00:00:00
I gotta wrap, guys. I got to catch a
00:00:01
flight to Miami. Let me do a closing
00:00:03
here. If you want to keep going, you're
00:00:04
welcome. Two. Three. Two. The plane.
00:00:05
Just wait. Just text the pilot and just
00:00:07
tell them you're all right. Listen. I'm
00:00:09
not burning all the allin credits, so to
00:00:12
speak, and all of our tokens. I'm
00:00:14
kidding. I'm kidding. I'm kidding. I'm
00:00:16
not flying private to everything and
00:00:19
then putting it on the all-in budget.
00:00:21
The rest of us are flying Southwest
00:00:23
for
00:00:24
your dictator Jim Pol. It's a strange
00:00:28
concept. Yeah. David S. I What does that
00:00:30
mean, Dave? When's the last time you
00:00:31
flew a commercial? Clinton, I haven't
00:00:34
missed a flight in about 15 years.
00:00:37
[Music]
00:00:39
Let your winners ride.
00:00:42
[Music]
00:00:46
We open sourced it to the fans and
00:00:48
they've just gone crazy with it.
00:00:54
All right, everybody. Welcome back to
00:00:55
the number one podcast in the world.
00:00:58
We're back. We're back and what an
00:01:00
amazing panel we have today with us.
00:01:03
Ryan Peterson, friend of the pod, is
00:01:05
back on the show. He's the CEO of
00:01:07
Flexport. How are you doing, Ryan? Did
00:01:09
you get any skiing in this year? I know
00:01:11
you like to ski in the deep powder like
00:01:13
I tried, man, but it was a busy year for
00:01:16
work and I got two little kids. I I did
00:01:18
a few days. Okay. So, you're Oh, yes. We
00:01:20
all forgot you gave control of your
00:01:23
company to somebody. It got a little uh
00:01:26
shaky, got a little contentious, and
00:01:28
then you took the reigns back. How's it
00:01:30
been being back in the pilot seat? Oh,
00:01:33
that was a year and a half ago. So, it's
00:01:35
a distant memory for in in flexport
00:01:37
time. That's like a decade. We've uh
00:01:39
yeah, really had an amazing run.
00:01:41
Although, these tariffs, I mean, I guess
00:01:42
that's why you guys invited me on. These
00:01:44
tariffs have kind of made a lot of
00:01:45
created a lot of new uncertainty in the
00:01:47
flex sports world. Okay, so we'll
00:01:49
definitely get into that. and of course
00:01:51
fan favorite back for his fourth
00:01:53
appearance on the pod. I I I was so
00:01:56
first of all I saw the comment I I saw
00:01:57
the comments last time I was on. I'm I'm
00:01:59
officially not a fan favorite but uh
00:02:01
glad to be back on and I will be I will
00:02:03
be representing free markets uh in uh in
00:02:06
this uh uh in this version. What do the
00:02:09
comments say about you? It was like uh
00:02:11
you know like uh loves Biden uh you know
00:02:16
to totally beta you know all the uh
00:02:19
lover beat soy boy you're filling in for
00:02:23
me though. Um I think so. Yes. I was
00:02:25
trying to represent uh I was trying to
00:02:27
represent libertarian values at the
00:02:28
time. But uh but I love this that the
00:02:30
leftists are embracing Milton Freeman. I
00:02:33
think it's all worth it if that's what
00:02:34
comes out of all of AOC AOC is going to
00:02:37
be a complete free market uh person
00:02:39
soon. Yeah, free market months are
00:02:41
coming soon. Embracing free market
00:02:42
values and the stock market, right? Yes.
00:02:44
Exactly. Because any decline in the
00:02:46
stock market is Trump's fault. So now
00:02:48
they're they're embracing the stock
00:02:49
market. Well, unfortunately
00:02:50
unfortunately the one day that he said
00:02:52
it's Biden's market, it was a it was a
00:02:53
green day. So that uh that didn't help
00:02:55
the case. I mean gosh uh well listen,
00:02:57
it's another massive green day already.
00:03:02
All that matters to me is that Uber is
00:03:04
the anti-tariff stock. It just does
00:03:07
great. It's not impacted by tariffs. So
00:03:08
here we go. with Chimoth. It started
00:03:10
already. We have TDS on both sides.
00:03:12
We've got Trump derangement syndrome
00:03:14
from Aaron. Wait, no, no, no, no. I want
00:03:16
to be defender syndrome from Sachs.
00:03:18
We've got both. Defender and
00:03:20
derangement. Here we go. Syndrome. Trump
00:03:23
bias syndrome on your part. Who? Me? Me?
00:03:25
I call balls and strikes. What are you
00:03:27
talking about? Let's get started. It's
00:03:30
starting already, folks. It's going to
00:03:31
be a great episode. Lots of excitement
00:03:34
with us again. Jason has the uh the rain
00:03:37
self-sabotage. Find every way to not get
00:03:39
rich syndrome. I do. What are you
00:03:42
talking about? You guys said you're buy
00:03:43
me out of this thing and I can get the
00:03:45
hell out of here. You know how much my
00:03:46
shares are in all in worth? For the love
00:03:49
of God, write a check.
00:03:52
Get me the hell out of here. I just may.
00:03:55
Oh god. I mean, I'm going to be a
00:03:57
terror. If Uber breaks 88, that's my
00:04:00
number. 88 is the number. You're all
00:04:03
when that happens. Uh, and we're getting
00:04:04
close. All right, let's get started
00:04:07
here. We have so many topics to get
00:04:10
through with us again. David Saxs. Hey
00:04:13
David, you're doing uh more episodes
00:04:15
now. The audience wants to know. I don't
00:04:18
know if we're allowed to make any
00:04:19
initial announcements, but people are
00:04:20
asking me on the streets, in the
00:04:22
airports, in the comment threads. Is
00:04:24
Sachs back?
00:04:26
Well, the ratings are back ever since I
00:04:28
came back to the show. That's for sure.
00:04:31
The ratings are back, show, but is
00:04:34
measurable back? Is Sachs back?
00:04:37
Well, I'm back as much as I can. Mhm.
00:04:40
And you are a partial employee of the
00:04:42
government. You can do 130 days a year
00:04:44
or something. Is that still the status?
00:04:46
Yeah, it's roughly half after work days.
00:04:48
Got it. And so what do you do? You have
00:04:49
a you have a punch clock there. When you
00:04:51
get to the White House, you punch in,
00:04:52
you punch out like Fred Flintstone or
00:04:54
what? Are you keeping track of these
00:04:55
days? How do you do it? I know why you
00:04:57
don't know this because you have yet to
00:04:58
be invited to the White House. But
00:04:59
that's interesting. I got not how it
00:05:02
works. Normal people just people just
00:05:04
badge in and badge out like that. Badge
00:05:06
it and badge. It's a natural place,
00:05:07
Jason. I I mean, literally, it's
00:05:09
interesting. There's a new private club.
00:05:11
It's incredible that you have thoroughly
00:05:14
prepared for this week, just like
00:05:15
always. I am always prepared.
00:05:17
Interestingly, I don't know if you
00:05:18
gentlemen know this, Ryan and Aaron.
00:05:20
There's a new private club in DC that uh
00:05:23
Don Jr. is doing and Sax is a member.
00:05:25
Chimat's a member. And I just checked my
00:05:27
Gmail. I checked all three of my Gmail
00:05:29
accounts, everything. No invite. You
00:05:31
must have gotten lost again.
00:05:34
Did you send a paper one? Was it like
00:05:35
you sent a goal card or something? Sax,
00:05:37
how do I get invited to this private
00:05:39
club? What is this private club?
00:05:40
Everybody wants to know. Well, we'll be
00:05:42
happy to have you as a guest. Okay. Do I
00:05:45
have to wear a MAGA hat and have the
00:05:46
courtesy MAGA hats at the door? If you
00:05:48
want to be a member, obviously there are
00:05:50
dues and a membership fee, and Okay. I
00:05:53
just didn't want to waste your time with
00:05:55
an offer that I knew you wouldn't uh be
00:05:58
willing to accept. It's only $500,000 is
00:06:01
what I read. Is that true?
00:06:03
That's true for for founding members who
00:06:05
have additional benefits, but there's
00:06:06
also a lower level that's the more
00:06:08
reasonable membership level. So, I think
00:06:11
people are getting a little bit carried
00:06:12
away with that number. Got it. Okay.
00:06:13
That's why I wanted to clarify. Yeah.
00:06:15
Yeah. There's like 10 founding members
00:06:17
who have that level and then there's a
00:06:19
lower level for more average member.
00:06:22
Chimatha, are you one of those 10? Yes.
00:06:25
Do you pay more if you have TDS or how
00:06:27
does that work? TDS premium. What do you
00:06:29
talking about? Jal specifically or what
00:06:30
are we talking about? the TDS search
00:06:33
charge. Asking for a friend. It's a TDS
00:06:35
search charge. You put the tariff search
00:06:38
we just we want a place to hang out in
00:06:41
DC. All of us have been to clubs like
00:06:43
the Battery or I don't know if you go to
00:06:45
LA like the I think places. There's
00:06:48
Malibu Beach House. There's Bird Street
00:06:50
Clubs. There are places in Palm Beach
00:06:52
that are really cool. In any event, we
00:06:54
wanted a place to hang out. And the the
00:06:56
clubs that exist in Washington today
00:07:00
have been around for decades. They're
00:07:01
kind of old and stuffy. To the extent
00:07:03
there are Republican clubs, they tend to
00:07:05
be like more Bush era Republicans as
00:07:08
opposed to Trump era Republicans. So, we
00:07:11
wanted to create something new, hipper,
00:07:13
and Trump aligned. Since I'm in the
00:07:16
government, I can't be an owner, but I
00:07:18
told him I'd be happy to be member
00:07:20
number one. And so, I, you know, said,
00:07:22
"Great, let's let's do it." And so,
00:07:24
we're creating a place for us to hang
00:07:25
out. That's basically it. We want a
00:07:27
place to go where you don't have to
00:07:28
worry that the next person over at the
00:07:30
bar is a fake news reporter or even a
00:07:33
lobbyist or something like that who we
00:07:35
don't know and we don't trust. Got it.
00:07:38
So, it's like any private club. You want
00:07:40
to go somewhere that's highly curated.
00:07:45
This private club uh movement is
00:07:46
happening all over the country, not just
00:07:48
Washington. But we're creating something
00:07:49
that didn't exist before in DC, which
00:07:52
again is younger, hip, Trump aligned,
00:07:54
Republican. We're uh I I actually
00:07:57
started a Kamala club in um in the Bay
00:07:59
Area. So um so we're Yep. I don't think
00:08:03
anyone would pay to join that though is
00:08:05
the problem, right? I mean, it's an open
00:08:08
bar, that's for sure.
00:08:10
Where where do where do you guys meet
00:08:12
up? In like Redwood City. We actually
00:08:13
meet up at the uh at the at the trade
00:08:15
ports. All right. was we're 100 days
00:08:18
into Trump 2.0. It's just a random 100
00:08:22
day thing, but everybody's talking about
00:08:24
everybody's hand ringing. What has it
00:08:26
been like for this first 100 days? How
00:08:28
does it compare to Biden? How does it
00:08:29
compare to Trump 1.0? 143 executive
00:08:32
orders, the most ever in the first 100
00:08:35
days. And they're moving obviously at a
00:08:37
at a different pace to uh be generous.
00:08:41
Major indices are down 7 to 10%.
00:08:43
Obviously, this trade war and tariffs,
00:08:47
the yield on the the yield on the
00:08:49
10-year, it's down about 40 basis
00:08:51
points. There's a lot going on. Let's go
00:08:53
around the horn. Ryan Aaron, you're our
00:08:55
guest. What's your take on the first 100
00:08:57
days? Is it what you expected, good,
00:09:00
bad, and otherwise, wins and fails,
00:09:02
everything. I'll go first. I think it's
00:09:04
a whirlwind. I mean, if you look at the
00:09:07
uh the John Boyd, the fighter pilot, has
00:09:09
this concept of the UDA loop, which is
00:09:11
observe, orient, decide, and act. And
00:09:14
the concept is that if you're in dog
00:09:15
fighting, if you're able to maneuver
00:09:18
through those UDA loops at a faster pace
00:09:20
than your than your competition, they
00:09:22
get disoriented and don't know what to
00:09:23
do. And I I think that that's got to be
00:09:26
how Democrats in Washington and maybe
00:09:28
mainstream Republicans in Washington.
00:09:30
Certainly journalists are all feeling
00:09:32
this like there's that the Trump Trump
00:09:34
administration takes action and before
00:09:37
anybody can respond to that they have
00:09:39
already done like four more things and
00:09:40
you're like wait I forgot to actually
00:09:42
follow up on the other thing that they
00:09:43
did that I didn't like. Uh, and so it's
00:09:45
yeah, it's pretty disorienting if you're
00:09:47
if you're trying to they they can't find
00:09:49
a line to fall back to and go, "Hey,
00:09:50
we're going to push back against this
00:09:51
policy because they're already moving on
00:09:53
to the next one, the next one." Um, so
00:09:55
that that's like my high level
00:09:56
interpretation. Obviously, I come at it
00:09:58
from a trade angle. I think everybody
00:10:00
knew that Trump was going to be he he
00:10:02
told us during the campaign that the the
00:10:04
most beautiful word in the English
00:10:05
language is
00:10:06
tariff. Don't tell them it's an Arabic
00:10:08
word, but the most beautiful word in the
00:10:11
English language. And so we knew that
00:10:12
was coming. I think that the the the
00:10:16
suddenness of it all caught people by
00:10:18
surprise. I mean, they told us April
00:10:20
1st, April 2nd would be Liberation Day.
00:10:23
They didn't tell us that it would go
00:10:25
live the next week, you know, and effect
00:10:27
you've already ordered these goods. So,
00:10:28
that's one aspect that people are kind
00:10:31
of disoriented about. And we're gonna
00:10:33
unpack. Yeah, we're going to unpack
00:10:35
that. Aaron, your thoughts on the first
00:10:36
100 days? Obviously you are a Democrat
00:10:40
and uh you were pretty vocally not in
00:10:44
support of Trump. So what's your take on
00:10:46
the first 100 days? Any any bright spots
00:10:48
for you things you you know support?
00:10:50
Actually Sax's world I'd say has has
00:10:53
been a bright spot. So especially I mean
00:10:55
I think we have a very clear message on
00:10:57
AI and uh and that that that is that's
00:11:00
been I think a huge net positive is um
00:11:03
you know if you look at the the past you
00:11:05
know few months uh out of all the the AI
00:11:08
push from the administration it's
00:11:09
unmistakably you know pro open source
00:11:11
you know pro you know bring as much AI
00:11:14
innovation you know to the US obviously
00:11:16
that the tariffs you know add a little
00:11:17
bit of a headwind to that. I have some
00:11:19
very strong asks, you know, around high
00:11:21
skilled immigration because I think
00:11:22
that, you know, AI talent is going to be
00:11:25
super critical to to actually win the AI
00:11:27
war. So, so I'd say that that
00:11:28
directionally has has had some positive
00:11:30
momentum. You know, from my perspective,
00:11:32
this is kind of playing out almost
00:11:35
exactly how I thought it would 6 months
00:11:36
ago. And then 3 months ago, I I think
00:11:40
there was some signs that maybe maybe,
00:11:42
you know, it wouldn't play out this way.
00:11:44
um just based on some of the some of the
00:11:46
you know kind of early groups that were
00:11:48
coming to the White House the the the
00:11:49
the sort of deep business you know kind
00:11:52
of centricity of the White House you
00:11:54
know I think it was day one or two that
00:11:56
Stargate was announced you know at the
00:11:58
White House we're going to go build
00:11:59
massive infrastructure the case I'd like
00:12:01
to make you know once we talk about
00:12:02
tariffs is is I think there's an
00:12:03
alternative universe where you just lean
00:12:05
into acceleration as opposed to adding
00:12:07
headwinds but but so that would be that
00:12:10
would be the case of what what maybe
00:12:11
could have been you know very different
00:12:12
is we just keep double down on doubling
00:12:14
down on what's working while fixing the
00:12:16
parts that aren't working. But uh but
00:12:19
that would be, you know, my my uh my
00:12:20
judgment so far. Chimath, I mean, you've
00:12:23
been talking about it here every week.
00:12:26
You and I have uh been talking about it
00:12:28
pretty consistently, so I don't think
00:12:29
there'll be many surprises here, but
00:12:31
take a second and maybe assess what you
00:12:33
think if you had to pick a singular
00:12:35
thing that's gone really well and a
00:12:37
singular thing you think could be
00:12:38
improved. What What do you got? Let me
00:12:40
give you my overall grade.
00:12:44
And then I'll tell you how I get to
00:12:45
that. I think the first 100 days have
00:12:49
been a B+.
00:12:54
And here's how I get to that
00:12:56
score. There have been two things where
00:12:58
I think Trump
00:13:01
has frankly hit a home run. The first is
00:13:06
all of the direct
00:13:08
investment and specifically the foreign
00:13:10
direct investment into the United
00:13:13
States. I think it's approaching if not
00:13:15
it has already exceeded a trillion
00:13:17
dollars from corporations and
00:13:20
organizations and individuals from
00:13:22
around the world who have committed to
00:13:25
bringing money into the United States.
00:13:27
And I think strategically that's a
00:13:29
legacy that will live past him. So, I
00:13:32
think that's been an
00:13:33
A+. The second is we had a very unsafe
00:13:38
border
00:13:39
situation and he ran on shutting it
00:13:43
down. I'm not talking about the
00:13:45
execution of the deportations. I'm just
00:13:47
saying getting the illegal crossings to
00:13:50
zero and he's done that. So, that's been
00:13:53
an A+.
00:13:55
I think what's going to be more
00:13:57
controversial are these next three
00:13:58
things though. But in my interpretation,
00:14:01
I think the tariffs have been an
00:14:04
A and I think that the market reaction,
00:14:07
the stock market is only down 4%. And
00:14:10
the interest rate markets are, you know,
00:14:12
4 and a/4%. I think those have been an
00:14:14
A. Now the reason I think tariffs have
00:14:18
been an A is because it is uncovered in
00:14:21
my opinion how beholden we are to a
00:14:25
brittle supply chain and specifically to
00:14:28
China who is a friend but who's also an
00:14:30
enemy and I think that that's going to
00:14:32
really severely complicate
00:14:35
our flexibility and optionality in the
00:14:38
future as they do what is in their best
00:14:40
interests. Okay. So where have they then
00:14:43
not done so well? I think the documents
00:14:47
have been frankly a D. We were supposed
00:14:50
to get the Epstein files. We haven't
00:14:52
yet. We were supposed to get the Martin
00:14:55
Luther King files. We haven't. We did
00:14:57
get the redacted JFK files. I don't
00:14:59
think there's been very good
00:15:00
communication about why it's taking so
00:15:02
long. So I think it's a very small
00:15:04
narrow thing, but I think it had a lot
00:15:06
of attention on the way in. I think
00:15:09
the communications of the
00:15:12
tariffs and the back and forth have been
00:15:14
a C. I think the markets were not
00:15:19
led in enough of a way where they could
00:15:22
absorb the
00:15:25
volatility. But if you take it all in
00:15:27
its totality, I would give it a B+. I
00:15:30
think it's been a very productive 100
00:15:32
days. And when you look back, I think
00:15:34
in, you know, 3 years, four years, 5
00:15:36
years. Okay, we've made some important
00:15:39
progress. Saxs, obviously you're part of
00:15:42
the administration, so I'm not sure
00:15:44
exactly how to ask you this, but you s
00:15:45
you heard some nice compliments about AI
00:15:47
from Aaron. I I happen to agree with
00:15:49
those. I actually agree with uh a good
00:15:51
portion of the crypto stuff, too. I
00:15:52
think actually getting those uh
00:15:54
tightened up, which are your two zones
00:15:56
of excellence and your area that you're
00:15:59
focused on. I think you've done a great
00:16:00
job there. So, just bestie to bestie,
00:16:02
great job there. What's your Thank you.
00:16:04
What's your take overall? You know, it's
00:16:06
kind of hard, I guess, to ask somebody
00:16:08
in the administration to criticize the
00:16:09
administration, but hearing everybody
00:16:10
else's take, what's your response,
00:16:12
maybe? Well, I would I would highlight
00:16:14
three main areas that I think are big
00:16:17
accomplishments for the Trump
00:16:18
administration in the first 100 days.
00:16:19
So, so number one has to be the border.
00:16:21
Like Jamas said, I think you have to
00:16:23
give the administration an A+ on this.
00:16:25
They've completely stopped the border
00:16:26
crisis. I think we all knew that Trump
00:16:29
would take action on this because it's
00:16:30
one of the main issues he campaigned on.
00:16:32
I think if you had asked any of us, you
00:16:34
know, 4 months ago, would this problem
00:16:36
be completely solved? Meaning, border
00:16:39
apprehensions completely stopped, border
00:16:41
completely sealed within the first 100
00:16:43
days, I don't think we would have
00:16:45
believed necessarily that it would get
00:16:46
done so quickly, but it has. Uh, recall
00:16:50
that for 4 years during the Biden years,
00:16:52
we were told for the first 3 years that
00:16:55
the problem didn't even exist. Whenever
00:16:57
the videos were published of caravans
00:17:00
coming or throngs of people running
00:17:02
across the border, we were told that
00:17:04
these were cherrypicked videos on Fox
00:17:06
News. It wasn't real. Finally, in the
00:17:09
last year of the Biden administration,
00:17:11
they said, "Okay, we're finally going to
00:17:12
do something about it." They took some
00:17:13
limited actions and they said that doing
00:17:15
more than that would require new
00:17:18
legislation. Well, all of that was just
00:17:20
gaslighting. It turns out Trump came in,
00:17:22
he restored remain in Mexico and other
00:17:24
policies, completely stopped it. He had
00:17:26
this line at the state of the union
00:17:27
which I think is exactly right which is
00:17:28
we didn't need a new law we just needed
00:17:30
a new president. So I think that's area
00:17:32
number one. Area number two I would say
00:17:35
would be the vibe shift in the culture
00:17:38
around wokeism and DEI. You know how
00:17:41
quickly we forget about this but wokeism
00:17:44
has completely collapsed. Uh I don't
00:17:46
know that anyone is endorsing in a
00:17:49
fullthroated way. Moreover, beyond just
00:17:51
sort of the cultural aspect of it, I
00:17:54
think we've had significant policy
00:17:55
changes on DEI. Trump has basically
00:17:58
ended DEI at the government level. He
00:18:01
also signed an executive order ending
00:18:04
the use of disparit impact for
00:18:06
affirmative action. This is the policy
00:18:08
that said that even if you have a policy
00:18:11
that's applied in a completely neutral
00:18:13
and objective way, if it results in a
00:18:16
disparate impact where different groups
00:18:17
are represented in a different way in
00:18:19
the outcomes, then somehow that must be
00:18:22
racist. And that led to essentially
00:18:25
engineering the results of various
00:18:28
populations to basically fit quotas. And
00:18:32
I think all of that now has fallen by
00:18:34
the wayside. And I think that
00:18:35
meritocracy and colorblindness are back.
00:18:37
The only hold out really has been these
00:18:40
universities where Trump is now taking
00:18:42
action against Harvard and I think that
00:18:45
ultimately we will win that battle. You
00:18:47
see that even in relatively liberal
00:18:50
companies the DI departments have been
00:18:52
cancelled and they're moving back
00:18:53
towards more of a meritocracy. So I
00:18:56
would say that that's like big shift
00:18:57
number two. And I think if any of us had
00:18:59
tried to predict that 100 days ago, we
00:19:02
would have thought yes, Trump will do
00:19:03
something about it. But I don't think we
00:19:04
would have predicted the total collapse
00:19:07
of wokeism and DEI so quickly. And then
00:19:10
I'd say the third area which is still in
00:19:13
flight is the rep privatization of the
00:19:15
economy. That's a term that Scott Besson
00:19:17
used. I think that the Trump
00:19:19
administration needs to reprivatize the
00:19:21
economy. And I like that framing of it.
00:19:24
And there's a bunch of different pieces
00:19:26
under that. I'd say number one is
00:19:28
Doge again ending this hogw spending. I
00:19:33
do think that Trump has come into office
00:19:36
inheriting a very weak Biden economy
00:19:38
that was being propped up by massive
00:19:40
amounts of government spending that was
00:19:42
not only stimulating the public sector,
00:19:44
but it was also gooseing the employment
00:19:46
numbers as well. And we knew that that
00:19:49
spending was unsustainable. We have to
00:19:51
do something about it. So, I think for
00:19:53
the first time in decades, we've
00:19:54
actually started to make real cuts in
00:19:57
government, real cuts in the federal
00:19:59
workforce. And look, we'd like to do
00:20:01
more, but that is a huge shift in the
00:20:03
conversation. There's other pieces of it
00:20:06
as well. I mean, President Trump has
00:20:08
signed a significant number of executive
00:20:10
orders on deregulation. There's also
00:20:13
been unleashing energy. He ended Biden's
00:20:15
EV mandate and a lot of these like green
00:20:18
new scam projects, offshore wind, and
00:20:21
he's been encouraging oil and gas
00:20:23
exploration. So, I think there's that.
00:20:25
And then I appreciate what Aaron said
00:20:27
about tech innovation. We did repeal
00:20:30
Biden's exec order on AI, which was, you
00:20:32
know, 100 pages of unnecessary
00:20:34
regulation on AI. We've ended the war on
00:20:36
crypto, and I think we're trying to stop
00:20:39
the regulatory capture that benefits
00:20:40
large incumbents. So, you have all these
00:20:43
things, and there's been other things
00:20:44
that that have been done on the economy
00:20:45
as well, but I I do think that this sets
00:20:47
us up for a Trump boom in the future.
00:20:51
It's just that a lot of these changes
00:20:53
take time to to play out. Okay, great.
00:20:56
Well done. And well, I think I think I
00:20:59
think we knew Sax would be very pro.
00:21:01
Chamas uh Chamas seems really pro other
00:21:03
than he wants like the alien conspiracy
00:21:05
files released which we'll get soon.
00:21:08
What is the what is the view from Jay
00:21:10
Cal when the where you're the the the
00:21:13
leftleaning guy in the in the room? Uh
00:21:16
you know I'm kind of independent but
00:21:17
yeah social liberal. You know I I look
00:21:19
at what all Americans believe and and
00:21:22
try to build some consensus here. It's
00:21:24
one of the things I've been trying to do
00:21:25
on the pod is look for where we
00:21:27
agree. Americans universally want the
00:21:30
border secured. They don't want illegal
00:21:33
immigration and they don't want
00:21:34
fentinel. So this is the biggest win I
00:21:36
think for Trump which I think everybody
00:21:37
on the panel pointed out and Sachs you
00:21:40
were dead right like when we were seeing
00:21:41
those videos some of them were 5 years
00:21:42
old some of them were recent. Biden
00:21:45
really covered up what was going on in
00:21:46
the border and it took years to figure
00:21:48
out what was exactly going on there. So
00:21:50
that's the biggest win possible. I I
00:21:52
give overall just to be brief a B
00:21:55
for this first 100 days and I give you
00:21:58
know Biden like a C minus. The second
00:22:01
thing that everybody agrees on is they
00:22:02
want to downsize the government. They
00:22:03
don't want waste and fraud. So I think
00:22:04
Doge is the other huge win. The things I
00:22:08
think that could be improved really just
00:22:10
three simple things. The economic
00:22:12
uncertainty is really terrible for
00:22:14
running a business. I'm seeing a lot of
00:22:15
folks in my circle on my podcasts this
00:22:18
week in Startups and here telling me,
00:22:20
"Oh, I don't know how to plan for the
00:22:22
future." And we're going to get into
00:22:23
that with this tariff stuff and the
00:22:24
trade war. And so, I think economic
00:22:27
uncertainty, we have to sort of slow
00:22:30
down and maybe make it easier for people
00:22:32
to understand what the administration is
00:22:34
trying to do. I think rule of law really
00:22:36
matters to people. People didn't like
00:22:38
Biden's pardons. They didn't like
00:22:40
covering up his mental acuity. And I
00:22:42
don't think people like the deportations
00:22:44
without due process. We talked about
00:22:45
that on a previous episode.
00:22:47
Overwhelmingly, people want Trump and
00:22:50
the administration to obey what the
00:22:52
Supreme Court says. They really want
00:22:54
rule of law, the third term talk. Like
00:22:57
eight out of 10 Americans don't like
00:22:58
that kind of talk. Um, and then
00:23:01
conflicts of interest. Obviously, people
00:23:02
hated the Hunter Biden stuff. They hate
00:23:04
the memecoin stuff. And so that's where
00:23:07
it could improve. crisper
00:23:08
communications, more thoughtful
00:23:10
execution, maybe less trolling. I don't
00:23:12
like the White House Twitter account
00:23:14
trolling. And then focus on what got
00:23:16
Trump here. You know, you all said the
00:23:17
same thing. What got Trump in here was
00:23:19
the economy. And one thing that wasn't
00:23:22
mentioned by everybody is the peace
00:23:24
dividend. And Trump is making massive
00:23:26
progress in Ukraine, apparently. I don't
00:23:29
know if it's on the docket today or not,
00:23:30
but stopping the wars and making the
00:23:33
economy boom, those are the two most
00:23:34
important things that he could do. Build
00:23:36
on that. I I totally missed that. You're
00:23:38
absolutely right. That's another one
00:23:39
where I would give Trump an A+. Nat and
00:23:41
I had dinner with POTUS two weeks ago.
00:23:45
And wait, you had dinner with Trump?
00:23:47
This is breaking news. Well, okay,
00:23:49
whatever. Yes. Well, I think it's I
00:23:50
think it's remarkable how much of a
00:23:52
Putin apologist Jake House's become. I
00:23:54
mean, you want to end the war in Ukraine
00:23:55
now? Well, you're going to you're going
00:23:56
to give it You're going to give it to
00:23:58
Putin? You're not going to stop Putin.
00:24:01
I'm totally in favor of what Trump's
00:24:03
doing in in negotiating a deal to get
00:24:05
more money. Oh, you want to talk to
00:24:06
Putin now? I've always wanted to talk to
00:24:08
Putin. I just don't trust him. But you
00:24:10
you can trust him. Let me tell you what
00:24:12
Trump said. So, there we go. There was a
00:24:14
handful of us at dinner and then he got
00:24:18
up to say a few words at the end and he
00:24:21
reminded me why I was so inclined to
00:24:27
vote for him, which is he talked about
00:24:29
his uncle
00:24:31
and he talked about how his uncle taught
00:24:33
him about the severity of nuclear war
00:24:36
and how people don't understand how
00:24:39
intense and how destructive it is and
00:24:42
the power of these weapons and he left
00:24:47
that speech at the end saying and this
00:24:49
is why I'm so fundamentally against this
00:24:51
thing and it reminded me to your point
00:24:54
Jason it is so easy to forget that
00:24:57
there's only one existential risk
00:25:00
save like aliens coming from the
00:25:02
heavens, right? There's only one
00:25:04
existential risk where all these issues
00:25:06
become fringe issues. You know, you
00:25:07
mentioned rule of law, border security,
00:25:09
foreign direct investment, tariffs, it
00:25:12
all goes out the window in a nuclear
00:25:14
war. And I was like, I am so glad this
00:25:17
guy's in charge because this one issue,
00:25:20
he never waivers.
00:25:23
Yeah. And I think there's all kinds of
00:25:25
complicated moments that could make this
00:25:29
an issue. And this was where my biggest
00:25:31
issue with Biden was was I did not know
00:25:33
who was in control. And I think that
00:25:35
Trump in the first 100 days, to your
00:25:37
point, I think has completely reinforced
00:25:39
that there are no conditions under which
00:25:42
he'll go to war. He has time and time
00:25:43
again showed find the off-ramp. And I
00:25:46
think that that's really healthy for
00:25:47
Americans to see. Yeah. And let me build
00:25:49
on that point with respect to to Ukraine
00:25:51
is we were on a glide path before the
00:25:54
Trump presidency that Biden had put us
00:25:57
on a certain path. Kla Harris gave every
00:26:00
indication she would have continued it.
00:26:02
What was that path? It was a path of
00:26:04
continued escalation and doubling down
00:26:06
in Ukraine. Recall that it was Biden
00:26:08
himself at the beginning of the war who
00:26:10
said that if we give Ukraine Abrams
00:26:13
tanks and F-16s or attackums or high
00:26:18
Mars or if we allow them to hit targets
00:26:20
inside of Russia, it would lead to World
00:26:23
War II. He actually used the word
00:26:25
Armageddon. So at the beginning of that
00:26:27
administration, they were very concerned
00:26:30
about how an escalatory path could lead
00:26:32
us into direct conflict with Russia and
00:26:35
World War II. And yet, despite that, at
00:26:38
every fork in the road where they had a
00:26:39
choice, they ended up doubling down.
00:26:41
They gave the Abrams tanks. They gave
00:26:42
the F-16s. They gave the High Mars. They
00:26:44
gave the attacks. And finally, when
00:26:47
Biden was a lame duck in his last couple
00:26:48
months in office, they did the most
00:26:50
reckless and irresponsible thing, which
00:26:52
is allow American weapons to be used to
00:26:55
strike targets on Russian soil. Not just
00:26:57
fighting in Ukraine, but on Russian
00:26:59
soil. Moreover, we now know from a New
00:27:01
York Times article that just came out in
00:27:03
the last few weeks that it was American
00:27:05
generals and American intelligence who
00:27:06
are planning this war. So, when you're
00:27:08
talking about striking Russian targets
00:27:10
on Russian soil, it's not just the
00:27:12
Ukrainians using our weapons. They're
00:27:14
using our targeting, they're using our
00:27:16
guidance, they're using our satellites.
00:27:18
I mean, we are deeply integrated in the
00:27:19
kill chain. This is the United States
00:27:21
being a co-elligerent in the war,
00:27:23
hitting Russian soil. That is incredibly
00:27:25
reckless and dangerous. I have no doubt
00:27:27
that if the Democrats were still in
00:27:30
office, we would be in an escalatory
00:27:32
spiral right now with the destination
00:27:34
being World War II. And I do think that
00:27:36
Trump has pulled us back from the brink
00:27:38
there. There's obviously still more work
00:27:40
to do. But I really appreciate the
00:27:42
efforts that Steve Wickoff has
00:27:44
undertaken where for the first time in 3
00:27:46
years, we've at least had direct
00:27:48
diplomacy with the Russians. We weren't
00:27:50
even talking before. We weren't even
00:27:52
talking before. I mean, talking is a
00:27:53
great thing and and and apparently we're
00:27:55
going to keep supplying with them with
00:27:57
weapons as long as they pay for them.
00:27:59
So, it's going to be very interesting to
00:28:00
see how this all hashes out over the
00:28:02
next 100 days or so. Let's keep moving.
00:28:05
I don't think we know that yet. Let's
00:28:07
let's wait and see on that. Okay. Yeah.
00:28:08
I mean, I think that's Yeah. Uh what was
00:28:11
reported, but you're right. We should
00:28:12
wait and see. Okay. Downstream tariff
00:28:15
impacts. We got to talk about this, and
00:28:16
this is why we have you here, Ryan,
00:28:18
since you're in the thick of it. you
00:28:20
tweeted a thread last week about the lag
00:28:23
time uh of shipments from China and when
00:28:26
you were on I guess during COVID you
00:28:28
really educated us to how the supply
00:28:30
chain how the supply chain
00:28:32
works and according to the thread that
00:28:36
you shared somewhere around early June
00:28:38
we're going to expect warehouses
00:28:40
trucking the entire supply chain maybe
00:28:43
to start to seize up or layoffs I don't
00:28:47
know how you would frame it Ryan but are
00:28:49
we asked the point of no return with
00:28:52
regard to the supply chain. Is there an
00:28:55
off-ramp for this tariff conflict war
00:28:58
negotiation with China in your mind?
00:29:00
What are you seeing on the streets and
00:29:03
in the purchase orders and the invoices
00:29:06
at Flexport? Definitely not past the
00:29:08
point of no return. I think we're still
00:29:10
right in the middle of the don't judge
00:29:12
the cook while he's cooking is one, you
00:29:13
know, like let's see what the we'll see
00:29:14
what it tastes like at the end is I
00:29:16
think a starting point here and we're
00:29:17
still they're still in active
00:29:18
negotiations. So I don't think today's
00:29:20
it's not static. Now the world does want
00:29:23
a lot more certainty and that's a big
00:29:25
cause of what's happened here and what
00:29:27
has happened is a 60% decline in
00:29:29
bookings of ocean freight from China to
00:29:31
the US. I mean so that's really really
00:29:33
pretty dramatic like probably exceeding
00:29:36
what was expected. I don't think, you
00:29:39
know, when they issued when they rolled
00:29:40
out the initial reciprocal tariff plans
00:29:42
on on April 2nd, it was meant to be a
00:29:44
54% tariff on China. Then, you know,
00:29:47
there's multiple cycles of escalation.
00:29:49
We ended up at what's now 154% tariff.
00:29:53
So, this is this is a lot higher than
00:29:56
anybody planned for. And so therefore, I
00:29:58
don't think anyone's planning for a 60%
00:30:00
decline in ocean freight. Ryan, let me
00:30:03
ask you a question about that. Are
00:30:05
people actually paying that 154%?
00:30:07
There's been this discussion online and
00:30:09
it's it's sort of unclear from the
00:30:11
administration and from
00:30:14
retailers stuff that's landing that
00:30:16
people ordered before April 2nd. Are
00:30:20
they actually paying the 154% on top of
00:30:22
what's landing? It's it's live now. Um
00:30:24
it is it was based on departure date. So
00:30:26
goods that departed China after midnight
00:30:30
Eastern time on April 9th are subject to
00:30:33
the tariffs upon arrival. And so now
00:30:35
enough time has passed that pretty much
00:30:36
all the ships that are arriving now left
00:30:38
China after April 9th when that started.
00:30:40
Um so yes. So what happens? People are
00:30:42
paying it or are people saying I won't
00:30:43
take delivery because it's Jason you
00:30:45
have you have to pay it Ryan correct me
00:30:47
if I'm wrong but you have to pay it at
00:30:49
the dock in order to get the goods
00:30:50
released. More or less more or less
00:30:52
that's true. They they allow you have a
00:30:55
bond in place so you can pull the goods
00:30:56
out before you pay but it the money's
00:30:58
owed at that time and then you get you
00:31:00
get like a two week time frame to
00:31:01
actually make the payment. But there are
00:31:03
strategies here. a lot of people are
00:31:04
doing that you can use what's called a
00:31:06
bonded warehouse and move cargo into
00:31:08
this warehouse uh and then you only owe
00:31:09
the duties when the cargo leaves that's
00:31:12
what I was asking like is there a hack
00:31:14
here to it's not that that lets you
00:31:16
defer things and it's very very common
00:31:17
right now people are searching
00:31:19
everywhere for bonded warehouse capacity
00:31:20
because in a bonded warehouse not only
00:31:22
you defer payment to when the cargo
00:31:24
leaves the warehouse but you only owe
00:31:26
the duty amount based on at that date so
00:31:30
if the duties come back down which a lot
00:31:32
of people are betting they will on the
00:31:33
China specific speific duties, you'll
00:31:35
actually lower your tariff burden. And
00:31:37
then there's another hack for this,
00:31:38
which is effect use a Mexican or
00:31:40
Canadian bonded warehouse. So you move
00:31:42
the goods into Mexico and then you
00:31:44
actually only technically import them
00:31:46
into the US at a future date when
00:31:48
tariffs are lower. So I understand a lot
00:31:49
of a lot of companies are doing that
00:31:51
right now, too. Um we're helping some
00:31:53
people with that type of strategy,
00:31:55
but yeah. Is that Sorry, Ryan. Do you
00:31:57
think that the government will they view
00:31:59
that okay that kind of hack and or like
00:32:03
you know like if you look at the GDP
00:32:04
numbers one of the craziest things was
00:32:06
the inventory pull forward that people
00:32:08
did to your point like trying to get as
00:32:10
much stuff into the United States before
00:32:13
April 9th as an example. Yeah. I mean
00:32:16
it's not a hack. It's a bonded
00:32:17
warehouses are been around for decades
00:32:19
and they're they're very commonly used.
00:32:21
I don't know that it'll be that material
00:32:22
in the scheme of things that it would,
00:32:25
you know, cause a change in the law
00:32:27
around bonded warehouse. So, you don't
00:32:28
think, for example, the Department of
00:32:31
Commerce will have an issue with the
00:32:33
strategy of sending inventory into
00:32:36
Mexico that essentially you're
00:32:37
essentially like, isn't it, it's a work?
00:32:41
Like, instead of paying the China
00:32:42
tariff, now you pay a Mexico tariff,
00:32:44
which should be less. Is that the idea?
00:32:46
Well, you can move it into a bonded
00:32:47
warehouse in Mexico even and not pay
00:32:49
Mexican tariffs either. and you just
00:32:51
wait until it imports. But I mean,
00:32:52
what's the Department of Commerce or the
00:32:55
customs to do? It's sort of you just
00:32:57
delayed importing the goods. You've
00:32:58
imported them in the future and you
00:33:00
know, it doesn't I I wouldn't even call
00:33:02
it a hack. It's just sort of like people
00:33:04
people are going to get creative here.
00:33:05
You know what I mean? Like that's the
00:33:06
job. Actually, the government should set
00:33:08
the rules and the rest of us got to
00:33:10
figure out, all right, how are we going
00:33:11
to compete and make money in this
00:33:12
environment that they've created? Ryan,
00:33:14
in that tweet you redid, which was a
00:33:16
pretty dramatic tweet painting a very
00:33:18
like I don't know, you know, like a
00:33:21
pretty dire situation. Where are we at
00:33:24
in terms of how uh dire this will get or
00:33:28
resolvable? paint us the the the best
00:33:31
case scenario and what you expect could
00:33:33
happen in that case or if this gets
00:33:35
extended are we going to see as you know
00:33:37
people are hand ringing empty store
00:33:40
shelves Christmas gets ruined and all
00:33:43
these layoffs start happening in the
00:33:45
supply chain take us through the two
00:33:46
scenarios that people are debating yeah
00:33:48
I mean the the the bleak scenario which
00:33:50
is I don't really think it's going to
00:33:52
happen I think that the administration
00:33:53
doesn't want this to be their legacy
00:33:54
that they like created a policy that
00:33:56
just like kind of tanked small business
00:33:58
and supply chain. So, I don't I don't
00:33:59
actually think this is going to happen,
00:34:00
but the bleak scenario is tariffs stay
00:34:03
at this level for 145% on China. The 10%
00:34:06
goes way back up to what it was
00:34:08
originally announced in reciprocal
00:34:09
tariffs. So, there's no like safe haven
00:34:10
for tariffs and trade just falls off a
00:34:13
cliff and a lot of companies go bankrupt
00:34:14
in our in our especially small companies
00:34:16
are the ones that are importing from
00:34:17
China. Reality is like tariffs have been
00:34:20
high on China for a long time. Labor
00:34:22
costs in China are not are not there for
00:34:24
cheap labor. You're there for quality
00:34:25
manufacturing at this point. like
00:34:27
there's much cheaper labor in Southeast
00:34:29
Asia, other parts of the world than
00:34:30
there is in China. So you're in China
00:34:32
because of the manufacturing
00:34:33
capabilities, the ecosystem, not just
00:34:35
for cheap labor. Uh and if you could
00:34:37
have moved, you would have already with
00:34:38
the 25% tariffs from the Trump's first
00:34:41
terms were pretty were high enough
00:34:43
incentive. And so that's the bleak
00:34:46
scenario is that small business starts
00:34:48
getting wiped out. The ones that are
00:34:49
buying from China and it's a lot of
00:34:51
brands like it's not just Amazon seller
00:34:53
selling stuff that you don't need. It's
00:34:56
like all the brands that you know are
00:34:58
like you know fashion brands, apparel
00:35:00
brands. I had cuts clothing on this
00:35:03
weekend startups last week and he said
00:35:05
there's going to be like if this doesn't
00:35:07
get revol resolved in like let's say two
00:35:09
to four weeks in his group chats people
00:35:11
are going to start layoffs and they they
00:35:13
can't
00:35:15
physically restart the supply chain in
00:35:18
Vietnam or wherever to make t-shirts. So
00:35:20
Aaron what's your thought on this as
00:35:22
well just bringing you in. Sure. Well,
00:35:24
well, first of all, I mean, Ryan has
00:35:26
supplied me with a high degree of doom
00:35:27
scrolling and uh it's just like a horror
00:35:30
show reading his tweets. First of all, I
00:35:32
I like I would feel better if the
00:35:34
messages out of the administration were
00:35:36
either more kind of consistent or that
00:35:39
there was a logical connection between
00:35:42
do we either want to raise the kind of
00:35:46
you know tariff revenue stream or do we
00:35:48
want free trade like like those things
00:35:50
are are working against each other
00:35:52
because like depending on who you talk
00:35:53
to they they say this is a a mechanism
00:35:55
to bring down income tax which obviously
00:35:57
then by definition means that they
00:35:59
expect the tariffs to sort of persist.
00:36:01
um which is totally different from let's
00:36:03
go negotiate deals that just allow for
00:36:05
the you know free trade to actually
00:36:07
increase and so are we worried about the
00:36:09
reciprocity or we worried about kind of
00:36:11
revenue stream so that's a whole whole
00:36:12
issue you also have this issue which is
00:36:15
the messaging from the government and
00:36:16
this is the meta point I'll make in a
00:36:18
second is about is about how we could
00:36:19
have actually accelerated into the
00:36:21
transformation of the economy but you
00:36:23
know you have folks like Lutnik etc you
00:36:25
know going on on TV talking about the
00:36:27
the end state of our economy which are
00:36:28
are actually probably you fine messages,
00:36:31
but but we haven't seen what that vision
00:36:34
looks like, you know? So, everybody is
00:36:35
kind of confused like does this mean
00:36:37
that we literally go into manufacturing
00:36:38
plants and we're like the ones literally
00:36:40
doing the screws on an iPhone or is it a
00:36:43
bunch of jobs which are next generation,
00:36:45
you know, jobs which is like we're
00:36:46
managing robots and and like shipping
00:36:48
and logistics grows as a result of this
00:36:50
and all of the surrounding kind of
00:36:52
supply chains, you know, start to grow.
00:36:53
So like like you know to like there's a
00:36:56
there's an underlying I mean you know mo
00:36:58
most people on this call have managed
00:37:00
teams like you do change management you
00:37:02
lead people to the end state that that
00:37:04
you want them to sort of see the
00:37:06
potential in and I think some something
00:37:08
that kind of gets missed is and the part
00:37:10
that kind of confuses me is like I don't
00:37:11
know if if to you know exactly to Ryan's
00:37:14
point like people are in China because
00:37:15
of the ecosystem of manufacturing yet
00:37:17
the messages you get out of the
00:37:18
administration are like oh you know
00:37:20
we're going to have fewer toys at
00:37:22
Christmas time. It's like, no, that's
00:37:23
not the big picture. Like, the big
00:37:25
picture is is like, like, you know, this
00:37:27
is supplying the parts that go into
00:37:29
building a manufacturing plant and
00:37:31
building a car that that allows us to
00:37:33
actually, you know, be even remotely
00:37:35
competitive in car manufacturing. So,
00:37:37
where should in your mind this all lead?
00:37:39
Because you have some thoughts on
00:37:41
American exceptionalism and maybe
00:37:43
skating to where the puck is going. So
00:37:45
for if you were to become an adviser
00:37:47
Yeah. on this as a technology expert and
00:37:50
and somebody who spent their whole
00:37:51
career in it, what would you advise them
00:37:53
to do? I'd get rid of Navaro immediately
00:37:55
and and you would basically say, you
00:37:58
know, Mia Kulpa, like oops. And like
00:38:00
obviously you need to like land that
00:38:02
with some really cool trade deals that
00:38:03
make everybody kind of feel happy. And
00:38:05
you basically say, you know what, like
00:38:07
let's go back to the first two days of
00:38:08
Trump, which is let's announce massive
00:38:11
deals. We're bringing manufacturing here
00:38:14
with Stargate. We're doing TSMC. We're
00:38:16
building Nvidia chips. We're going to do
00:38:18
a deal which is you get like 5% tax uh
00:38:21
rate if you build in America. And so you
00:38:23
just stimulate a a manufacturing boom in
00:38:26
in the country. You know, we
00:38:28
incentivize, you know, automation across
00:38:30
the manufacturing. We use that as a
00:38:32
competitive weapon to go and compete
00:38:34
with with the the sort of lowerc cost
00:38:36
labor that that that happens
00:38:37
internationally. You know, we we find
00:38:39
every incentive and tool we can. we
00:38:41
deregulate, you allow people to build
00:38:43
these plants and so you don't have to go
00:38:45
through the, you know, three-year EPA
00:38:46
process like like you you just
00:38:48
accelerate from from this position and
00:38:51
you see you see it all as upside and so
00:38:54
then business leaders, you know, if you
00:38:56
go talk to the Fortune 500 company that
00:38:58
actually has to build, you know,
00:38:59
anything right now, you give them a path
00:39:01
to say, listen, we're going to help you
00:39:03
transition away from your current supply
00:39:04
chain and we're going to make it even
00:39:06
more competitive and more compelling in
00:39:07
America to do that. You know, there's a
00:39:10
reason that Elon builds in America like
00:39:12
like he has he's actually made it be
00:39:14
more effective to to be able to to, you
00:39:16
know, bring automation to manufacturing,
00:39:18
to be able to to build locally, but he
00:39:20
wasn't forced to do that. And so so I
00:39:22
would just I would argue like you you
00:39:24
use as many carrots as possible in some
00:39:27
surgical areas. And Chimath, I've heard
00:39:28
your points about the like, you know,
00:39:30
chips, pharma, you know, AI, like in
00:39:32
those surgical areas, we get tough where
00:39:35
where necessary. Um, and and if we have
00:39:37
to do, you know, a couple sort of very
00:39:39
surgical tariffs, you know, to kind of
00:39:41
make make people move the the direction
00:39:43
that we want, that's totally fine. But I
00:39:45
I mean like like it's like the even
00:39:47
arguing the premise is hard because
00:39:48
because we act like we're like like it's
00:39:50
like countries that are screwing us, but
00:39:52
actually businesses are are
00:39:53
independently making decisions about
00:39:55
where they want their supply chain to
00:39:57
exist. They they have in a free market
00:39:59
they've made that decision. They don't
00:40:01
need the government to tell them where
00:40:03
where are they supposed to or where are
00:40:04
they allowed to to have their supply
00:40:06
chain operate. that that ends up with
00:40:08
just lots of economic distortions that
00:40:10
everybody on the the right would have
00:40:12
called, you know, the left socialists
00:40:14
for trying to kind of implement central
00:40:16
planning around supply chains. So that's
00:40:17
my piece. Well, no. What do you think,
00:40:19
Chimath, here of this sort of
00:40:21
reframing/offramp and sort of maybe the
00:40:23
positive spin on it, hey, if you want to
00:40:25
make t-shirts, you know, you want to
00:40:26
make commodity items, have at it. Free
00:40:29
trade, you know, reciprocal tariffs,
00:40:31
great checkbox there. But here is a
00:40:35
series of incentives and a path forward
00:40:37
to do the advanced stuff to do robotics
00:40:40
etc. Let me answer this in a different
00:40:42
way. Okay, a lot of those things he's
00:40:45
actually doing. I think this is where we
00:40:49
are is we're beyond
00:40:52
TDS. There's something that comes after
00:40:54
it. And I think that the mainstream
00:40:57
media has just lost their mind to a
00:41:01
degree that they hadn't even lost their
00:41:03
mind in Trump one. I'll give you a
00:41:05
couple of examples. Um well, one
00:41:08
example, by the way, just a shout out to
00:41:09
our friend, completely brazen,
00:41:12
ridiculous, shitty reporting by the Wall
00:41:14
Street Journal last night. When they
00:41:16
were told that this, you know, this
00:41:18
whole Tesla thing was a total farce,
00:41:20
they continued to publish it. Okay,
00:41:22
fine. They're they were referring to
00:41:24
Elon. The board Yeah. starting a search
00:41:28
to replace Elon and then the board said,
00:41:29
"Well, wait. We told you we weren't
00:41:30
doing that." And they didn't even
00:41:32
mention they communicated that directly
00:41:34
to the Wall Street Journal. The Wall
00:41:35
Street Journal said, "I don't care. I
00:41:37
have an axe to grind." Correct. Yeah. I
00:41:41
think that Trump has a strength, which
00:41:43
is he shapes these potholes for the
00:41:45
mainstream media to fall into. The
00:41:48
downside of that though is that that the
00:41:49
mainstream media then doesn't do the
00:41:51
other part of the job which is to tell
00:41:53
the things that are important. So for
00:41:55
example, we spent a lot of time
00:41:57
breathlessly talking about the MS-13
00:42:00
knuckles of the guy, right? Or then we
00:42:04
spent a bunch of time talking about how
00:42:06
MSNBC blurred out the names of the
00:42:09
placards on the lawn. Okay, but here's
00:42:12
the other part where then they get so
00:42:15
tilted. Here's what they don't report.
00:42:17
They didn't report that, for example,
00:42:19
when Trump took a shot at Harvard, he
00:42:23
also reinforced and strengthened
00:42:25
historically black colleges and
00:42:26
universities. Totally did not get
00:42:28
written. I'll give you another
00:42:31
example. This past
00:42:33
week, Bessant said that the tax bill
00:42:37
will allow you to fully deduct all the
00:42:41
PPE and all of the incidental costs of
00:42:45
building a factory. I heard that, I
00:42:48
immediately went to my wife. She runs a
00:42:50
pharma business. This is exactly what
00:42:52
she's trying to figure out. And we now
00:42:56
are like, how do you build a business
00:42:57
case if this actually gets effectuated?
00:43:00
The point is that
00:43:03
thing would create an absolute economic
00:43:07
bonanza if it were to get
00:43:10
passed. Other than people hearing it on
00:43:12
this pod or randomly maybe finding it on
00:43:15
a direct clip that Bessant puts out on
00:43:17
X, there has been zero coverage by the
00:43:20
mainstream media. Yeah. But but like
00:43:22
like the the uh first of all that Yeah.
00:43:25
Okay. MSM and whatever we want to call
00:43:27
it aside like the that is that is still
00:43:30
on the administration for driving a a
00:43:33
change management process that that
00:43:35
causes people to build on momentum and
00:43:37
not causes boards to basically say oh
00:43:40
are we going to pivot our entire supply
00:43:41
chain this week because because Trump
00:43:43
you know didn't get a call back from
00:43:45
from sheet like like that that is like
00:43:48
this is really not a a a TDS MSM issue
00:43:51
this if you talk to Fortune 500 CEOs did
00:43:54
you know about the PPE thing No, but
00:43:55
that's not but like I'm I don't need to
00:43:57
like the thing that I know but there are
00:43:59
many other CEOs that do they they're
00:44:01
controlling trillions of dollars of
00:44:02
capital allocation. It's an important
00:44:04
thing. If we had if we had Mary Bar on
00:44:06
this call and and we said Mary has has
00:44:10
you know Trump increased your ability to
00:44:12
execute and operate and accelerate the
00:44:13
the transition to the US or has he had
00:44:16
headwinds that make it tougher to
00:44:18
navigate right now? Which which way do
00:44:20
you think she'd go?
00:44:22
I think that she would give you a
00:44:23
calculated answer that neither is pro or
00:44:26
con. I think I think that answer changes
00:44:27
by the day. I think that if you talked
00:44:29
to her last week, she would say this has
00:44:30
been a major headwinds and then
00:44:32
yesterday to Chimal's point, they did
00:44:34
this thing where you can depreciate or
00:44:35
fully expense in year one capital
00:44:37
improvements or building out factories.
00:44:39
But also earlier this week, they made it
00:44:42
so that auto auto parts are not subject
00:44:44
to the tariffs. They created a huge
00:44:46
exemption that wasn't there everything
00:44:49
like they should have. By the way, This
00:44:51
all speaks if this was a bit planned, it
00:44:53
should have been there in the beginning
00:44:54
because these auto auto companies were
00:44:56
saying, "Hey, this is going to bankrupt
00:44:57
us if we have to pay taxes on where I
00:44:59
would take circling back to
00:45:01
communication and making a crisper and
00:45:03
clear, Erin, where you are right
00:45:06
expectation is it's my job to stay
00:45:08
informed." Okay. As a CEO of my company,
00:45:11
I try to stay informed and you're right.
00:45:13
It is hard because sometimes I find
00:45:15
myself hunting and pecking to find the
00:45:17
things that matter. But I do put a bunch
00:45:20
of that responsibility into the lap of
00:45:22
the people that are supposed to actually
00:45:24
report the facts. They can choose. They
00:45:27
didn't have to run that article about
00:45:29
Elon, which turned out to be total
00:45:31
and horseshit on the front page
00:45:33
of the Wall Street Journal. They could
00:45:34
have talked about what Ryan just
00:45:36
mentioned as the first article and said,
00:45:38
"Here completely changes your ro and roe
00:45:42
calculations for 90% of the S&P 500."
00:45:45
That was not the article they chose to
00:45:47
write and to publish. I but I also think
00:45:49
it comes back to my original point
00:45:50
around the UDOT loops that they're the
00:45:52
Trump is the administration is running
00:45:54
these very tight hey let's take an
00:45:56
action let's see what happens let's see
00:45:58
the reaction and then take another
00:46:00
action and Washington's used to doing
00:46:01
all these committees that plan
00:46:02
everything for 5 years or something or
00:46:04
whatever 18 months and then roll it out
00:46:06
slowly and they're going hey let's roll
00:46:08
it out oh crap we're about to cause this
00:46:10
huge problem in the auto manufacturers
00:46:12
and they're all telling us they're going
00:46:12
to go bankrupt okay 3 days later they
00:46:15
push an update and it's kind of it feels
00:46:17
chaotic But yeah, to summarize, Ryan and
00:46:20
Aaron, your position so we can keep
00:46:21
going through the docket. Hey, a little
00:46:23
less shock and awe, maybe a little more
00:46:25
predictability, a little crisper
00:46:27
communication, and Chimath, I think your
00:46:28
position is, hey, maybe the mainstream
00:46:30
media can play a better role here in
00:46:32
focusing us on what matters. That
00:46:34
wouldn't be my takeaway. So, yeah. Okay.
00:46:36
Um, my my What's your takeaway? I mean,
00:46:38
like like zero shaken like like not a
00:46:41
little less like like I like my strategy
00:46:44
would be 100% different. Actually, uh,
00:46:46
Scott Besson has an incredible podcast
00:46:48
from like September of last year, and he
00:46:50
basically said, you know, Bid Biden got
00:46:53
it all wrong. And I was like listening
00:46:54
to it. I was like, oh, okay, actually,
00:46:56
this is kind of cool. Like, he basically
00:46:58
says, deregulate the US, make it easier
00:47:01
to build manufacturing in the US,
00:47:03
increase the GDP, and then you'll be
00:47:05
able to take in less tax revenue, spend
00:47:06
less in the government. And it was like,
00:47:08
oh, this is actually like a glide path.
00:47:10
We could take we could take the fact
00:47:12
that we did a soft landing relative to
00:47:14
the rest of the globe. We're winning in
00:47:16
AI. We're winning in it in in you know,
00:47:19
you know, number of categories. We
00:47:20
obviously need more energy. We need we
00:47:22
need to bring in manufacturing into the
00:47:24
US. And and so you you have this great
00:47:26
momentum which is where we are the tech
00:47:28
leader, you know, in the world. Let's
00:47:29
like just pour fuel on that. And so to
00:47:32
pour fuel on that, you you do you just
00:47:34
do a series of carrots and and the and
00:47:36
the winds that build a flywheel of
00:47:38
positive energy. Like the reason why I
00:47:40
take I take a little bit of exception to
00:47:42
to Chimas's MSN point is that I think to
00:47:44
some extent Fortune 500 CEOs are not the
00:47:46
ones like oh my gosh like like Rachel
00:47:49
Matto said this like I'm going to go and
00:47:51
and you know worry about this topic now
00:47:52
like like the information coming at them
00:47:55
is is I'm not talking about the
00:47:56
information that's presented. I'm
00:47:58
talking about the information that's
00:47:59
excluded. How do you get the information
00:48:02
that's not published and shared broadly?
00:48:03
No, no, but but come on. Like like
00:48:05
Goldman Sachs and JP Morgan are not
00:48:06
writing reports on the fact that we
00:48:08
might enter a recession because of of
00:48:10
MSNB's MSNBC's reporting on this topic.
00:48:13
Like like again, but again, that's not
00:48:15
what I'm talking about. I'm saying like
00:48:17
gladhanding some highlevel
00:48:19
prognostication which nobody ever gets
00:48:21
right is in my opinion worthless. What
00:48:24
I'm talking about is the details. So
00:48:26
when you talk about something as narrow
00:48:28
and specific as excluding PPE or
00:48:32
allowing you to double or triple
00:48:33
depreciate something in a given calendar
00:48:36
year, fantastic. That is narrow. It's
00:48:38
precise. It's specific. It's actionable.
00:48:40
And what I'm saying is if I surveyed the
00:48:43
500 CEO of the S&P 500, dollars to
00:48:46
donuts, the overwhelming majority would
00:48:48
not have known. And had they brushed up
00:48:50
against that somehow in their normal
00:48:52
media consumption to then ask their
00:48:54
teams, the odds of that would have been
00:48:56
zero as well. So I guess then who do we
00:48:59
who do we blame for this chim? Is it the
00:49:01
administration's job or mainstream
00:49:02
media? But I bet you everybody knows
00:49:04
about the blurring out of the stupid,
00:49:06
you know, pictures on the lawn and the
00:49:08
MS-13 knuckle tattoos. Yeah. Okay. So
00:49:10
let's uh wrap up on this just really
00:49:12
lightning round here. Amazon
00:49:14
flip-flopped on a new tariff
00:49:15
notification on their websites. Trump
00:49:19
said he had a great discussion with
00:49:20
Bezos. He solved the problem very
00:49:22
quickly. He did the right thing, good
00:49:24
guy, etc. If you haven't seen this, it's
00:49:28
something that Teimu is doing. Here's
00:49:29
what Timu does uh currently today. Nick,
00:49:33
you have that image. If you could pull
00:49:34
it up of just when there is a tariff,
00:49:37
they explain the tariff coming into the
00:49:39
country. They put it as like a line
00:49:41
item. I thought this was actually kind
00:49:44
of cool. I don't know why people take
00:49:45
offense, Brian. This is pretty standard
00:49:47
stuff. So Amazon's competitor Tim Teimu
00:49:49
is putting in the import charges. They
00:49:51
don't say tariffs. They don't say taxes.
00:49:53
Import charges. This is like a standard
00:49:54
thing. This happens in other countries
00:49:56
too. What is this? What is Teeu? This is
00:49:58
like a dollar store. It's basically like
00:50:00
a dollar. Is it this the last place you
00:50:01
would ever
00:50:02
buy jeans? Yeah, you can buy $12 jeans.
00:50:06
Basically, your left sock from Laura
00:50:08
Piana costs less than Timu's entire
00:50:12
inventory of jeans. The point being, um,
00:50:15
I thought this was actually a plus for I
00:50:19
think they totally misplayed this. They
00:50:21
they they did it, they rolled it out,
00:50:23
then they got criticized. I think they
00:50:24
were called a treasonous company from
00:50:25
the White House press uh, you know, by
00:50:27
the press secretary. Uh, they totally
00:50:29
misplayed this because they should have
00:50:30
gone and leaned into it and said, "Yeah,
00:50:32
we're showing you all these tariffs when
00:50:34
you buy from China. If you buy from
00:50:36
America, you don't have to pay any
00:50:37
tariff and look at all these other
00:50:39
products." Come on. Come on. That would
00:50:40
have lasted three and a half seconds.
00:50:41
This is exactly consistent with the
00:50:43
other issue which is they're playing
00:50:45
whack-a-ole. Okay, we're going to do
00:50:46
something with the automakers. We're
00:50:48
going to we're going to try and solve
00:50:49
some problem with Amazon. Like like this
00:50:52
is a sign that that it like it's not a
00:50:55
good strategy if you have to do this
00:50:56
much whack-a-ole. Like like they they're
00:50:58
not like they can't cover up what Amazon
00:51:01
is going to end up dealing with because
00:51:02
there's going to be 500 other retailers
00:51:04
that that don't get the call with Trump.
00:51:06
So, so this is like that that like to me
00:51:08
that's evidence of of clearly this they
00:51:11
didn't think through the entire
00:51:12
downstream set of of of conditions that
00:51:14
are going to change as a result of this.
00:51:16
Sure. Yeah. I I thought this was a big
00:51:19
win uh Chimoth because they could then
00:51:23
have Amazon Here's um a mockup somebody
00:51:25
made. I'll pull it up here. It was
00:51:27
interesting. They could to Aaron's point
00:51:30
just show hey here's a bunch of American
00:51:32
companies by American when you do a
00:51:33
search. Ryan's point. I'm sorry, Ryan's
00:51:36
point. Hey, here's what it might look
00:51:37
like. Pull pull that um the OralB
00:51:39
toothbrush one up, Nick, if you got it
00:51:41
right there. So, somebody mocked this
00:51:43
up. I think this could be the hugest
00:51:44
win. You could have the retailers do buy
00:51:48
American, buy it once, buy a high
00:51:50
quality product from America. If you
00:51:52
look here, we don't have the products.
00:51:54
It wouldn't work. We don't have the
00:51:56
products.
00:51:58
Well, I mean, we do have for some
00:51:59
products, you know, Americanmade
00:52:01
products. Uh, you know, I buy my boots
00:52:03
from Danner and those are all American.
00:52:05
Yeah. So, we should just go back to
00:52:07
communism and we're all going to make
00:52:08
our our shoes. Like, it's like that's
00:52:10
like we we're in a global market. Like,
00:52:12
we buy from everywhere. Chamop, any
00:52:14
thoughts on this? I mean, obviously
00:52:16
there's the whack-a-ole angle. There's
00:52:19
buy American and be proud of it. There's
00:52:22
communication. Here's the narrow
00:52:23
question. I I got a bunch of emails from
00:52:26
people and a bunch of them were Amazon
00:52:29
sellers and I don't know Nick if you can
00:52:32
find it but I posted their comments and
00:52:34
I reshared them just to kind of
00:52:36
highlight the issues that they were
00:52:38
going through and at the core of it was
00:52:41
a feeling by them that Amazon had
00:52:45
abandoned them as American purveyors and
00:52:48
sellers of goods and that Amazon on the
00:52:52
margins had attended
00:52:54
to
00:52:56
help competitors from abroad come stand
00:53:00
themselves up and compete and
00:53:03
essentially cannibalize on price and
00:53:04
margin. This is my view completely and
00:53:06
that this is the biggest opportunity
00:53:08
that I think the Trump administr
00:53:10
administration's flying at 40,000 ft
00:53:11
doing macrolevel negotiations and look
00:53:13
and failing to see some of these micro
00:53:15
optimizations that are really really
00:53:16
real. So you in the United States you
00:53:18
can import goods as a foreign company.
00:53:21
You don't not you do not have to create
00:53:22
an LLC or any sort of registered entity
00:53:25
in the United States to import goods.
00:53:27
Sometimes they say, "Oh, Americans pay
00:53:28
the tariff." Like that is not true. In
00:53:31
many many cases, the foreign company
00:53:32
just imports this stuff and they sell on
00:53:34
Amazon. And when they get caught
00:53:36
cheating, they can they can lie about
00:53:38
the valuation and pay a lower tariff.
00:53:40
They can change the classification and
00:53:41
pay a lower tariff. They can import
00:53:43
stuff that, you know, is harmful to
00:53:44
children, has lead, paint, whatever
00:53:46
else. There's no enforcement at all. You
00:53:48
can't. So you're saying Amazon third
00:53:50
party is like a bit of a backdoor to
00:53:52
abuse the system, right? I mean, Amazon
00:53:53
is sort of just playing the game that's
00:53:55
on the field, but they're this is legal
00:53:57
in the United States is these companies
00:53:58
import stuff and I think it's 60% of all
00:54:00
of all the sellers on Amazon are these
00:54:02
Chinese registered. They're not
00:54:03
registered in the United States at all.
00:54:04
Just Chinese companies. Which sounds
00:54:06
profoundly unfair in terms playing
00:54:09
field. Can we just take a step back and
00:54:10
also acknowledge that we are talking
00:54:12
about a level of detailed issues that we
00:54:16
would never have talked about 6 months
00:54:17
ago. That there was
00:54:19
no interest in even bringing this up.
00:54:22
Like if if Ryan wanted to bring up the
00:54:25
sort of hollowing out of
00:54:28
American salesmanship, let's say, if you
00:54:31
will, because of like this arbitrage
00:54:33
that Amazon does for GMV. That would
00:54:36
have been a snoozefest. except today it
00:54:39
can actually get a lot of awareness and
00:54:41
Erin mentioned this and I've mentioned
00:54:43
this before but like I think that there
00:54:45
are four things that really matter
00:54:46
batteries AI pharma APIs and rare earths
00:54:50
that now is on the agenda I think the
00:54:51
the positive way to look at this is the
00:54:54
American economy is too complicated if
00:54:56
you had waited Aaron for a study for all
00:54:59
of the implications we would have been
00:55:01
waiting forever and nothing would have
00:55:03
happened and I think that we've made
00:55:05
macrolevel moves you're right and now we
00:55:07
are finding what the implications are in
00:55:10
course correcting in real time. And I
00:55:12
hope what happens though is when we find
00:55:14
these big thorny issues, I think the
00:55:16
Amazon thing is a is a pretty
00:55:18
interesting issue actually about like
00:55:19
American competitiveness. Now the
00:55:21
question is do we follow through and get
00:55:23
to the root cause of it and and fix it.
00:55:26
And those feedback loops are there. I
00:55:27
mean the Trump administration is going
00:55:28
to act on this and if then there's an
00:55:30
act that's coming out of Congress as
00:55:31
well to shut down the foreign import of
00:55:33
records. So those feedback loops are
00:55:34
there in ways that I don't I don't know
00:55:36
if they were there in the past. Go
00:55:37
ahead, Erin. We can totally chaos monkey
00:55:38
the the economy and just see what
00:55:41
breaks. I think that the you know you
00:55:44
you you sort of phrase the we could, you
00:55:46
know, do the research paper and we could
00:55:48
do, you know, Aspen Institute as a bad
00:55:50
thing, but also it can be a bad thing if
00:55:53
you're the small business owner right
00:55:54
now who has 30 employees and you just
00:55:56
literally don't know what you're going
00:55:57
to do next month. And so, so that that's
00:56:00
that's sort of then the argument to not
00:56:02
counterbalance. That's like why you do
00:56:04
have some bureaucracy and and you don't
00:56:05
chaos monkey the economy and why Rand
00:56:07
Paul is literally saying we shouldn't
00:56:09
actually let you know have unilateral
00:56:11
you know control over tariffs. So you
00:56:13
know interesting that that like dynamic
00:56:16
there. All right, you want to uh wrap us
00:56:19
up here or you want to pass? Oh, am I
00:56:22
still on the pod? Yeah. Well, you turned
00:56:25
your camera off. So, look, I spent 80
00:56:28
minutes debating this topic with Larry
00:56:30
Summers three weeks ago.
00:56:32
The point I made then is that we had in
00:56:36
this city at for 25 years a globalist
00:56:38
consensus on trade that distorted a lot
00:56:42
of outcomes. And I don't need to rehash
00:56:44
that debate, but I'll just recall that
00:56:47
Larry Summer's main argument for why
00:56:50
this would not work out is that the
00:56:52
market was down. Do you remember that?
00:56:53
That was his evidence. Yeah. That this
00:56:56
wasn't going to work. Okay? And it was
00:56:58
all about the market not pricing in
00:57:00
lower expectations. Well, guess what?
00:57:02
The market is actually up since
00:57:04
Liberation Day on April 2nd. So, what
00:57:07
happened 3 weeks ago was basically a
00:57:10
panic in the market over this policy and
00:57:13
the media has been trying to fuel that
00:57:15
panic. Now, what I said as well is we do
00:57:18
have to stick the landing on this. I
00:57:20
mean, President Trump shifted the
00:57:21
conversation away from this globalist
00:57:24
consensus and he's now redefined the
00:57:27
debate, but it is now up to Scott
00:57:29
Bessant, the Treasury Secretary, Howard
00:57:31
Lutnik, the commerce secretary, Jameson
00:57:33
Greer, the US trade rep, and so on, the
00:57:35
Trump trade team, to now negotiate these
00:57:38
deals, stick the landing. And I agree
00:57:40
with you to the extent that the sooner
00:57:42
that is done, the better because it is
00:57:44
good to provide business certainty. But
00:57:47
the idea that so far this hasn't worked.
00:57:50
I think again the main argument against
00:57:51
that was the market reaction that now
00:57:54
the market's not positive. So I think my
00:57:56
point is just we need to give this time
00:57:59
to work. I think it's too soon to be
00:58:01
judging this policy as if it hasn't
00:58:02
worked yet. It needs to be executed
00:58:04
properly. And quite frankly, Ryan, I
00:58:06
mean, I remember the last time you were
00:58:07
on this pod, you were coming on about,
00:58:11
wasn't there like some union deal that
00:58:13
was supposed to shut down all the ports
00:58:14
and all the shelves would be empty? That
00:58:16
never happened either. It did happen. It
00:58:18
did. They shut down for 3 days. I don't
00:58:20
Okay. I don't remember the the shelves
00:58:21
being empty, which is now the new panic
00:58:24
the media is trying to create. So, look,
00:58:26
there's a lot of pants wedding that's
00:58:28
occurring here that's being fueled by
00:58:29
the media. Well, well, I I do want I
00:58:31
want I do want to I just want to
00:58:32
bookmark one one thing. The only thing I
00:58:33
actually um I was more frustrated
00:58:35
listening to the Larry Summers and and
00:58:37
you conversation because I was like, why
00:58:38
Larry make this point like come on.
00:58:40
Like, what? Don't go down the WTO rat
00:58:42
hole. Like, that's that's not relevant.
00:58:43
So, here here's the other thing. Of
00:58:45
course, it's relevant. It's how we got
00:58:46
here. No, no. Like like as in like
00:58:49
that's 25 years ago. Like let's worry
00:58:50
about literally today and like what what
00:58:52
do we do going forward today? And the
00:58:55
only thing I just want to say because I
00:58:57
because I I do think that that you know
00:58:59
I appreciate your point about hey
00:59:01
there's like you know everybody's
00:59:02
freaking out whatever but to be totally
00:59:04
fair that some of that freak out whether
00:59:07
we can decide how emotional it needs to
00:59:08
be is the reason that then Trump walks
00:59:10
back the things that then cause the
00:59:12
market to correct. So, so we can't just
00:59:14
say the market's back and and like see
00:59:16
we didn't need to freak out cuz it was
00:59:18
literally I said we have to make the
00:59:19
deals. We have to stick to the landing.
00:59:21
But look, China over the last 25 years
00:59:23
has been able to strategically
00:59:25
annihilate our rare earth processing
00:59:28
capability and our ability to cast rare
00:59:30
earth magnets. We just sat back and
00:59:32
watched as the market basically went to
00:59:35
the lowest bidder which was being
00:59:36
subsidized by the Chinese government
00:59:38
which the WTO allowed them to do. And
00:59:40
now we have a critical dependency. Yep.
00:59:42
In our supply chain on China for
00:59:45
basically every electric motor in every
00:59:47
product, including cars. That was crazy.
00:59:49
We should not have allowed that to
00:59:50
happen. How are you going to change
00:59:52
that? So, we needed to shift the
00:59:54
political conversation to recognize the
00:59:57
ways in which free trade led to unfair
01:00:00
trade and created unacceptable
01:00:02
dependencies on the for the American
01:00:04
economy. Wait, wait, just I just want to
01:00:07
make one one point. It's more than that.
01:00:09
point and then I'll but this is the
01:00:11
national security of the United States
01:00:13
that that's at stake. Let's go take your
01:00:16
favorite pet issue. China invades
01:00:18
Taiwan. Okay. And we have to take a
01:00:21
side. Jason, my pet is hold on. Let me
01:00:24
just finish. And the and the and the
01:00:27
Chinese say here are the implications of
01:00:30
supporting Taiwan on this A, B, C, and
01:00:33
D. You don't get any pharma APIs. You
01:00:35
don't get any rare earths. You don't get
01:00:36
any batteries.
01:00:38
Okay, it'll it'll send life back 50
01:00:41
years. Or let's say China and India get
01:00:44
into a fight and we're forced to pick a
01:00:46
side. Same situation. The point is
01:00:48
there's all these scenarios that we
01:00:50
never even
01:00:52
considered us being able to have
01:00:55
strategic optionality to make the
01:00:56
decision that's morally and ethically
01:00:58
right for the United States. And I think
01:01:00
that we have learned through this lens
01:01:03
that these are huge issues. The thing
01:01:05
that the Chinese did that was so
01:01:06
brilliant, which we still don't have an
01:01:08
answer for is they have these national
01:01:10
champions. And being a national champion
01:01:13
allows you, and we'll talk about this in
01:01:14
AI, it allows you to blur the lines
01:01:17
between the public and private
01:01:18
partnership. It allows you to blur the
01:01:20
law. It allows you to blur capital. And
01:01:22
I'm not saying we have to do that, but
01:01:25
what I am saying is we need to have our
01:01:27
own answer to it. And that was never on
01:01:30
the table until April 9th.
01:01:33
All right. So yeah, but but like 100%
01:01:36
like do that strategy and then and then
01:01:38
and then don't have a mad rush. What is
01:01:40
the strategy? No, no, no. Because
01:01:41
because that's a you if you have a, you
01:01:43
know, Ryan, what what is the number? I
01:01:44
don't know, trillion of imports or
01:01:45
whatever, like you don't need everybody
01:01:47
then than then jamming the system to
01:01:49
build their supply chain, you know, in
01:01:51
the US to solve that problem
01:01:52
immediately. If we're sitting here in 9
01:01:54
months and you're saying this and there
01:01:56
are no deals, I would say that you're
01:01:58
right. what Howard Lutnik said last week
01:02:01
and again we may have all gotten caught
01:02:03
up in the knuckle tattoos and we missed
01:02:04
this but he was very clear we have a
01:02:07
country a deal is already done we're
01:02:09
convening parliament it's going to be
01:02:11
the first of many so for all we know
01:02:14
there's like 30 deals that are waiting
01:02:16
in the wings and the first one will set
01:02:18
the tone and and I think that Sax is
01:02:21
right here which is it's way too early
01:02:23
to declare defeat and that it was quote
01:02:25
unquote chaos I think if we're sitting
01:02:27
here in 9 months and foreign direct
01:02:29
investment has shriveled up and domestic
01:02:32
investment has shriveled up because
01:02:34
there is just no continuity. You have a
01:02:36
claim. But that's No, no, because I I I
01:02:38
don't I don't think that'll happen. I
01:02:40
don't I don't think that'll happen. I
01:02:41
think we will end up in like I'm I'm
01:02:42
with Ryan like we will end up in a good
01:02:44
spot because we'll iterate through this.
01:02:45
My my only point is there's an
01:02:46
alternative path that that could have
01:02:48
occurred. It could have been done in a
01:02:50
more thoughtful well-communicated
01:02:52
pattern instead of hey let's do barrel
01:02:53
rolls with the airplane. I don't
01:02:55
disagree with you, Aaron, and I can tell
01:02:56
you we've already started to see layoffs
01:02:58
and nobody wanted to even initiate the
01:03:00
barrel roll, guys. Yeah, listen, we get
01:03:03
it. It's like, hey, I don't want
01:03:05
anything to change. I think we agree to
01:03:07
disagree on this. We got to move on to
01:03:09
the next one. Hold on. This one last
01:03:11
point. Excuse me. Excuse me. You didn't
01:03:13
call me for 40 minutes. I just want to
01:03:14
make one final point. Eron, here he
01:03:16
goes. Erin, where were you with this
01:03:18
perfect plan? Yeah. Where were you with
01:03:21
this perfect plan before Liberation Day?
01:03:23
I was telling Kla about it. You were
01:03:25
telling Kamla. Okay, great. They were
01:03:27
having nobody nobody ties and they were
01:03:31
talking about this specific issue. All
01:03:33
the people who suddenly know what the
01:03:35
perfect plan is and how to perfectly
01:03:37
execute it. No barrel rolls had nothing
01:03:39
to say about this topic for 25 years.
01:03:41
Now all of a sudden they've come forward
01:03:43
with their perfect plans. I would say
01:03:46
that's victory for Trump. Finish. It's
01:03:47
like look at this. The best thing of all
01:03:49
of this is you've got the Liberals
01:03:50
embracing Milton Friedman and their
01:03:51
backgrounds on there. I love it. Yes. Uh
01:03:54
All right. Listen, we're going to agree
01:03:55
to disagree. We're going to agree to
01:03:56
disagree on this one. You know, the
01:03:58
Liberals love the stock market. Listen,
01:04:00
Sachs Kla is coming on next week. We're
01:04:02
going to make some cocktails. It's going
01:04:03
to be wonderful. You we'll ask her some
01:04:06
direct questions about it. But I want to
01:04:08
talk about AI agents. 2025 shaping up to
01:04:11
be the year of AI agents. Tons to talk
01:04:14
about here. Open AI is planning to
01:04:16
charge between two and 20K a month for
01:04:18
different levels of AI agencies would be
01:04:21
basically cron jobs they would run in
01:04:22
the background and do things for your
01:04:24
company that uh humans are doing right
01:04:26
now. You may have heard of this uh
01:04:28
agentic tool. Again agentic is just a
01:04:30
fancy word for agent which is a fancy ro
01:04:32
word for like a cron job that just runs
01:04:35
uh perpetually. Yep. And uh Manis is uh
01:04:39
the company in China that started this
01:04:41
weirdly benchmark invested in it that's
01:04:43
created a whole bluff on the side and
01:04:46
Manis' website has a really good
01:04:47
visualization of what these agents would
01:04:49
look like. So first of all I think
01:04:50
you're giving a little too much credit
01:04:51
to Manis. They didn't come up with the
01:04:53
agents but I do think that they have a
01:04:55
very good demo and it's hard to know
01:04:57
exactly how real it is because not
01:04:59
everyone's used it and it's from China.
01:05:01
It's from China. I'll get to that in a
01:05:02
second. If you go to their website, you
01:05:05
can see a bunch of their demos. And I do
01:05:08
think that what they deserve credit for
01:05:09
is advancing the ball on the UI
01:05:13
paradigm. And it's not that other people
01:05:16
weren't doing this. I mean, I think
01:05:17
notably Anthropic was doing this with
01:05:18
its operator product, but the basic idea
01:05:21
is that you've got this two-pane view
01:05:23
and in one window, you've got the
01:05:25
standard chatbot interface and then in
01:05:27
the other view, you can see what the
01:05:29
agent is doing. And that agent has the
01:05:32
ability to toggle between it currently
01:05:34
four apps. There's search, browser,
01:05:37
code, terminal, and document editor. And
01:05:40
so when you give Manis a task, the first
01:05:42
thing it does is create a to-do list in
01:05:44
the document editor. You can kind of see
01:05:46
it there. And then it works sequentially
01:05:48
to achieve each of those tasks and then
01:05:49
puts an X on them there. And you can
01:05:52
kind of see it working. And I think
01:05:53
what's cool about the demo is just the
01:05:55
way that it seamlessly toggles between
01:05:57
those four apps. and you can see what
01:05:59
the AI agent is doing, you know, and
01:06:01
it's browsing the internet, it's
01:06:02
searching for things, it's writing
01:06:04
documents, it's crossing things off its
01:06:06
to-do list. Now, I think it's pretty
01:06:08
easy to imagine where this goes, which
01:06:10
is you'll be able to connect an agent to
01:06:13
all of your SAS apps. So, it won't just
01:06:15
be four applications. It'll now be
01:06:17
connected to dozens of applications,
01:06:19
including ones that already have your
01:06:21
data. And it's going to know what
01:06:23
actions it's possible to take in those
01:06:24
apps. So when it creates its to-do list,
01:06:26
there's a much wider range of things
01:06:28
that it can accomplish. And in fact,
01:06:30
there's a new standard called MCP which
01:06:33
is taking off like wildfire which is
01:06:35
built specifically to enable agents to
01:06:37
connect with applications and understand
01:06:39
the data and understand the actions that
01:06:40
are possible in those SAS applications.
01:06:43
So look, Manis is just at the tip of the
01:06:45
iceberg here. I think this will become a
01:06:46
very standard UI paradigm. That's the
01:06:48
reason why I mention it. Not because I
01:06:50
am predeclaring them to be the winner in
01:06:52
the space, but just because I think
01:06:54
there's a lot of talk about agents and I
01:06:56
think it's hard to conceptualize what
01:06:59
that means without just seeing it
01:07:00
visually. A great great summary there,
01:07:02
Saxs. And Aaron, I want to get your
01:07:04
thoughts on it because obviously you're
01:07:06
running Box and and you have your your
01:07:08
finger on the pulse of this. We actually
01:07:10
started building one of these in our
01:07:11
venture firm. We have 20,000
01:07:14
applications a year and we have updates
01:07:16
coming in from investments. We are now
01:07:18
taking those sacks Aaron and we are
01:07:20
having an agent sort them and then look
01:07:23
for competitors and compare them to the
01:07:25
last update and we're looking into our
01:07:27
notion our kod and saying what else have
01:07:30
what other communications have we had
01:07:32
what questions should we ask about the
01:07:33
startup and about their strategy and
01:07:35
then we're presenting that in Slack to
01:07:37
our team. So this is coming fast and
01:07:39
furious and we spend I don't know
01:07:42
probably 15 minutes on each of those
01:07:43
incoming applications. You start doing
01:07:45
the math on that. Talk about 5,000 hours
01:07:47
of work. Aaron, what are you seeing on
01:07:49
the street? What are you doing at Box in
01:07:51
terms of agents landing right now in Q2
01:07:55
of 2025? Uh, yeah. I mean, I think I
01:07:58
think Sax represented it well, which is
01:08:00
which is, you know, you have to now
01:08:02
think about AI as as effectively being
01:08:04
able to do anything on a on a computer
01:08:06
or another piece of software as as a
01:08:08
human can do. And the little distraction
01:08:11
that that I think happened two years ago
01:08:13
after the chatbt moment was we sort of
01:08:15
thought about that as oh we're just
01:08:17
going to now you know do like typing
01:08:18
information retrieval and that's a new
01:08:20
paradigm for user interfaces let's say
01:08:22
so you just like talk to your software
01:08:24
and you like search Zillow via chat that
01:08:27
was sort of a little bit of a
01:08:28
distraction that that's super helpful
01:08:29
like when you want basic information
01:08:31
lookup or whatnot the big breakthrough
01:08:32
was starting to think through these
01:08:34
things as as full you know effectively
01:08:37
uh uh you know agentic systems that that
01:08:40
operate on any amount of data, any
01:08:42
amount of tools for as long as you want
01:08:44
to complete any task that you want. And
01:08:46
this is sort of the big year where
01:08:48
agents are starting to, you know, enter
01:08:50
the vocabulary of enterprises, of IT
01:08:53
people, of, you know, larger and and
01:08:55
certainly small organizations. Um, and
01:08:58
and it kind of requires you to have a
01:08:59
little bit of a of a reset moment on how
01:09:01
you think about AI, which is which is
01:09:03
it's not just now a kind of a co-pilot
01:09:05
that you talk back and forth with. it's
01:09:07
actually something running behind the
01:09:09
scenes that's now actually starting to
01:09:11
deliver, you know, real automated, you
01:09:14
know, kind of work for you. And so lot
01:09:16
lots of implications like, you know,
01:09:18
massive implications to what the
01:09:19
software business model is in the
01:09:20
future. You know, I I I would argue, you
01:09:22
know, strongly that's a massive TAM
01:09:23
increase um because now software starts
01:09:26
to go after labor spend. It completely
01:09:28
changes the dynamics of then, you know,
01:09:29
how do you build a moat in a world of AI
01:09:31
agents? Um uh but but I think unpack
01:09:34
that piece there Erin. You said
01:09:35
something very interesting how you how
01:09:37
software then is going to go over human
01:09:40
spend. Yes. Explain that concept. Unpack
01:09:43
it for a second. David and I, you know,
01:09:46
we we we go back way back in SAS land,
01:09:48
but like you used to basically just, you
01:09:50
know, you built you built a piece of of
01:09:52
software and you sell it for the number
01:09:54
of of people in the organization. And
01:09:56
so, you know, company has 500 employees
01:09:59
and uh and you sell that thing for,
01:10:01
let's say, $10, you know, a user a
01:10:03
month, you know, 120 bucks a year and
01:10:05
you make 60,000 bucks. Um, and so so
01:10:07
that that's kind of the business model.
01:10:09
Uh, now when your software actually
01:10:11
brings the underlying workflow to the
01:10:14
customer or the underlying outcome to
01:10:16
the customer, so you know that that
01:10:18
company might have 10 lawyers and so
01:10:20
previously if you were selling software
01:10:21
for lawyers, you had a maximum amount of
01:10:23
10 seats that you could sell. Now all of
01:10:25
a sudden if your AI agents are doing the
01:10:27
equivalent of let's say parallegal work
01:10:29
or some form of professional services
01:10:31
all of a sudden you might be able to
01:10:32
sell a multiple of the initial kind of
01:10:34
10 seats that you would have sold
01:10:35
previously. So you see it as a huge
01:10:37
opportunity because now you're not
01:10:39
enabling a human to be 5% more
01:10:41
productive. You're replacing a human or
01:10:42
you're replacing one out of 10. Yeah.
01:10:44
And and actually I'm going to take a
01:10:45
massive I'm going to do an underscore on
01:10:47
this point though. I I don't I don't
01:10:48
like the word replace because I I think
01:10:50
actually most of the upside is actually
01:10:52
going to be for companies that now
01:10:53
deploy labor at things that they
01:10:55
wouldn't have deployed labor at before.
01:10:57
And and you know, maybe I'm biased from
01:10:58
the view we have, but most of our
01:11:00
conversations with customers are when
01:11:02
they have AI agents, it they can
01:11:04
actually now go and actually deliver
01:11:06
work in areas that would have been
01:11:08
unaffordable previously. So they
01:11:09
actually they weren't doing the work.
01:11:12
True in logistics. So, like we we're
01:11:14
making thousands of phone calls a day
01:11:15
using AI, calling truck drivers, and we
01:11:18
if we have a load, we've got 400,000
01:11:20
truck drivers using the mobile app, the
01:11:22
Flexport mobile app. I don't have enough
01:11:24
loads to keep them all checking it every
01:11:26
day to see if there's a load that
01:11:28
matches them. And if they don't check
01:11:29
it, I they're they're useless to me now.
01:11:32
And it was too expensive to call the
01:11:33
truck driver and have a human call and
01:11:35
talk to them, even if it's a human in a
01:11:36
call center in the Philippines. Whereas
01:11:38
with AI, it's almost free. I'm calling
01:11:40
thousands of them a day going, "Hey,
01:11:42
this load looks like it's a good match
01:11:43
for you. Are you interested?" And and
01:11:45
then we activate them on the platform.
01:11:47
Um, that's new work that wasn't going to
01:11:49
happen before, not just a replacement.
01:11:51
Yeah, I I think that I think probably I
01:11:53
think like we we have, you know, in the
01:11:55
valley, unfortunately, we've been
01:11:56
co-opted a little bit somewhat with with
01:11:58
a little bit of a doomer, you know,
01:12:00
mindset in some areas and and we think
01:12:02
of then AI is, okay, like like it's all
01:12:04
fixed pie, it's going to replace things.
01:12:06
And on the ground with large
01:12:08
enterprises, the vast majority of the
01:12:10
use cases are are it's it's the you
01:12:13
know, it's the ability to finally review
01:12:14
the contracts that we never got around
01:12:16
to reviewing. It's the ability to
01:12:17
finally automate an invoice process that
01:12:19
we never did. It's the ability to go in
01:12:21
and just create marketing campaigns in
01:12:23
every language that we never got around
01:12:25
to. And so so I think that'll probably
01:12:27
be actually like 90% of the usage of AI
01:12:29
in the future will be things that if we
01:12:31
look back and we snap the line right now
01:12:32
and we said this is what knowledge work
01:12:34
is today. 90% of AI usage will be things
01:12:37
that we don't do today. 10% will will
01:12:39
replace you know what we're what we're
01:12:40
doing in some areas. I think that's the
01:12:42
right take cuz Chimath I can tell you in
01:12:45
our firm we would never have associates
01:12:46
or researchers or analysts sax you also
01:12:49
were in this line of work as well
01:12:51
venture capital you would never have
01:12:52
them review legal documents. That's
01:12:54
something lawyers would do in the legal
01:12:55
department. But now because of AI, we
01:12:58
have them say, "Here's the safe. Here's
01:12:59
the term sheet. Here's the edited
01:13:01
version, dump it all in, find out what
01:13:04
the changes are, what are the deltas
01:13:06
here, and and then let's have a
01:13:07
discussion about what the founder
01:13:08
changed in a Sander document, and we
01:13:11
don't have to bother with an attorney,
01:13:13
and maybe you wouldn't have even checked
01:13:14
those documents if you were, you know, a
01:13:16
seed fund or an angel fund. You would
01:13:18
just go along for the ride because
01:13:19
you're the 10th person signing the
01:13:21
documents." So what what do you think
01:13:22
Chimoth here in terms of the premise
01:13:24
that maybe it's 10% replacing work
01:13:27
that's happening but this is blue ocean
01:13:29
and we're going to do 90% of like new
01:13:30
stuff that we just never got to. Yeah, I
01:13:33
tend to I tend to believe that's true. I
01:13:34
think the customers that we sell into at
01:13:38
8090 are largely large enterprises as
01:13:41
well. So not dissimilar to Aeron's
01:13:42
customer base. What I would say is that
01:13:45
what they are encountering is the trough
01:13:49
of disillusionment.
01:13:52
And I don't know if Erin, you're seeing
01:13:53
this as well, but every single, you
01:13:56
know, CIO ran
01:13:58
around signing up some sort of AI
01:14:01
product in large part because their CEO
01:14:04
would say to them, hey, what's your AI
01:14:07
strategy? And the reason the CEO asked
01:14:09
them that is that at some point somebody
01:14:10
on the board said, what are we doing
01:14:12
about AI? So that's the the
01:14:15
cascade that that that we went through
01:14:17
in the last two years. And I think what
01:14:20
has happened now is people have spent
01:14:22
billions and billions of dollars. I
01:14:24
think you can see it in the revenue
01:14:25
traction of the AI
01:14:28
companies. But I think where we are
01:14:30
today is that there are some real
01:14:33
technical complexities that have not
01:14:35
been solved. I'll give you an example.
01:14:37
We have a lot of customers in regulated
01:14:39
industries which is to say that if you
01:14:42
make a
01:14:43
mistake you will get fined or you will
01:14:46
get shut down. Life sciences,
01:14:48
healthcare, financial services are three
01:14:52
examples. People still don't seem to
01:14:54
appreciate that when you replace
01:14:56
software that is deterministic with
01:14:59
software that is probabilistic, meaning
01:15:01
software that somebody wrote for
01:15:04
you, do A then do B, then do
01:15:08
C with an LLM that can
01:15:11
hallucinate, you'll have
01:15:14
errors. So what used to be a throwaway
01:15:17
thing, which is quality assurance and
01:15:18
QA, right? unit testing, integration
01:15:21
testing is now the only thing that
01:15:22
matters. Why? Because if you're a
01:15:24
financial services institution and
01:15:26
you're supposed to do KYC and hit BORS
01:15:28
and now all of a sudden you send a wire
01:15:30
somewhere in Syria, guess what? You're
01:15:32
in trouble. Bueno. Yeah. If you're a
01:15:34
healthcare company and you're supposed
01:15:36
to do some clinical diagnosis to send
01:15:38
out a drug on time and you don't do that
01:15:41
because the the model
01:15:42
hallucinates, that's a real problem. And
01:15:45
I am guaranteeing you, we have not seen
01:15:49
the class action lawsuits that will come
01:15:52
when those errors will eventually be
01:15:54
made. They're guaranteed to be made. We
01:15:57
just don't know the scope and the scale
01:15:58
of them. So that's why I'm sort of of
01:16:01
this posture where I think we've sold in
01:16:06
a ton of promise. I think the reality is
01:16:09
much more tactical. It's a little bit
01:16:11
more benal. I think we're sorting
01:16:13
through the exact use cases where you
01:16:15
can put guard rails around these error
01:16:18
rates where it's okay and tolerable.
01:16:19
Like Ryan will probably tell you there's
01:16:22
some number of phone calls that just
01:16:23
sound totally fcocked but he's okay with
01:16:27
that because the broader thing is okay
01:16:29
and Aaron will so I don't know. So I
01:16:32
don't want to talk to my customers
01:16:33
though. I have it calling truck drivers
01:16:35
to offer them you know offer them loads
01:16:37
but I'm not having to talk to my
01:16:38
customer. Sorry, I I meant your truck
01:16:40
drivers, but my my my point just is that
01:16:43
I think agents are
01:16:45
real, but I think that we are far away
01:16:48
from that because we're still at the
01:16:49
phase of how do you build reliable
01:16:52
software in production for an enterprise
01:16:55
versus the toy apps that you see on the
01:16:58
internet which is like let me vibe code
01:17:00
something. I think these things are
01:17:02
worlds apart still. Okay, so let me get
01:17:03
saxed in on here and just to inform the
01:17:05
audience you heard tri of
01:17:06
disillusionment. This comes from the
01:17:08
hype cycle. This is something Gartner
01:17:10
has been talked about for a long time.
01:17:11
So in case you're taking it for granted
01:17:13
if you're watching, you have some sort
01:17:15
of technology trigger like agents. You
01:17:16
have this like peak of inflated
01:17:18
expectations. Now we're in the trial of
01:17:20
disillusionment. Hey, this stuff doesn't
01:17:21
work. It's hallucinating. But we're kind
01:17:23
of going up that works. It just doesn't
01:17:26
say we're on the slope of I don't see
01:17:28
the disillusionment. I don't know where
01:17:29
this is coming from. I don't even think
01:17:30
we're at the peak yet. Oh, okay. So you
01:17:33
think we're still going up? cuz I I a
01:17:35
lot of people to Shimat's point were
01:17:37
buying stuff and saying, "Hey, it
01:17:38
doesn't work, you know, and now we're in
01:17:40
the mess." Let me say let me say it
01:17:42
differently, Sax. I think we have not
01:17:44
yet figured out how to move the budgets
01:17:46
from experimentation to mainline
01:17:49
production. Meaning where large chunks
01:17:51
of the US economy are comfortable enough
01:17:55
with the ways in which hallucinations
01:17:57
are managed such that they will replace
01:18:00
legacy deterministic code with this new
01:18:03
probabilistic model generated code
01:18:06
meaning model enabled code. Let's just
01:18:08
put it that way.
01:18:11
Where are we on the slope here? Yeah.
01:18:12
Look, I would I would separate change
01:18:15
management issues, which are always
01:18:16
going to be important and there's always
01:18:18
going to be big ones whenever there's a
01:18:19
big disruption, especially in enterprise
01:18:21
and especially around compliance and
01:18:22
legal and all that kind of stuff. I
01:18:23
would separate that from the impact of
01:18:25
the underlying technology trend. And I
01:18:28
don't think the impact has come anywhere
01:18:30
close to peaking yet. And in fact, I
01:18:33
would say the rate of progress is
01:18:35
exponential right now on at least three
01:18:38
key dimensions. So number one is the
01:18:40
algorithms themselves. The models are
01:18:42
improving at a rate of I don't know
01:18:43
three to four times a year. They're not
01:18:45
just getting faster and and better, but
01:18:49
qualitatively they're different.
01:18:50
Remember, we started with pure LLM chat
01:18:52
bots. Then we went to reasoning models.
01:18:55
And the difference there is with a
01:18:56
chatbot, I just it's like kind of a
01:18:58
smart PhD or college student giving you
01:19:00
an answer off the top of their heads.
01:19:02
The reasoning models, it's more like the
01:19:04
PhD saying, "Okay, let me go off and
01:19:06
think about that. Let me do a project on
01:19:07
that." And it could work for 30 seconds
01:19:10
or a couple of minutes. I mean, as much
01:19:11
compute as you want to throw at it and
01:19:13
it will break down your complicated
01:19:15
question into a bunch of sub questions
01:19:16
and then it'll try different approaches
01:19:18
and it can validate some of those
01:19:19
approaches and come back to you with a
01:19:21
much more impressive answer. And if
01:19:22
you've been using like the Gro 3 deep
01:19:25
research or the new Chad
01:19:29
GBT3 to do these types of new reasoning
01:19:32
models, it's pretty mind-blowing what
01:19:34
they're capable of. Have we even come
01:19:36
close to figuring out how to tap the
01:19:38
potential there, especially in an
01:19:39
enterprise context? No, but my point is
01:19:42
that the rate of progress on the
01:19:43
algorithms is again three or four
01:19:47
times
01:19:49
here. Okay, go finish about sex and then
01:19:52
I'll I'll take it and pass it. Go ahead.
01:19:53
Well, I was trying to lay out the
01:19:54
dimensions of which progress is
01:19:56
proceeding exponentially. Okay, so one
01:19:58
is the algorithms. Okay, which is not
01:20:00
just quantitative, it's also
01:20:01
qualitative. We didn't even get to the
01:20:03
agents part of it yet, but that's the
01:20:05
next big leap after reasoning models.
01:20:06
We're just starting to scratch the
01:20:08
surface there. Then you've got the
01:20:10
chips. I mean, the chips are getting
01:20:12
better at, I don't know, three to 4x a
01:20:14
year. We've gone from, you know, the
01:20:15
H100 to the H200. Now we're on the
01:20:18
GB200. We'll be a GB300 soon. We'll be
01:20:22
on to three times better per year or
01:20:24
they get better three times per year.
01:20:26
No, no, no. They're getting the chips
01:20:27
themselves, depending on how you measure
01:20:29
it. Each generation of chips is probably
01:20:31
three or four times better than the
01:20:32
last. Okay. and Nvidia is back to
01:20:36
rolling out new chip, new generation of
01:20:38
products roughly annually and I'm just
01:20:40
using them as one example. Obviously
01:20:42
there are other companies as well. So
01:20:44
basically the lead from Hopper to
01:20:46
Blackwell to got it Reuben I guess will
01:20:48
be in next year and and then I think
01:20:50
Fman's coming after that. I mean really
01:20:52
an astounding rate of progress. It's not
01:20:54
just the individual chips are getting
01:20:55
better. They're figuring out how to
01:20:56
network them together like with NVL72.
01:20:58
It's like a rack system to create much
01:21:01
better performance at the data center
01:21:03
level. And that would be like the the
01:21:04
third area where you're seeing basically
01:21:06
exponential progress. Just look at the
01:21:09
number of GPUs are being deployed in
01:21:11
data centers. So when Elon first started
01:21:13
training Grock, I think they had maybe
01:21:15
100,000 GPUs. Colossus was 100,000.
01:21:17
Correct. Right now they're up to
01:21:18
300,000. They're on the way to a
01:21:20
million. Same thing with OpenAI's data
01:21:23
center, Stargate. And within a couple
01:21:25
years they'll be at I don't know 5
01:21:27
million GPUs, 10 million GPUs. So and
01:21:29
you see that on the power side, right?
01:21:31
You're going from 100 megawatt data
01:21:32
centers to 300 megawatts to we're just
01:21:35
starting to now see the first gigawatt
01:21:37
power data centers. I don't even think
01:21:39
they're live yet, but this is where
01:21:40
they're trying to get to. And I don't
01:21:42
think it's beyond the real possibility
01:21:44
that we could be at 5 or 10 gawatt data
01:21:46
centers in the next I don't know several
01:21:48
years. So, so my point is just look, the
01:21:50
algorithms, the chips, and the data
01:21:53
centers are all improving or scaling at
01:21:56
a rate of, I don't know, 3 to 4x a year.
01:21:58
That's 10x every 2 years. Okay? Where
01:22:01
people don't understand exponential
01:22:02
progress is that if you're getting
01:22:04
better at 10x every 2 years, that
01:22:06
doesn't mean you'll be at 20x in four
01:22:08
years. It means you'll be at 100x. 100x.
01:22:11
So the models, the chips, and the data
01:22:12
centers will all be 100 times more
01:22:14
powerful in four years, let's say at the
01:22:16
end of of this presidential term. So you
01:22:18
multiply those things together, the
01:22:21
algorithms, the chips, and then the raw
01:22:23
compute that's available, you're talking
01:22:25
about a millionx increase, some of which
01:22:28
will be captured in price reductions,
01:22:30
some of it will be in the performance
01:22:31
ceiling, and then some of it will just
01:22:33
be in the overall amount of of AI
01:22:37
compute that's available to the economy.
01:22:40
But the impact of this thing is going to
01:22:42
be absolutely massive and I think people
01:22:43
still don't even appreciate that fact
01:22:45
because they don't understand
01:22:45
exponential progress. Yeah. And I think
01:22:47
maybe just to square the circle the the
01:22:49
because because everything is that that
01:22:52
you just said tax is what I think is
01:22:54
propelling the industry and then the
01:22:55
reality on Jama's side like like just
01:22:57
just to connect the dots. So uh we have
01:23:00
an eval test that we do uh where we run
01:23:03
enterprise data through every model to
01:23:05
to kind of figure out its accuracy rate
01:23:07
and and you know how much it's not even
01:23:09
hallucination but just literally how
01:23:10
much data does it miss when we ask for
01:23:12
facts. The best model in the world um
01:23:14
actually interestingly we was gro three
01:23:16
on on this particular test. We send it
01:23:18
500 documents and we ask for 40 data
01:23:21
fields back from the documents and so it
01:23:23
has to get every single data field
01:23:24
correct and we only do a single pass. So
01:23:27
we send the document to to the to the
01:23:29
model, we get a single pass back. Right
01:23:31
now the the best score is about 90%. Um
01:23:34
and so you can imagine a number of
01:23:35
industries where you can't have 90%
01:23:37
accuracy, you know, if you give
01:23:39
something, you know, a question on 40
01:23:41
data fields. Now there's ways to solve
01:23:42
it is you rerun it multiple times or you
01:23:45
chunk up the document into smaller parts
01:23:47
and so it doesn't get confused by the
01:23:49
large context window. But a lot of the
01:23:51
people that were deploying AI a year ago
01:23:53
or a year and a half ago weren't doing
01:23:55
that. And so they they you know they did
01:23:57
have a a kind of a a pilot run of
01:23:59
something and it it kind of worked okay.
01:24:01
And what they have to realize back to
01:24:03
your point sax is like this space is
01:24:05
literally exponentially you know
01:24:06
changing and so if you don't use the the
01:24:08
latest methods of okay you have to
01:24:10
actually like run the data through the
01:24:12
model multiple times um and you have to
01:24:14
chunk up the data into smaller parts and
01:24:16
you have to use a reasoning model and
01:24:17
you have to make sure that your your
01:24:18
prompt is like hyper tuned for the
01:24:20
particular use case. If you haven't done
01:24:21
those four things, then you probably
01:24:23
will actually end up with a project that
01:24:25
fails. And and even so, even when you do
01:24:27
all those things for even, you know,
01:24:29
harder problems, you know, you're still
01:24:30
going to run into issues. So, I think I
01:24:32
think the challenge is that everybody's
01:24:33
running a million miles an hour right
01:24:34
now and they're trying a lot of things.
01:24:36
Some work, some don't work at the same
01:24:38
time that the space is actually, you
01:24:40
know, you know, changing at a at a
01:24:41
pretty uh, you know, kind of crazy rate.
01:24:43
and let's take a look at our partner
01:24:45
Poly Market and uh which company they
01:24:47
think will have the best AI model by the
01:24:49
end of 2025 and get feedback from our
01:24:51
panel on if you think this is accurate
01:24:54
and who you would pick here. Looks like
01:24:57
Google is in the lead here. 41% of
01:25:01
people believe that they will have the
01:25:05
best AI model by the end of the
01:25:06
question. What is what is the dimension
01:25:08
like you know we use Gemini so for many
01:25:11
tasks at 8090 we use Gemini it's
01:25:13
incredible but for most of our codegen
01:25:15
we use anthropic and claude kicks ass
01:25:17
it's it's exceptional this is uh based
01:25:20
on the best scores in the chatbot arena
01:25:23
which just became a fourth company so
01:25:26
that is slightly different because
01:25:27
people have
01:25:28
gamed tests so that is a rub there
01:25:31
people are now building their a model
01:25:33
for the AVAL unit right all the models
01:25:34
are way overfitted for these eval so but
01:25:36
if you had to pick who's your I mean so
01:25:38
I guess Chimath you're saying you have
01:25:39
to take a task by task it's what depends
01:25:41
on task I agree what Zach said is right
01:25:43
so it's kind of like what what problem
01:25:45
are you trying to solve and then you
01:25:46
have to ride this technology wave that
01:25:48
is compounding very quickly all I was
01:25:50
just trying to get across is that the
01:25:51
error rates have been diminishing but
01:25:54
not nearly as fast as you need for some
01:25:55
sectors of the economy so you can use a
01:25:58
model to generate deterministic code
01:26:00
that's great and as long as you unit
01:26:02
test it and integration test it'll be
01:26:04
fine but I'm saying if you're going to
01:26:05
use a model in production
01:26:07
in an environment where if there are
01:26:10
consequences
01:26:12
we're not there
01:26:15
yet but you could use it for writing or
01:26:18
writing jokes or
01:26:20
maybe but that's that's too binary like
01:26:23
it's already used right now in
01:26:24
healthcare but it's just the doctor's
01:26:26
meeting notes that that would normally
01:26:28
take 30 minutes to go and transcribe
01:26:30
yeah so so it's we you can't be too too
01:26:32
black and white on that one yeah what's
01:26:33
happening right now the reason why the
01:26:35
the progress is so rapid in coding
01:26:37
assistance and I I think you know you're
01:26:39
right that enthropic with was it claw
01:26:43
3.7 37 yeah yeah I think they're the
01:26:45
leader and in fact I think the manis
01:26:47
demo that we showed it's not entirely a
01:26:49
rapper on clawed because they actually
01:26:51
they do a number of different things but
01:26:52
I think they are significantly using
01:26:54
anthropic for the code assistant part of
01:26:56
it in any event the reason why the
01:26:59
progress is so rapid with coding is
01:27:01
because code compiles and you can
01:27:04
determine objectively whether it works
01:27:06
or not, you can validate it. And so that
01:27:09
makes it a perfect area for AI to get
01:27:13
better at through reinforcement learning
01:27:14
and test time compute is AI tries a
01:27:16
bunch of things. It sees what works. It
01:27:18
sees what compiles, sees what the user
01:27:20
then accepts, and then be able to learn
01:27:22
and iterate based on that. That's why
01:27:24
coding right now is really the big
01:27:26
breakthrough application and use case.
01:27:28
But it's not going to be the only one. I
01:27:29
mean, math is another good area where I
01:27:31
think AI is improving rapidly again
01:27:33
because math you have proofs and you can
01:27:37
look at the results and see if it
01:27:39
validates. Now, I think one of the big
01:27:41
questions in terms of AI progress is how
01:27:44
extensible is the progress to other
01:27:46
areas that don't easily validate that
01:27:47
way. So, for example, legal work is I
01:27:52
think a really good area for AI, but how
01:27:54
do you validate that it's correct? You
01:27:56
would have to go to a court, right? the
01:27:57
court is the compiler like a lawsuit is
01:27:59
the compiler or maybe the laws or you
01:28:02
could hire like a thousand lawyers or
01:28:04
experts in an area to basically do you
01:28:07
know reinforcement people are doing
01:28:09
people people are doing but it's not
01:28:10
like a compiler to your point it's not
01:28:12
like a it's not the progress isn't going
01:28:13
to be as rapid because it's harder to
01:28:15
validate but absolutely but my guess is
01:28:17
that once they figure out how to nail
01:28:20
coding math and the things that are
01:28:22
easily validated they can move to the
01:28:23
things that are harder to validate and
01:28:26
but I I think this is one of the big
01:28:28
questions is whether I think people just
01:28:30
kind of assume that AI progress will be
01:28:32
you know equally fast in all areas and I
01:28:34
think it's possible that AI gets really
01:28:36
good in some areas better than human but
01:28:38
it's sort of childlike in other areas
01:28:40
and um narrow possible outcome is make
01:28:43
your point right like with a finite
01:28:45
answer or an answer we know is the
01:28:47
definitive well this is but but this is
01:28:49
the the important thing about the agent
01:28:52
kind of you know let's just say uh
01:28:54
framework or architecture uh you
01:28:56
momentum was was instead of just saying
01:28:59
okay we're going to do a single pass
01:29:00
through the model and then whatever it
01:29:02
comes back with is is we're going to be
01:29:03
satisfied um like you know the legal
01:29:06
work might be might be reviewed by
01:29:08
another agent whose job is to review
01:29:10
legal work and so we we can just throw
01:29:12
more and more compute at the problem uh
01:29:14
and we're just early in figuring out how
01:29:16
to architect those or multiple models
01:29:18
right you could
01:29:20
have you have enthropic checkp check
01:29:23
Gemini yeah when you have some anomaly
01:29:25
it spits back out to the user. So human
01:29:27
in the loop still matters in this type
01:29:28
of process. In the early days of OCR,
01:29:31
you would have a computer say here's the
01:29:33
characters in this legal document. Then
01:29:34
you'd have two humans type it in and
01:29:36
then you would get a certain level of
01:29:37
certainty. You'll quickly find that when
01:29:39
you layer these models on top of each
01:29:41
other, the test time compute costs are
01:29:43
astronomical. Y and Aaron's probably
01:29:46
dealt with this like it's like I get a
01:29:47
bill from AWS and it's like oh wait,
01:29:51
hold on a second. I just, you know, per
01:29:54
100,000 this month. what's going on. So
01:29:56
we have to get to the bottom. That that
01:29:58
by the way is another major trend line
01:30:00
which is that the new applications that
01:30:01
we talked about are all much more token
01:30:03
intensive. So we went from basic LLMs
01:30:06
totally you know which don't require
01:30:08
that many tokens to give you an answer
01:30:09
to the reasoning model where you can
01:30:11
spend a thousand times more tokens just
01:30:13
getting one answer to a question and now
01:30:15
the agents are going to be even more
01:30:17
token intensive than that. So the amount
01:30:19
of compute required to serve all these
01:30:21
new applications is going to be massive
01:30:23
which is why I think the capex buildout
01:30:25
actually makes sense when you do a deep
01:30:27
research to your point David you're
01:30:29
firing off maybe 200 queries and it's
01:30:31
asking them the AI is saying hey what
01:30:33
query should I ask on behalf of the user
01:30:35
and then you go down that rabbit hole
01:30:36
you it's basically like doing 200 of
01:30:38
them at once Ryan uh your thoughts here
01:30:40
on AI first companies and agentic
01:30:42
computing well the one I really wanted
01:30:44
to tie back to was actually our earlier
01:30:45
conversation on tariffs and there's a
01:30:47
very real use case of of LLMs is how do
01:30:50
you classify a product and you'd like to
01:30:52
get to what we see today when we did our
01:30:55
first machine learning based natural
01:30:57
language classification of a product you
01:30:59
take a product URL listing page a
01:31:01
Shopify page or Amazon page and say hey
01:31:03
what what classification code is this
01:31:05
what duty is owed six years ago in a
01:31:08
hackathon we got to like 70% accuracy
01:31:11
we're now in high 90s accuracy versus
01:31:14
what a human trained expert will get to
01:31:16
but you actually get to which is not
01:31:18
good enough. You know, you're wrong 3%
01:31:20
of the time. You know, you might have
01:31:22
committed a violation of the law for
01:31:23
sure. But actually, what is truth in
01:31:26
that regard? It's there's a lot of gray
01:31:27
area in this. And truth ultimately is
01:31:29
what does customs say? What does the CBP
01:31:32
determine is correct? And those guys are
01:31:35
using software, right? That's pretty
01:31:38
that's a very simple algorithm and it's
01:31:39
a decision tree that's going, okay, is
01:31:41
it a shoe? Yes. Is the top made of
01:31:43
leather? Yes. You know, is the bottom
01:31:45
made of rubber? and they just go through
01:31:47
a very simple and that outputs it. And
01:31:49
so on some level, if you convince the
01:31:51
government to use your LLM, it becomes
01:31:53
true whether it's true or not. I think
01:31:55
there's going to be some interesting
01:31:56
cases like that that we haven't really
01:31:57
thought through of like when does the
01:31:59
government adopt these to be the source
01:32:01
of truth. Okay. All of this speaks to
01:32:03
this thing that's going to sound totally
01:32:05
esoteric, but like we all used to on QA,
01:32:10
right? Like the least talented engineers
01:32:12
were allocated to QA. I think in the
01:32:15
world of AI it's it'll end up being the
01:32:17
most talented. You know we internally at
01:32:19
8090 we call it improvement engineering
01:32:22
and it's a total specialty. It's similar
01:32:24
to when I kind of coined the growth team
01:32:27
at Facebook. I feel it's the same kind
01:32:29
of moment where improvement engineering
01:32:31
is really the skill that translates toy
01:32:35
apps and vibe coding into something
01:32:37
that's very practical and real. And we
01:32:40
like and my team and the leader of this
01:32:42
team, he's like steeped in things like
01:32:44
Japanese kata management from like
01:32:46
Toyota and quality systems and these are
01:32:48
all the things that matter when you're
01:32:50
trying to just shrink the error rate
01:32:52
down to zero so that you can use it in a
01:32:54
reliable way and also to document it so
01:32:57
that if people want to question what
01:33:00
happened or have you know recompense or
01:33:03
some some way to come back and say hey
01:33:05
that that really harmed me, how do you
01:33:07
even do that like these are all very
01:33:10
complicated issues that that will get
01:33:12
sorted up. Super I think interesting.
01:33:14
Okay, four. I got to wrap guys. I got to
01:33:17
catch a flight to Miami. Let me do a
01:33:19
closing here. If you want to keep going,
01:33:20
you're welcome to. Three, two, the plane
01:33:21
just waits. Just text the pilot and just
01:33:23
tell them you're all right. Listen, I'm
01:33:25
not burning all the allin credits, so to
01:33:29
speak, and all of our tokens. I'm
01:33:30
kidding. I'm kidding. I'm kidding.
01:33:34
private to everything and then putting
01:33:35
it on the allin budget and the rest of
01:33:37
us are flying southwest
01:33:39
for
01:33:41
your dictator Jim Paul. I miss a flight.
01:33:44
It's a strange concept. Yeah, David s.
01:33:46
What does that mean? David, when's the
01:33:47
last time you flew commercial? Clinton,
01:33:49
I haven't missed a flight in about 15
01:33:51
years
01:33:53
for Ryan Peterson from Flexport. Aaron
01:33:56
Lee from the amazing box. Chim pitia.
01:33:59
David Saxs your chairman dictator Zar. I
01:34:02
am the world's greatest moderator. Love
01:34:04
you, boys.
01:34:07
We'll let your winners ride.
01:34:09
Rainman David S.
01:34:14
We open sourced it to the fans and
01:34:16
they've just gone crazy with it.
01:34:19
Queen of
01:34:21
[Music]
01:34:27
besties
01:34:27
[Music]
01:34:29
are my dog taking notice your driveways.
01:34:34
Oh man, my habitasher will meet up. We
01:34:37
should all just get a room and just have
01:34:39
one big huge orgy cuz they're all just
01:34:40
useless. It's like this like sexual
01:34:42
tension that we just need to release
01:34:43
somehow.
01:34:45
Wet your feet. Wet your feet. Your feet.
01:34:50
That's going to be good. We need to get
01:34:51
merch. I'm going all
01:34:55
[Music]
01:34:59
in. I'm going all in.
