5 seconds
Get Our Chrome Extension
Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.
Chrome Extension