00:00:00
All right, everybody. Welcome back to
00:00:01
the number one podcast in the world.
00:00:03
Freeberg's out saving the world,
00:00:04
creating new potatoes or I don't know,
00:00:06
quinoa, maybe some Brussels sprouts. I'm
00:00:08
not sure what he's working on at this
00:00:10
point. In his place, his personal
00:00:12
favorite bestie. He always says that
00:00:13
when I'm not here, I want Brad Gersonner
00:00:16
in the seat. Welcome back. We haven't
00:00:18
seen you on the pod since your shout out
00:00:22
at the State of the Union. Take us
00:00:24
behind the scenes for a brief moment
00:00:25
here, Brad, of what it's like to get a
00:00:28
shout out from POTUS at the state of Did
00:00:31
you know it was coming? Did you
00:00:32
choreograph to this thing? Did you Did
00:00:34
you choreograph that or was that more
00:00:35
spontaneous?
00:00:36
>> I honestly had no idea it was coming.
00:00:38
>> And in fact, I I found out after the
00:00:40
fact that it wasn't in the speech and
00:00:42
the president added it to the speech.
00:00:44
So, I don't even think it was, you know,
00:00:46
few days ago few days before going to
00:00:48
happen. But we got an invite to the
00:00:50
State of the Union and you know, listen,
00:00:52
it's an institution. And this has
00:00:53
happened every year for 250 years in the
00:00:58
country. I've never been. I thought I
00:00:59
did know he was going to talk about
00:01:00
Trump accounts. So I figured if I'm ever
00:01:02
going to go, that's the time to go. And
00:01:04
I have to say, you know, I'm just a
00:01:06
sucker for Democratic institutions and
00:01:08
Democratic traditions. It it was an
00:01:10
extraordinary night. Set aside, you
00:01:12
know, the headlines about what Democrats
00:01:14
did or Republicans did. Just the whether
00:01:15
it's a Democrat president or Republican
00:01:17
president that this happens every year.
00:01:19
You have to go report on the State of
00:01:20
the Union. So it was a it was a special
00:01:22
night. did dinner ahead of time. We're
00:01:24
in the chamber. The chamber, as you all
00:01:25
know, is very small. And so, you know,
00:01:28
just to your right was the first family
00:01:31
and and Jared and Ivanka.
00:01:34
And so, you know, we were there as to
00:01:37
observe like everybody else. And wow, it
00:01:40
was it was quite a moment. And
00:01:42
>> I want to just say you did a great job
00:01:44
because when you sent your heart out to
00:01:47
all of America,
00:01:48
>> I took it. I took it. I took it. I was
00:01:49
like it out,
00:01:51
>> but you kept it at the right angle.
00:01:53
>> Right. Right.
00:01:54
>> You had just gone up a little bit extra
00:01:56
and out.
00:02:00
>> Would have been no bueno.
00:02:01
>> Those would be some super racist Trump
00:02:03
accounts.
00:02:04
>> Keep your protractor and your ruler out
00:02:07
when you send your heart out. Okay.
00:02:08
>> One final thing on it, Jason.
00:02:10
You know, we're signing up over a
00:02:12
100,000 kids a day to these Trump
00:02:14
accounts. Fantastic. We have millions of
00:02:16
kids who've already claimed their
00:02:17
account. We have nearly 30 million kids
00:02:19
in America who are eligible for at least
00:02:21
$250 if they just go claim their
00:02:24
account. These things are going to go
00:02:25
live on July 4th. And what it really
00:02:27
showed, I think, the country, it
00:02:28
accelerated after the State of the Union
00:02:30
because the president, you know, really
00:02:32
believes this is a way to get everybody
00:02:34
main street America into the game of
00:02:36
capitalism and get them all directly
00:02:38
owning, you know, the great companies in
00:02:40
America. So, it meant a lot to me in
00:02:42
that regard that it highlights the
00:02:44
importance of the program. So, I was
00:02:45
deeply grateful to the president for not
00:02:48
only making sure this happens, but the
00:02:49
shout out is pretty cool.
00:02:50
>> Good for you, bro.
00:02:51
>> I have um an interesting uh idea for
00:02:53
you. I'm sure it's come up already, but
00:02:56
with this whole discussion of UBI,
00:02:58
somebody said to me, "Oh, you know, I
00:02:59
really like these Trump accounts your
00:03:00
friends did the the Invest America
00:03:02
because it's like the start of UBI." And
00:03:04
I was like, "Well, that's not exactly
00:03:05
the intention, but I get it." And with
00:03:08
wealth disparity going on in the
00:03:09
country, that has a lot of people
00:03:11
concerned. What if there was a giving
00:03:13
pledge around equities and people could
00:03:17
opt into it? They don't have to, but if
00:03:18
somebody like, I don't know, Larry and
00:03:21
Sergey or Zuckerberg said, I want to
00:03:23
pledge 5% of my shares to go into kids
00:03:26
accounts over the next 20 years. What an
00:03:30
amazing, beautiful thing that could be.
00:03:31
And it would be incredibly material to
00:03:33
get whatever it is, a tenth of a share,
00:03:35
a hundredth of a share, a thousandth of
00:03:36
a share of whatever company. Has that
00:03:38
come up yet as an idea? I'm sure it's
00:03:40
obvious, right?
00:03:41
It's come up. Stay tuned. But yes, we're
00:03:44
going to have some banger announcements
00:03:47
as we head toward July 4th.
00:03:49
>> All right, let's talk about the war in
00:03:50
Iran. Obviously, there are much more
00:03:52
important issues than financial ones.
00:03:54
Life, death, the freedom of the people
00:03:57
of Iran, but we're uniquely qualified, I
00:03:59
think, to talk about the economic
00:04:01
fallout, second order effects, first
00:04:03
order effects. And there has been
00:04:05
massive volatility over the last five
00:04:07
trading days. Just talking about Brent
00:04:10
crude oil and we'll we'll key the
00:04:12
discussion off of that type of oil. It
00:04:15
spiked to $84 on Friday. That was day
00:04:17
seven of the war, 119 on Monday, day 10
00:04:21
dropped back down to 84. Jump backed up
00:04:23
to 100 after three commercial ships were
00:04:26
hit in the straight on Wednesday. Those
00:04:28
ships, by the way, were not oil tankers.
00:04:30
Uh they were carrying cargo. They were
00:04:32
flagged as Thai Japanese and Marshall
00:04:34
Islands. Brent crude currently at 99
00:04:36
when we're taping this. It'll be at
00:04:38
something different by the time you
00:04:39
listen to the pod, I'm sure. But it's uh
00:04:42
quite a spike. And here's a second
00:04:44
chart. This shows you the spikes over
00:04:46
time. I was old enough to remember the
00:04:47
oil shock of 1978. Uh we had to like get
00:04:50
in line at the gas station based on your
00:04:52
license plate number and uh you had to
00:04:54
wait an hour or two to get gas. Gulf War
00:04:56
obviously it hit $100 in 20 $26.
00:05:00
2008 we hit kind of a peak moment. $216
00:05:04
in today's dollars. That was the peak
00:05:06
oil discussion. Demand from China went
00:05:08
off the charts. When Russia invaded
00:05:10
Ukraine, we hit 115, which would be 133
00:05:14
in today's dollars. So, this is uh not
00:05:16
new, but it is significant. And breaking
00:05:21
news today, Iran's new Supreme Leader
00:05:23
Mosh Taba says he's keeping the straight
00:05:26
closed as a tool to pressure the enemy.
00:05:29
Wall Street Journal on Thursday quoted a
00:05:31
senior fellow at the Middle East
00:05:33
Institute saying that reopening the
00:05:35
strait will require ground troops. Poly
00:05:37
market 27% chance that US forces enter
00:05:41
Iran by the end of March and 57% by the
00:05:44
end of the year. So the sharps over at
00:05:45
poly market believe we will have boots
00:05:48
on the ground. Let me stop there. Brad,
00:05:50
your thoughts on what happens when oil
00:05:52
hits this kind of number and we have
00:05:54
this uncertainty of, hey, this could
00:05:56
last, you know, two more weeks or it
00:06:00
could last 6 months, it could last a
00:06:02
year. Nobody seems to know and how it
00:06:05
resolves. We just had a really
00:06:08
interesting talk with Graham Allison.
00:06:10
Um, how it resolves is also a major
00:06:13
unknown. Your thoughts?
00:06:14
>> Right. So first is obviously there are
00:06:17
huge direct costs as oil prices go up
00:06:20
right oil is a component of a lot of
00:06:21
consumer you know and enterprise
00:06:23
products and it also hurts consumer
00:06:25
confidence enterprise confidence Goldman
00:06:27
Sachs is out today with some analysis
00:06:29
where they updated kind of the economic
00:06:31
knock-on effects right so they raised
00:06:33
their PCE inflation forecast from 2.1 to
00:06:36
2.9 right for the year so that's a huge
00:06:39
jump right in terms of their expected
00:06:43
you know PC inflation core PCE which
00:06:45
excludes oil, okay, is they forecasted
00:06:48
up uh from 2.2 to 2.4. So they're saying
00:06:51
even if you excluded the direct price of
00:06:54
oil, the knock-on effects is going to
00:06:56
cause a little more inflation. They
00:06:57
lowered their GDP forecast by 30 basis
00:06:59
points for the year. And they also
00:07:02
expect higher unemployment as a result
00:07:04
of this for the year. All of that is
00:07:06
weighing on the sentiment of the market.
00:07:07
Remember, just a few months ago, the S&P
00:07:09
peaked at 24 times. Now we're at 21
00:07:11
times. But I think the market may be
00:07:13
getting it a little bit wrong. Right?
00:07:15
The Trump doctrine, you know, I tweeted
00:07:17
about this last week. I think the Trump
00:07:18
doctrine is far more pragmatic than the
00:07:22
neocon doctrine. Right? I think we I
00:07:24
think Trump has a very limited set of
00:07:26
goals. He wants to destroy and degrade
00:07:27
threats to America's national security
00:07:29
interests. He doesn't want to spread
00:07:31
democracy. So my suspicion is that these
00:07:34
impacts are shorter duration, but right
00:07:36
now the market's having a little bit of
00:07:38
post-traumatic stress flashbacks to
00:07:40
Afghanistan and Iraq and wondering if we
00:07:42
might be wandering into a quagmire.
00:07:45
>> All right. And just he may not um in
00:07:49
terms of the doctrine, he has said he
00:07:50
wants to see uh the people rise up
00:07:52
there. So might be splitting hairs, but
00:07:54
I think uh he might actually be for
00:07:57
regime change. He says he wants the
00:07:58
regime to change
00:07:59
>> being equal. I don't think he minds if
00:08:01
the people bring it to themselves. The
00:08:02
question is whether the US is going to,
00:08:04
you know, put boots on the ground and
00:08:06
try to spread misonian democracy like
00:08:08
like the Cheney doctrine was. And I
00:08:10
think this is very different.
00:08:11
>> Chimat, uh, your take on the economic
00:08:13
impact and, you know, and any other uh,
00:08:16
things you'd like to add about the war
00:08:19
in Iran?
00:08:20
I think the most important thing that I
00:08:23
saw this week was I think President
00:08:26
Trump was asked about the war and he
00:08:28
said the war would be over very soon.
00:08:32
What did the market do? The market
00:08:34
literally took oil from 120 a barrel to
00:08:37
90 a barrel
00:08:40
almost in you know a nancond.
00:08:43
I think that that sort of tells you what
00:08:46
everybody thinks. To the extent that the
00:08:50
market really didn't believe it, oil
00:08:52
would not have budged and if anything,
00:08:54
it would have faded those comments and
00:08:56
you probably would have seen oil stay at
00:08:58
around 120 or even go slightly higher.
00:09:02
So the fact that there was this
00:09:03
reflexive move, I think is a belief by a
00:09:06
lot of the sharps that there is no path
00:09:09
to a sustained conflict. there's going
00:09:11
to be a lot of chest bumping from the
00:09:13
Iranians obviously because they need to
00:09:14
save face and they will want to set up
00:09:17
whoever comes next to have the most
00:09:19
successful chance of governing. So my
00:09:21
perspective is that was a trial balloon.
00:09:24
I think it validated what everybody
00:09:25
thought which is that this is going to
00:09:27
be a shortrun thing. I agree with that.
00:09:31
The downstream impact is, I think,
00:09:33
correct what Brad said, which is it
00:09:35
could show up in some short-term price
00:09:38
spikes. But then on March 11th, you saw
00:09:41
what Chris Wright did, which is the
00:09:44
president activated a whole bunch of
00:09:46
member countries in the IEA, and I think
00:09:48
Chris released 172 million barrels. I
00:09:51
think there's a coordinated release of
00:09:52
about 400 million barrels of petroleum.
00:09:56
That's going to dampen the effect of any
00:09:57
price spike. On top of that, I think the
00:10:00
estimate is there's probably another
00:10:02
billion or so more barrels that one
00:10:04
could
00:10:05
release out of strategic stock piles.
00:10:08
So, I think that both of these two
00:10:10
things together kind of paint a picture
00:10:12
that probably the worst is behind us.
00:10:14
And I think now it's about finding the
00:10:16
off-ramp.
00:10:19
Sax, your thoughts?
00:10:22
>> Well, I agree that we should try to find
00:10:23
the off-ramp. I mean, I agree with what
00:10:25
Brad and Jamas said about that. Look,
00:10:28
we've degraded Iranian capabilities
00:10:30
massively. Their army, navy, air force,
00:10:33
all been destroyed. This is a good time
00:10:36
to declare victory and get out. And that
00:10:37
is clearly what the markets would like
00:10:39
to see. You are seeing, however, a
00:10:42
faction of of people, I'd say largely,
00:10:45
but not exclusively in the Republican
00:10:47
party, who want to escalate the war and
00:10:49
who are calling for things like ground
00:10:52
troops or regime change or they simply
00:10:56
want the pounding of Iran to just keep
00:10:58
going on and on. I saw an oped in the
00:11:00
Wall Street Journal to that effect that
00:11:02
we shouldn't try and find an offramp, we
00:11:03
should just keep going with this. And I
00:11:06
just want to lay out, I think, some of
00:11:08
the risks of what an escalatory approach
00:11:12
could entail. So, first of all, we're
00:11:14
all seeing that the the straits for
00:11:16
moves are closed right now. We don't
00:11:18
want that to persist longer than it has
00:11:21
to, but there are actually worse
00:11:23
outcomes than that. So if the Iranians
00:11:27
get hit, if their oil and gas
00:11:28
infrastructure gets hit, they've already
00:11:31
said they're going to engage in tit
00:11:33
fortat retaliation against the Gulf
00:11:34
States. And we saw there was recently um
00:11:37
the Iranians blew up this giant oil
00:11:39
depot in Oman. You saw some of those
00:11:42
images. They could continue to target
00:11:44
the oil and gas infrastructure across
00:11:47
the Gulf States. And if that happens, it
00:11:50
won't really matter if the straits get
00:11:51
reopened because you won't be able to
00:11:53
restart oil and gas production in the
00:11:55
Middle East. So that would be, I think,
00:11:57
a much worse outcome that could result
00:11:59
from escalation.
00:12:01
Furthermore, there's an even worse, I
00:12:03
think, scenario from there, which is the
00:12:06
region is very dependent on desalination
00:12:08
plants. I think something like 70% of
00:12:11
Riad gets their water from desalination.
00:12:13
I think it's something like 100 million
00:12:14
people on the Arabian Peninsula that get
00:12:17
their water from Delaw. I mean, it's
00:12:19
basically a desert, right?
00:12:21
>> And those DEL plants are soft targets.
00:12:23
You already saw there was I think there
00:12:25
was one diesel plant in Iran that got
00:12:27
hit and then it caused Iran again tit
00:12:29
for tat to hit a delaw plant. I think it
00:12:32
was in Kuwait. I could be off about
00:12:33
that.
00:12:34
>> But any event, if you see that type of
00:12:36
destruction continue, you could
00:12:38
literally render the Gulf almost
00:12:41
uninhabitable. I you're not going to
00:12:43
have enough water for 100 million people
00:12:45
and human beings just cannot survive
00:12:46
very long without water. So that would
00:12:49
be a truly catastrophic scenario and
00:12:52
we're talking about destroying the Gulf
00:12:55
States economically and then also from a
00:12:57
humanitarian perspective. So I think we
00:12:59
have to take things like this into
00:13:00
account when you hear people preaching
00:13:02
for or advocating for escalation. You
00:13:05
also have to I think consider the
00:13:06
impacts on on Israel. I mean, it's hard
00:13:09
to know exactly how much damage Israel
00:13:11
is taking right now. There's a social
00:13:13
media blackout, but what you're starting
00:13:16
to hear trickle out is that Israel is
00:13:18
getting hit harder than they've ever
00:13:19
been hit before in their history. And
00:13:21
we're only two weeks into this. If this
00:13:22
war continues for weeks or months,
00:13:26
then Israel could just be destroyed.
00:13:29
There are very large parts of it. Now, I
00:13:31
think Israel is a harder target than the
00:13:33
Gulf States. Their infrastructure is
00:13:35
more hardened. Also, they're further
00:13:36
away. The Gulf States are vulnerable to
00:13:39
drones and short-range missiles, whereas
00:13:42
Israel is mainly vulnerable to
00:13:43
long-range missiles. Nonetheless, at
00:13:46
some point, their air defenses could
00:13:48
become exhausted if it hasn't happened
00:13:50
already. And Israel could get seriously
00:13:52
destroyed. And then you have to worry
00:13:53
about Israel escalating the war by
00:13:56
contemplating using a nuclear weapon,
00:13:58
which would truly be catastrophic. So
00:14:02
there's a lot of scenarios here, a lot
00:14:04
of really
00:14:06
frightening scenarios about where
00:14:07
escalation could lead. And even though
00:14:10
the United States is a much more
00:14:11
powerful country than Iran, they
00:14:13
essentially have a dead man's switch
00:14:16
over the economic fate of the Gulf
00:14:18
States and even potentially beyond that,
00:14:21
you know, the habitability of some of
00:14:23
these these countries. So I do tend to
00:14:26
think that this is a good time to
00:14:28
declare victory. I think Brad, you're
00:14:30
right that the president has never said
00:14:32
that democracy promotion is one of his
00:14:34
objectives. Yes, Jal, obviously everyone
00:14:37
would welcome if the people rose up and
00:14:39
chose a new regime, but that's not
00:14:41
something that we've said we have to
00:14:43
accomplish. And this would be a really
00:14:46
good time to take stock of where we are
00:14:48
and try, I think, to seek an off-ramp.
00:14:51
And look, if escalation doesn't lead
00:14:53
anywhere good, then you have to think
00:14:55
about, well, how do you deescalate? And
00:14:57
deescalation, I think, involves reaching
00:14:59
some sort of ceasefire agreement or some
00:15:02
sort of negotiated settlement with Iran.
00:15:05
And we can get into more of what that
00:15:06
looks like. But I think that's the big
00:15:08
picture is that if escalation could lead
00:15:11
in all these horrifying directions, then
00:15:12
I think that's not the right approach.
00:15:14
You have to look at deescalation.
00:15:16
>> Jake, where are you on this?
00:15:18
>> Complicated. I have my personal feelings
00:15:19
on regime change. And since we don't
00:15:22
have the information uh the MSAD and the
00:15:25
CIA and Trump has, I do think Trump
00:15:28
would only do this if he had a very high
00:15:30
probability of success and an off-ramp.
00:15:33
However, it's not looking good with the
00:15:35
offramp right now and it could be quite
00:15:37
chaotic. I think if the neocons get
00:15:40
their way and the people on poly market
00:15:43
are correct, the sharps who say 57%
00:15:46
chance we'll have boots on the ground, I
00:15:47
think this is kind of the end of Trump's
00:15:50
second term. And if you were to put
00:15:52
together the series of mistakes that
00:15:55
he's made um and the administration's
00:15:58
made, they're really at the heart of why
00:16:02
people voted for him. You take starting
00:16:04
a war like this specifically with Iran.
00:16:06
That's what we were told was the reason
00:16:09
to vote for President Trump. He was not
00:16:11
going to take us down this path. He was
00:16:13
not going to
00:16:16
risk World War II. He was not going to
00:16:17
risk a nuclear possibility as as Sax
00:16:21
correctly points out. And now you have
00:16:24
all the MAGA supporters from, you know,
00:16:26
Tucker to MTG to Rogan, Matt Walsh,
00:16:29
Megan Kelly, they are all up in arms
00:16:31
about this is the end of MAGA and this
00:16:34
is, you know, a massive betrayal.
00:16:36
There's the 1B betrayal where Trump
00:16:39
wouldn't release the Epstein files.
00:16:40
We'll put that aside because I don't
00:16:41
think that's as important as starting
00:16:43
World War II. Um, and then obviously the
00:16:46
insane unnecessary cruelty we talked
00:16:48
about many times on this podcast of ICE
00:16:50
agents which he has corrected by getting
00:16:53
rid of Christy Gnome. So you start
00:16:54
putting these things together. If this
00:16:58
continues for another 6 months, it's
00:17:00
basically going to result in the
00:17:04
Democrats doing a clean sweep in the
00:17:07
midterms. Here's the chart that I think,
00:17:10
you know, the Republicans really need to
00:17:12
look at at how misguided,
00:17:15
you know, uh, this all is. This is the
00:17:18
chart that should be absolutely
00:17:20
terrifying. Nobody wanted this war or
00:17:22
very few people wanted it besides the
00:17:24
neocons and the people of Iran probably
00:17:26
and the Israelis. But the chances of the
00:17:29
Democrats sweeping now is up to 45%.
00:17:32
This just happened. The Democrats are
00:17:35
going to sweep, then they're going to
00:17:36
win in 2028. and the entire agenda of uh
00:17:40
MAGA and Trump's 2.0 will be gone. And
00:17:43
then you look at just absolutely
00:17:45
ignoring the working man. Inflation
00:17:48
going up above 3% as you pointed out is
00:17:50
a likelihood, Brad. Unemployment ticking
00:17:53
up. Still very low, but it's ticked up
00:17:55
10%. Worth keeping an eye on. These
00:17:57
foreign affairs things are the least
00:18:00
important to the American people. It's
00:18:02
very, very low on the list of
00:18:04
priorities. And people are looking at
00:18:07
Trump and what they believe is the
00:18:09
enriching of his family and all these
00:18:11
business deals.
00:18:12
>> You're in the kitchen sinking it.
00:18:13
>> It's literally what I was thinking.
00:18:15
>> You're bringing everything at 1, two,
00:18:17
three, four. Number one, doing the war
00:18:20
that everybody said he should not do,
00:18:22
include and that that was why we should
00:18:23
vote for him. Number two, the Epstein
00:18:25
files. Number three, the ICE cruelty.
00:18:26
And number four, not working for the
00:18:30
American working man who doesn't own
00:18:32
equities. Those are four. 1 2 3 four.
00:18:34
It's not a kitchen sink. This is not my
00:18:36
personal feelings on this. This is my
00:18:37
assessment of the situation. If he
00:18:39
doesn't find an offer ramp quickly,
00:18:42
they're going to lose both houses in the
00:18:44
midterms. That's I think the thing Trump
00:18:47
needs to really consider and I think he
00:18:48
will consider I think he's going to find
00:18:50
an offer.
00:18:50
>> Right. That is the topic. The topic was
00:18:52
are we going to find an offramp or not
00:18:53
find an offramp in Iran? And I think Sax
00:18:56
made the argument that there's danger
00:18:59
that neocons and others are arguing that
00:19:01
we expand, put boots on the ground.
00:19:03
You're saying if he doesn't it'll be a
00:19:05
disaster. Chimath and I both say he
00:19:07
will. Right. And so the
00:19:10
>> wait will what?
00:19:11
>> He will find an off-ramp over you know
00:19:13
in the nearer term because the Trump
00:19:15
doctrine is not the neocon doctrine. As
00:19:18
much as people want to talk about Iran
00:19:21
Iran Iran. I think as I explained last
00:19:23
week I think this is about China China.
00:19:26
And you have to remember at the end of
00:19:28
this month he has a pivotal 3 days with
00:19:32
Xi Jinping in China. This is going to be
00:19:34
an absolutely historic convening
00:19:38
of the two superpowers
00:19:42
that run the world. One which is us, we
00:19:45
are the established and one which is
00:19:47
China who wants to be reasscended.
00:19:50
And I would bet dollars to donuts that
00:19:54
there is going to be an enormous
00:19:55
incentive for Xi to negotiate a grand
00:19:59
bargain in those three days
00:20:02
and do something historic for himself.
00:20:06
And I think that the president will use
00:20:08
that if he thinks that it creates
00:20:09
leverage. I
00:20:10
>> I think it's is a great insight. How
00:20:12
does the straight of Hormuz open? If
00:20:14
this war is dragging on and Israel,
00:20:16
which seems to be the driving force in
00:20:19
this, if Israel keeps it up with uh
00:20:22
Iran, how do we ever get the straight
00:20:24
open again?
00:20:25
>> I think the offramp is that the United
00:20:27
States uh you know, declares victory,
00:20:29
does what Zach says and says, "Listen,
00:20:31
we degraded and we destroyed. That's
00:20:33
what we came here to do. We did not come
00:20:35
here for some experiment in in
00:20:37
democracy. We wish the best to the
00:20:39
Iranian people to do the things they
00:20:41
need to do. And if Iran does not back
00:20:43
down, if after that declaration, Iran
00:20:45
continues to destroy cargo containers
00:20:48
moving through the narrow straits, I
00:20:50
think you're going to see Iran's
00:20:52
neighbors in Israel and others get very
00:20:56
involved as it pertains to Iran because
00:20:58
it's in their interest. Listen, the
00:20:59
United States produces 20 million
00:21:01
barrels of oil a day and we consume 20
00:21:03
million barrels a day. This is a this is
00:21:05
a modest problem for the United States.
00:21:07
This is a massive problem for China.
00:21:09
This is a massive problem for Asia. This
00:21:11
is a massive problem for all of our
00:21:13
friends in the Gulf who are trying to
00:21:14
dodge Iranian missiles right now. So,
00:21:16
there are a lot of people in the world
00:21:18
who will take up arms to deal with the
00:21:20
Iranians if the United States isn't
00:21:22
there because we can take care of
00:21:23
ourselves. Your position Brad just to
00:21:25
confirm it is we are going to leave the
00:21:28
war in the next 30 days and then if the
00:21:30
straits are not open then China, India
00:21:34
and all the Gulf uh countries that are
00:21:36
impacted by it they will protect it.
00:21:39
They will fight Iran.
00:21:40
>> I think they'll put a lot of pressure on
00:21:41
Iran not to continue firing missiles at
00:21:44
their ships. Right? At the end of the
00:21:45
day this is not just an American
00:21:47
problem. Right? And let's be clear we're
00:21:50
always involved in this part of the
00:21:52
world. The only question is, are we
00:21:54
going to have an active armada that's
00:21:56
engaged in active military activities
00:21:58
against Iran? And what I'm suggesting
00:22:00
again, and listen, anytime you try to
00:22:03
clean up a mess like this, there is
00:22:04
risk. This is not a riskfree, you know,
00:22:08
initiative by the United States, nor was
00:22:10
Venezuela. But let me steal man the
00:22:13
alternative. doing nothing and allowing
00:22:16
Iran to procure, you know, the
00:22:18
ingredients for a nuclear missile when
00:22:19
they are set on the destruction of the
00:22:21
United States and US interests. Doing
00:22:23
nothing in Venezuela while the Monroe
00:22:25
Doctrine is totally wrecked and we let
00:22:27
our adversaries take up, you know,
00:22:29
positions in in South America. Those
00:22:31
also have risks, right? Those carry a
00:22:34
lot of risks. And so we're we're
00:22:36
weighing these two risks. Again, for me,
00:22:38
I don't like the fact that we're engaged
00:22:40
in military activities here. But I will
00:22:42
tell you, I am very much on the side
00:22:44
that if we're going to go protect
00:22:46
American national security interest, you
00:22:48
go in, you do the the degrading of their
00:22:51
capability and you get out. And I think,
00:22:54
you know, that's what I hear out of the
00:22:55
president.
00:22:56
>> Jim, you had a followup.
00:22:58
>> All roads lead to China. I think that
00:23:00
there you're going to see Xi offer up a
00:23:03
grand bargain. And I think it's up to
00:23:04
the president to decide whether he wants
00:23:06
to take it and see what he wants to add
00:23:07
to it to get something done. But I just
00:23:09
don't see them meeting and coming out
00:23:11
with nothing. I think I see them going
00:23:13
in and coming out with something that's
00:23:15
historic.
00:23:17
And I think that all of this Venezuela
00:23:19
and Iran together is all about China.
00:23:22
>> Let me just say one thing as to that,
00:23:23
Chimath. I think because I think the
00:23:25
point is absolutely spoton, right?
00:23:28
Probably the single greatest takeaway
00:23:30
for us from an investment perspective at
00:23:32
the start of this war was that the
00:23:34
Chinese, right, didn't take up arms on
00:23:37
behalf of Iran
00:23:39
aren't defending Iran and they didn't
00:23:43
cancel the summit with the president,
00:23:45
right? The very fact that because they
00:23:47
need him, they need the oil. 20% of
00:23:51
their entire domestic consumption is oil
00:23:54
from Venezuela and Iran. 20%. But it's
00:23:57
not 20%. Because it's literally 100% of
00:24:00
anything that's feed stock, anything
00:24:02
that's transport, cars, buses, planes.
00:24:06
They are in an enormous world of hurt.
00:24:08
Now, they have a strategic petroleum
00:24:10
reserve as well, and it's quite robust,
00:24:13
but it's not robust enough to sustain
00:24:15
five or 6 months of this. It's not that
00:24:18
robust. So, at the end of the day, who
00:24:21
is going to be hurting the most? It is
00:24:23
China. And so if you play this game
00:24:25
theory out, the reason he kept it is
00:24:27
because now he needs a summit even more.
00:24:29
Could you imagine if the president
00:24:31
canled, that would be a disaster for the
00:24:33
Chinese.
00:24:35
So the fact that it's still on the
00:24:36
books, if I was Xi, I'd be like, how do
00:24:39
I negotiate and help find the off-ramp?
00:24:42
How do I end up fixing this faster? All
00:24:46
right, remember you have 25%
00:24:48
unemployment of young men inside of
00:24:50
China. 25% today. What do you think it
00:24:53
goes to in five months with no oil?
00:24:57
>> It's in the
00:24:58
>> That's the unemployment rate you should
00:24:59
be focused on, Jason.
00:25:02
>> Oh, the China issue is a separate one
00:25:04
just to
00:25:05
>> It's not separate.
00:25:06
>> No, no, that was separate from my point
00:25:08
is my point. I I was bringing up a
00:25:09
different point.
00:25:10
>> Yeah, your kitchen sink didn't include
00:25:11
the Chinese. I get that. I'm just adding
00:25:13
to your kitchen sink.
00:25:14
>> I didn't have a kitchen sink. I have
00:25:15
four very salient points. All right,
00:25:17
Sax, I'll give you the final word here.
00:25:20
Well, look, that was a bit of a a
00:25:21
broadside, Jay Cal, where you kind of
00:25:23
did kitchen sink it. But look, here
00:25:25
here's the part I'll agree with you
00:25:27
about, which is it doesn't take a
00:25:29
political genius to understand that long
00:25:31
wars are unpopular.
00:25:34
It will hurt the Republicans in the
00:25:36
midterms or 28 if this does turn into a
00:25:38
long war. Fortunately, I think the
00:25:40
president understands that his political
00:25:42
instincts are impeccable and he's always
00:25:44
favored
00:25:46
short, decisive, swift actions, military
00:25:49
actions, whether it was Midnight Hammer,
00:25:50
whether it was the Maduro raid. I think
00:25:53
that is his inclination and preference.
00:25:55
And I think we are pretty much at or
00:25:58
close to a point where the president's
00:26:00
going to have to decide on next steps. I
00:26:02
think he's indicated that we have
00:26:05
completed our objectives and I think
00:26:07
it's just important that we don't let
00:26:09
this neocon wing of the party try to
00:26:12
expand the objectives or aims of the war
00:26:16
because frankly they've always been
00:26:17
wrong about everything. I mean these are
00:26:19
people who never wanted to get out of
00:26:21
Iraq and Afghanistan would still be
00:26:23
there after 20 years if they had their
00:26:25
choice. So, I think it's just important
00:26:28
to not listen to those people. And look,
00:26:30
it's not just one op-ed in the Wall
00:26:32
Street Journal. The Wall Street Journal
00:26:34
is kind of the tip of the spear
00:26:35
representing that whole neocon
00:26:37
establishment. And I think it's just
00:26:39
important that this is the time to
00:26:41
frankly ignore those voices and let the
00:26:42
president do what I think his political
00:26:44
instincts are telling him to do, which
00:26:46
is to wrap this thing up.
00:26:48
>> I'm in strong agreement, and it is my
00:26:50
hope, too, that he wraps it up quickly
00:26:52
and that we don't have any more loss of
00:26:55
life. All right. We'll keep discussing
00:26:57
uh this ongoing breaking news story in
00:27:00
the coming weeks. But back to our zone
00:27:02
of excellence, AI and tech, OpenAI and
00:27:06
Anthropic are scaling revenue and costs
00:27:08
faster than we've ever seen in the
00:27:10
history of well business the world.
00:27:13
Revenue at these two companies growing
00:27:15
gosh like unprecedented levels. Here are
00:27:18
the reports and I believe your investors
00:27:20
in both these companies. Brad Anthropic
00:27:22
hit a $14 billion run rate last month,
00:27:26
February. That means they have grown
00:27:28
revenue from 1 billion uh to 14 billion
00:27:32
in 14 months. They have 12x
00:27:34
year-over-year.
00:27:36
They're valued at a meager 380 billion
00:27:39
uh last month. This feels like a bargain
00:27:41
given the growth. Openai ended 2025 at
00:27:44
20 billion annualized run rate and
00:27:47
they've grown revenue from 2 billion to
00:27:49
20 billion in 24 months. They're valued
00:27:51
at 840 billion last month. Uh, and man,
00:27:55
it looks like Sam Alman has Daario in
00:27:58
the rearview mirror. He could get lapped
00:28:02
any moment. Lots of debate.
00:28:03
>> Where did you find this? What the hell
00:28:05
is this?
00:28:05
>> That one? I made that. This is Daario
00:28:07
closing in.
00:28:09
>> What is that? A velociraptor? What is
00:28:11
that?
00:28:12
>> It's the famous scene from Jurassic
00:28:14
Park.
00:28:14
>> Oh, wow. But I mean, I don't think
00:28:16
anybody expected Daario to be coming
00:28:17
around the bend this fast, but he's
00:28:19
right behind apparently and they're
00:28:21
winning obviously the businessto
00:28:23
business side of the business. The J
00:28:25
curve on these companies is insane. Uh
00:28:28
250, 500 billion. Who knows what's get
00:28:30
gets invested before these companies
00:28:33
reach profitability, Brad, but you're
00:28:35
invested in these two companies.
00:28:37
>> Yeah.
00:28:37
>> Unless you sold when uh Sam Wman told
00:28:40
you he would buy his shares back on the
00:28:42
famous BG2 episode. I don't think you
00:28:44
sold it.
00:28:44
>> I bought a lot more since then, Jason. I
00:28:46
bought a lot more.
00:28:47
>> Fantastic. That's important information
00:28:48
for us to have. Quick question for you.
00:28:51
Number one, what's a better buy here?
00:28:53
>> Yeah.
00:28:53
>> Uh, Anthropic at 380, OpenAI at 840. Uh,
00:28:57
and then I think people want to know if
00:28:58
these companies are going to go public.
00:29:00
What you, if you think they should go
00:29:02
public, what is the chance of that
00:29:03
happening? Take those questions however
00:29:05
you like.
00:29:06
>> Well, I mean, listen, you know, Love
00:29:08
Your Children equally. They're both
00:29:10
incredible companies. Anthropic
00:29:11
unquestionably has a lot of financial
00:29:13
momentum, you know, and OpenAI is seeing
00:29:16
a lot of momentum themselves, right? But
00:29:17
the single most important question this
00:29:19
year, right, was would AI revenue show
00:29:22
up? And just 60 days ago, 90 days ago,
00:29:26
there was tremendous skepticism. No way
00:29:28
all of these infrastructure investments
00:29:30
were going to pay off. There's no
00:29:32
incremental revenue coming out of AI,
00:29:34
including many of our friends. But in
00:29:37
February, we had in January and
00:29:39
February, we really had kind of a
00:29:41
nuclear moment, right? The splitting of
00:29:42
the atom moment. I mean, we had a $6
00:29:45
billion month out of Enthropic in
00:29:48
February, right? What widely reported.
00:29:50
Okay, let that set in for a second,
00:29:53
right? $6 billion in a month, and it was
00:29:56
only a 28 day month. Okay, that's more
00:29:59
revenue than the annual revenue of data
00:30:01
bricks and snowflake that are two of the
00:30:03
greatest software companies of all time
00:30:05
after 12 years, right? They could do in
00:30:08
the first four or five months of this
00:30:09
year the total revenue of SpaceX this
00:30:12
year.
00:30:12
>> What is driving that? Just explain to
00:30:14
the audience what's driving it. Is it
00:30:15
token use? Is it clawed subscriptions?
00:30:18
>> We crossed a threshold with Opus 4.6,
00:30:22
right? And we saw it again with chat GBT
00:30:25
5.4 before where the models and the
00:30:27
agents on top of them whether it's cloud
00:30:29
codeex chatbt they're no longer
00:30:32
competing with IT budgets they're now
00:30:34
augmenting labor they're competing with
00:30:36
labor budgets you could not possibly
00:30:38
have a $6 billion month it is impossible
00:30:41
to do that by displacing IT budgets
00:30:44
millions of other companies across
00:30:45
America say oh my god let's spin up
00:30:48
these agents and have them do things for
00:30:49
us and we're willing to pay for it
00:30:51
because the product of that effort is
00:30:53
worth the money to And the revenue and
00:30:56
the usage momentum, I will tell you in
00:30:58
the month of March continues and it only
00:31:00
accelerates from here. As Kevin Wild has
00:31:03
said, the models in the agents are the
00:31:05
dumbest today they will ever be. Right?
00:31:08
We're in the early innings of compute
00:31:10
and algorithmic uh capability. And so,
00:31:14
you know, like that to me is the
00:31:15
observation at this moment. Should they
00:31:17
go public? I've said yes, they should go
00:31:19
public for several reasons. There's tons
00:31:20
of institutional demand. They need cheap
00:31:23
access to money to continue to build out
00:31:25
uh the compute they need to support.
00:31:27
They they there is more compute
00:31:28
constraint in these businesses this very
00:31:31
day than they've had any time in the
00:31:33
last 3 years. So they need access to the
00:31:34
capital. And then finally, I think you
00:31:37
have to have the retail investor in the
00:31:38
game. These are two of the most
00:31:40
important companies in the history of
00:31:41
capitalism in the history of America.
00:31:43
It's destabilizing not to have them
00:31:45
public. You know, Jensen said last week
00:31:48
that he expected the 40 billion he
00:31:50
recently invested in these two companies
00:31:52
would be his last money in because they
00:31:54
were both going to go public. He thought
00:31:56
they he said they would both go public
00:31:58
this year. I think that, you know,
00:31:59
they're they're preparing and heading
00:32:01
down that path, you know, and and
00:32:03
listen, I want to get some of these
00:32:05
shares in the accounts of all these kids
00:32:06
that were opening up because they're
00:32:08
really really important companies to the
00:32:10
future of the American economy. Shim,
00:32:12
you had some insight into the quality,
00:32:16
durability of this revenue. There's not
00:32:18
a single good example that we can find
00:32:22
of sustained positive
00:32:25
margin expansion and impact of AI inside
00:32:29
of a true corporate enterprise that is
00:32:32
not right now a small test. There's not.
00:32:35
>> So where does six billion come from?
00:32:38
because everybody has to show up to
00:32:40
their board and have an AI checkbox and
00:32:43
everybody is thousands and thousands of
00:32:46
companies. And when you have tens of
00:32:48
thousands of companies as customers
00:32:49
paying $200 plus a month, it's not that
00:32:52
hard to show up with that kind of
00:32:54
revenue. The real question is the
00:32:56
following. If you take you use the data
00:32:59
bricks and snowflake example, if you
00:33:01
look at the companies that use that
00:33:04
software,
00:33:06
those companies generate enormous
00:33:10
revenues and enormous margins and these
00:33:13
products are in critical production
00:33:16
workflows that underly those revenues
00:33:19
and profits. That is just not true with
00:33:21
AI today. We have all kinds of claims,
00:33:24
but we are still experimenting.
00:33:27
Why are we experimenting? Because we
00:33:29
know it's important, but we don't yet
00:33:32
really know what to do. You can't just
00:33:34
slot this in to a critical workflow in
00:33:36
healthcare and all of a sudden show up
00:33:39
where if you make a misdiagnosis or if
00:33:41
you make a a misccharacterization of a
00:33:44
procedure, you can get fined and go to
00:33:46
jail. The companies that are in
00:33:48
healthcare don't do that. If you're in
00:33:50
financial services and you make a
00:33:52
mistake about somebody's portfolio or
00:33:54
you make a misallocation and you point
00:33:55
to a model, you will get sued and you
00:33:58
will be in trouble. None of these things
00:34:01
have transitioned from it's interesting,
00:34:03
it's experimental to it's the core
00:34:06
critical operational workflow.
00:34:08
>> Just interesting.
00:34:09
>> There will be a transition in revenue
00:34:11
quality when that happens. A great
00:34:14
example of this is Amazon. Why does
00:34:16
Amazon issue an edict that says you
00:34:19
cannot use this stuff inside of AWS
00:34:22
unless a human now reviews and approves
00:34:24
it? Because what happened? They had
00:34:26
three or four sev one faults from a
00:34:29
bunch of code that was written by agents
00:34:31
that brought down AWS. Now look, I've
00:34:33
told you I love AWS for one reason,
00:34:36
because it's hyper reliable. I hate AWS
00:34:39
for the same reason that hyper
00:34:40
reliability comes at enormous cost. I
00:34:42
pay it, but I pay it to never have a SE
00:34:45
one. The reason they have 12 9s of
00:34:48
accuracy is because it's humans and
00:34:50
deterministic code that never fails. It
00:34:52
doesn't mean that two companies can't
00:34:54
get to 20, 30, 40 billion of revenue.
00:34:56
What it means is we have to be honest.
00:34:58
This is an industry that's early. We are
00:35:01
all figuring it out. There's a lot of
00:35:04
test budgets that are going at it. It
00:35:06
will slowly and methodically emerge into
00:35:09
production. But let's not oversell what
00:35:12
this moment is.
00:35:13
>> Okay, Brad, I want to give you a well-
00:35:14
constructed question here to respond and
00:35:16
then Sax, we'll go to you if you have
00:35:18
some input. Of the 20 billion, how much
00:35:20
of it do you think is experimental? What
00:35:24
percentage is experimental versus
00:35:26
production? Bas strip out the cons the
00:35:29
consumer spending for because like you
00:35:31
know that's half of it, right?
00:35:33
>> Okay, so let's I'll I'll put aside the
00:35:35
consu S. Great idea. We'll put the
00:35:36
consumer subscriptions aside. they're
00:35:38
obviously getting value or they wouldn't
00:35:39
be subscribing, but it's like Netflix.
00:35:41
>> And by the way, for that where it can be
00:35:44
extremely faulty and there's no SLA that
00:35:46
you're giving,
00:35:47
>> it's a phenomenal. These are phenomenal
00:35:49
products.
00:35:50
>> Yes. For 20 bucks a month, well worth
00:35:51
it. And consumers have decided it's
00:35:53
worth it. So, I think we're in agreement
00:35:54
there.
00:35:55
>> And also for the individual engineer, of
00:35:58
which I suspect there's a few million,
00:36:00
who get to pay $200 a month and have
00:36:02
their company subsidize it. The company
00:36:04
knows that these costs are being
00:36:06
incurred, but there is no tick and tying
00:36:09
at the end of it where then you review
00:36:10
the code in a different way because you
00:36:12
you you're worried that there's
00:36:13
hallucinations as Amazon just
00:36:14
demonstrated.
00:36:15
>> Yeah, there's a story in the there's a
00:36:17
story in the FT about Amazon having some
00:36:19
blast radius from some computerenerated
00:36:22
AI and they're putting controls in
00:36:24
place. We'll put that in the show notes.
00:36:25
Brad, let's get to the specific question
00:36:27
I asked. of the tens of billions of
00:36:29
dollars in revenue between the two
00:36:30
companies that's not consumer, what
00:36:33
percentage do you think is production
00:36:35
quality versus experimental to Chimat's
00:36:38
point?
00:36:38
>> Yeah, I mean listen and I've been, you
00:36:40
know, I coined the phrase experimental
00:36:42
run rate revenue, right, versus annual
00:36:45
recurring revenue, right? Like I think
00:36:46
it's I think Chimas's point is a really
00:36:48
important one. As an investor, I have to
00:36:50
discern what's repeating, right? What's
00:36:52
recurring and what's not. What I would
00:36:55
suggest is of course there's a lot of
00:36:56
experimentation because these things
00:36:58
haven't been around that long. But I
00:37:00
suspect right that Palunteer, the US
00:37:04
government, the US military, Nvidia, and
00:37:08
a lot of other major enterprises would
00:37:11
argue they've gone full production. In
00:37:14
fact, it's existential to the wartime
00:37:16
effort going on in Iran right now. That
00:37:18
doesn't sound to me like experimental as
00:37:20
much as it sounds like production
00:37:22
capability. And I will tell you what
00:37:24
will prove this one way or the other is
00:37:27
in the month of March, do revenues
00:37:29
continue and go up? Right? Do they
00:37:32
experimentation go on forever? That's
00:37:34
not true. That's not true. The
00:37:36
experimentation goes on forever. That
00:37:38
sounds like recurring.
00:37:39
>> We've scratched the surface of of the
00:37:41
number of companies that even know how
00:37:42
to adopt AI. So these numbers will go to
00:37:44
the stratosphere. I'm not debating that.
00:37:46
>> Okay.
00:37:47
>> Look, I'm on the same side of the bed as
00:37:49
you are. I want these numbers to keep
00:37:50
going to the moon. I'm just being much
00:37:52
more circumspect and honest with myself
00:37:54
to say I see it on the ground. I sit on
00:37:57
top of these models. I am paying these
00:37:59
models millions of dollars a year. I am.
00:38:01
>> Yeah.
00:38:02
>> And what I'm telling you is my revenues
00:38:03
don't go up faster than their revenues.
00:38:05
I'm consuming more tokens every single
00:38:08
day. Do I get more economic output? I am
00:38:11
not. And I would say that my team is at
00:38:13
the leading edge. And so I suspect a
00:38:15
Fortune 1000 company is steps behind my
00:38:18
team. And if I am spending triple every
00:38:21
3 months and not seeing my revenues
00:38:24
tripling, I suspect these other
00:38:26
companies are in a similar situation.
00:38:28
>> I'd finalize it. But ask it on Friday
00:38:30
when you're with Michael Dell because
00:38:32
I've had this conversation recently with
00:38:33
Michael Dell and Michael said a year ago
00:38:35
companies weren't seeing ROI. Today
00:38:37
they're seeing very big ROI in their AI
00:38:39
investments. But I think that's the
00:38:40
question on the table. But
00:38:42
>> what companies? Which companies? Of
00:38:44
course the second
00:38:46
of course is seeing an ROI. Sachs were
00:38:49
agreeing. This is experimental in large
00:38:52
part and this is a new tool. So by
00:38:56
definition you have to experiment before
00:38:57
you put it in production. What's your
00:38:59
take on this grand debate? How much of
00:39:01
this revenue is experimental versus
00:39:04
real?
00:39:04
>> Well look when you're talking about
00:39:06
enterprise revenue, what you're really
00:39:08
talking about is coding assistance.
00:39:10
That's been the breakout use case. It's
00:39:12
really the first big breakout use case
00:39:14
on the enterprise side. The consumer
00:39:16
side's been more of like you know
00:39:18
research and writing that kind of stuff
00:39:20
the chat bots but enterprise has all
00:39:23
been about coding assistance. My sense
00:39:25
is that the demand for code is very
00:39:27
scalable. Software engineers has always
00:39:29
been an area of the economy where
00:39:31
companies have never been able to hire
00:39:32
enough. Even in Silicon Valley which is
00:39:33
the most attractive place for software
00:39:36
engineers to work. We've never been able
00:39:37
to recruit and attract enough of them.
00:39:41
The rate limiting factor on the progress
00:39:43
of every startup I've ever invested in
00:39:45
is not having enough engineers to code
00:39:47
up the product roadmap. And then you
00:39:50
look at the rest of the economy, the
00:39:51
Fortune 500 and so forth and so on. They
00:39:53
have hardly been able to recruit
00:39:55
software engineers at all because
00:39:57
they've all gone to Silicon Valley. So I
00:39:59
think you're dealing with a part of the
00:40:00
economy where there's always been a
00:40:02
massive supply shortage. And I don't
00:40:04
know what the natural limit on that is,
00:40:06
but my sense is that there's a
00:40:08
tremendous latent demand for the ability
00:40:11
to generate code in large quantities,
00:40:16
create new products. You know, as the
00:40:18
cost of code goes down, as the coding
00:40:20
assistants get better, you can code up
00:40:22
new types of products. And then, of
00:40:24
course, it's going to lead to agents,
00:40:25
which is another way of basically using
00:40:28
the code that gets generated. So, my
00:40:30
sense is that this could be very
00:40:32
scalable. I don't know where it it taps
00:40:34
out exactly. Where I think Chimatha is
00:40:37
right is that I think there is a change
00:40:39
management aspect to this in Fortune 500
00:40:43
companies for example and they haven't
00:40:44
really wrapped their heads around how
00:40:46
exactly they're going to use this.
00:40:48
There's a McKenzie study that showed
00:40:50
that a lot of these pilot projects in
00:40:52
Fortune 500 companies were experimental.
00:40:54
A lot of them were proving not to be
00:40:55
successful. So I do think like as you go
00:40:58
beyond coding into you know company
00:41:02
transformation things like that it
00:41:03
becomes a little bit more speculative.
00:41:05
That's not to say it won't happen. I
00:41:07
think it will happen. I'm actually I'm
00:41:09
bullish. But I do think that
00:41:12
>> we're still waiting to see what the
00:41:13
breakout use cases beyond coding will
00:41:15
be. Probably agents will be the next big
00:41:17
one. But I think Brad's right that
00:41:18
that's big enough to see, you know, this
00:41:21
scale for a while because, you know, the
00:41:23
thing about the thing about code is
00:41:25
you're paying for code on a metered
00:41:27
basis right now. You're paying per
00:41:28
token, which is kind of an amazing deal
00:41:31
for companies, right? Because before
00:41:33
they had to go through this recruiting
00:41:34
process to find engineers, source them,
00:41:38
vet them, you know, keep them happy,
00:41:40
give them all the perks, the kind bars,
00:41:43
all, you know, uh, and so to be able to
00:41:45
buy code on a metered basis as the cost
00:41:48
per token keeps going down.
00:41:50
>> It's kind of metered before. It was just
00:41:52
metered on people. Now, now to your
00:41:54
point, it's it's metered. It's metered
00:41:56
in a different way, but it's still
00:41:57
metered.
00:41:58
>> Yeah. And let me just, you know, you use
00:42:00
this term labor displacement, Brad.
00:42:02
That's like the one part where I might
00:42:04
disagree with you is because there was
00:42:06
such a shortage.
00:42:08
>> Yeah.
00:42:08
>> Of software engineers. I think when
00:42:10
people hear the word labor or term labor
00:42:12
displacement, they might
00:42:14
>> start to think that 6 billion of
00:42:16
incremental revenue means 6 billion of
00:42:19
layoffs. No.
00:42:20
>> And I don't think it does. And and the
00:42:21
way to to thread that needle is the fact
00:42:23
that we were artificially constrained in
00:42:26
the number of software engineers and how
00:42:28
they could be used and how rapidly they
00:42:31
could be acquired and all that kind of
00:42:32
stuff. So to be able to now turn that on
00:42:35
like electricity and that's kind of what
00:42:36
we're talking about.
00:42:37
>> Exactly.
00:42:38
>> Is just such a huge
00:42:40
>> game changer and unlock for the whole
00:42:42
economy.
00:42:43
>> Yes.
00:42:43
>> And that's what I think is really
00:42:44
exciting about it.
00:42:45
>> It's augmenting it's augmenting human
00:42:47
labor, right? We're not at a place yet
00:42:49
where it's displacing. And this is the
00:42:51
definition of productivity gains.
00:42:53
>> I'm going to make just uh two quick
00:42:54
points here. The place to look for this
00:42:57
actually moving from experimental into
00:43:00
production is not not at big companies.
00:43:02
Big companies are actively resisting
00:43:04
this. Management in big companies will
00:43:06
resist it because it means lowering
00:43:07
headcount and it means the person who
00:43:09
implements it might actually implement
00:43:10
themselves out of a job. So that is the
00:43:13
natural resistance you'll see in big
00:43:14
companies. That's not where to look for
00:43:15
adoption of new tech.
00:43:16
>> That's not why it's happening. That's
00:43:18
not why
00:43:18
>> let me finish and then you can you can
00:43:20
counter it.
00:43:21
>> Startups are the place to look at this
00:43:24
and that's where I am on the ground. And
00:43:26
what I'm seeing there is that startups
00:43:29
are using this in production for their
00:43:31
legal work for their marketing for SDRs
00:43:35
uh for their accounting reviewing legal
00:43:38
documents. This is all uh work that
00:43:40
would normally they would hire
00:43:42
consultants for outsource uh or make
00:43:45
hires for. And what I'm seeing on the
00:43:47
ground is it's production ready in
00:43:49
startups who are using it in those
00:43:51
categories. HR as well, accounting,
00:43:54
marketing, all of that, all that
00:43:56
blocking and tackling, all those chores
00:43:57
are being done currently with these LLMs
00:44:00
and they're doing it in production and
00:44:01
they're doing it at scale. Just a quick
00:44:03
second point here. Here's the J curve.
00:44:05
And this is the question I think we'll
00:44:07
get to in our next segment. When does
00:44:08
this become a profitable business? If uh
00:44:11
and you asked this of Sam in that famous
00:44:13
clip on the BG2 podcast, RIP BG2, here
00:44:18
you go. The LLM industry, Jacob, I just
00:44:20
asked Claude to make this for me. Um, if
00:44:22
you have 500 billion, I think you would
00:44:24
agree it's probably going to be around
00:44:25
that number, Brad, invested in all of
00:44:27
this. And then
00:44:28
>> more a lot more.
00:44:30
>> Okay. So, five billion is an
00:44:32
underestimate here. And then when do we
00:44:34
actually see these large language model
00:44:35
companies hit profitability in a in a
00:44:38
calendar year? It took Tesla, Uber,
00:44:42
Amazon, you know, decade plus in each of
00:44:44
those cases
00:44:45
>> to win back their investment. 10
00:44:47
billion.
00:44:47
>> This is a really good chart. Here's the
00:44:49
precise math on this. So, I am building
00:44:51
a 1 gawatt data center in Arizona.
00:44:53
>> Okay.
00:44:54
>> When I green lit that project, I thought
00:44:56
it was going to be a four or five
00:44:57
billion investment. I was like, okay,
00:44:59
whatever. Then it went to 10. Then it
00:45:01
went to 15. Then it went to 20. And now
00:45:03
it's upwards of $50 billion for the
00:45:06
powered shell, for all the land, for all
00:45:08
the permits, then for all the
00:45:09
infrastructure, all the people, all of
00:45:12
it. Okay. Sarah Frier said, I think it
00:45:15
was about a year ago, maybe less than a
00:45:17
year ago, that for them, every gigawatt
00:45:19
is about 10 billion of annual revenue.
00:45:21
So if you think about that J curve,
00:45:23
Jason, really the way to think about it
00:45:24
is look, energy equals intelligence. For
00:45:27
every gigawatt that they're trying to
00:45:28
spend, they have a fiveyear payback is
00:45:30
roughly what it means just to get to
00:45:32
break even. And then year six, seven,
00:45:33
and eight will be where the profit is.
00:45:35
Now, how do you shrink the Jaker? You
00:45:37
have better silicon. We're going to see
00:45:39
something from Jensen in a in a week or
00:45:41
two that uses a bunch of the stuff that
00:45:44
we partnered with him at Grocon on.
00:45:46
There'll be other people. There'll be
00:45:48
open source. So, all those things can
00:45:50
shrink the depth and the surface area of
00:45:52
that JC curve so that you can get out of
00:45:54
it faster. But right now that that thing
00:45:56
is roughly accurate which is it's about
00:45:58
50 billion per gigawatt and it's about a
00:46:02
five to six year payback just to get
00:46:04
into the money and then it's about 10
00:46:05
billion a year and the technology
00:46:08
industry has to do something to make
00:46:10
this better. Could I though
00:46:13
take a step back and give you just a
00:46:14
different framing of all of this
00:46:17
>> please?
00:46:18
I think the big thing that we're
00:46:19
debating is actually something we've
00:46:22
seen in every other technology trend
00:46:26
when it started to get some really
00:46:28
meaningful traction. So in the first
00:46:30
generation of the internet when you
00:46:31
started to see e-commerce and all these
00:46:33
other business models then in the second
00:46:35
big wave of the internet around the move
00:46:37
to mobile and the move to social and
00:46:39
then now we're seeing this big wave
00:46:40
around AI and I think what happens is in
00:46:45
step one
00:46:47
entrepreneurs
00:46:48
are AB testing
00:46:51
what it takes to raise money. Okay,
00:46:54
that's step one.
00:46:56
And I think what has happened is that at
00:46:59
least some parts of the AI ecosystem
00:47:02
have decided that this crazy
00:47:06
scary doomerism is the best way to raise
00:47:09
money where every now and then they come
00:47:12
out and they say all the jobs will be
00:47:14
destroyed anthropic you know Daario says
00:47:16
that this thing is sentient and
00:47:18
investors are like okay here's 10
00:47:21
billion here's 50 billion here's 100
00:47:22
billion but then the second step happens
00:47:25
they get the money, they start to do the
00:47:27
training, they start selling, and then
00:47:29
the investors are like, "Hey, where's
00:47:31
the revenue?" And so then they start
00:47:32
selling everywhere.
00:47:34
And then if you see in the Department of
00:47:37
War example, all of a sudden you
00:47:39
flip-flop, you become sort of an
00:47:40
unserious dilotant like partner to the
00:47:43
American government. They're like,
00:47:44
"We're going to boot you out. That's
00:47:45
billions of revenue gone." And what
00:47:47
happens? Those same investors that gave
00:47:48
billions of dollars are like, "Hey, hold
00:47:49
on a second. That's absolutely not
00:47:51
allowed. You need to conform and get
00:47:53
back on track." And so what does Dario
00:47:54
do? He flip-flops and he's like, "Oh,
00:47:56
I'm really sorry. I didn't mean to.
00:47:58
Let's sort of make good." All of that to
00:48:00
me is an industry that's still in its
00:48:02
very early phases and still figuring out
00:48:05
what its place in society is. So, what
00:48:08
is the problem? The problem is the
00:48:10
following two clips, and I'll just have
00:48:12
Nick play these, and I'd love your guys'
00:48:14
reaction. The one thing though that I
00:48:17
think even now is underestimated by all
00:48:20
actors in industry and including in
00:48:23
Silicon Valley is how disruptive these
00:48:26
technologies are. If you are going to
00:48:28
disrupt the economic and therefore
00:48:30
political power significantly of one
00:48:33
party space, highly educated, often
00:48:36
female voters who vote mostly Democrat
00:48:39
and military and workingclass people who
00:48:42
do not feel supported. and you feel like
00:48:45
that's you believe that that's going to
00:48:46
work out politically, you're in an
00:48:48
insane asylum like that you cannot have
00:48:51
a this technology disrupts humanity's
00:48:54
trained largely democratic voters uh and
00:48:57
and makes the economic power less and
00:49:00
increases the power economic power of
00:49:03
vocationally trained workingclass often
00:49:05
male uh voters. and and and so these
00:49:09
disruptions are going to disrupt every
00:49:11
aspect of our society. And it to make
00:49:15
this work, we have to come to an
00:49:17
agreement of what it is we're going to
00:49:20
do with the technology. How are we going
00:49:22
to explain to people who are likely
00:49:24
going to have less good and less
00:49:26
interesting jobs from their perspective?
00:49:29
And how is it that we are going and by
00:49:32
the way on the military thing, these
00:49:34
technologies are dangerous society. the
00:49:37
only justification you could possibly
00:49:39
have would be that if we don't do it,
00:49:41
our adversaries and uh will do it and we
00:49:44
will be subject to their rule of law. So
00:49:47
you if you decouple this from the
00:49:48
support of the military, you're going to
00:49:50
have an enormous problem explaining to
00:49:52
the American people why is it that we're
00:49:55
absorbing the risk of disrupting the
00:49:58
very fabric of our society, including
00:50:00
the most powerful
00:50:02
parts of our society. uh if it's not
00:50:05
because it's about maintaining our
00:50:07
ability to be American in the near term
00:50:09
and and and long term.
00:50:10
>> Now watch Sam's reaction. Fundamentally,
00:50:13
our business and I think the business of
00:50:15
every other model provider is going to
00:50:17
look like selling tokens. But we see a
00:50:22
future where intelligence is a utility
00:50:26
like electricity or water and people buy
00:50:30
it from us um on a meter and use it for
00:50:34
whatever they want to use it for. So if
00:50:36
you take those three messaging veins on
00:50:38
a spectrum, one is we have a sentient
00:50:41
super god. We're the only ones that can
00:50:43
protect you from it, but you know your
00:50:45
days are numbered. That's Daario. Alex,
00:50:47
which is, hey, hold on a second. You
00:50:48
can't have it both ways. You can't both
00:50:51
say it on the one hand and then try to
00:50:52
run the fabric of society and flip it.
00:50:54
You need to be much more circumspect.
00:50:56
And then Sam's, which is we want to sell
00:50:58
tokens as a service.
00:50:59
I think the point is that this industry
00:51:01
right now that revenue traction if
00:51:04
anything else has distracted people from
00:51:07
actually getting on the same page and
00:51:08
being much more methodical and much more
00:51:11
reliable and trustworthy in explaining
00:51:14
all of this and managing the expansion
00:51:16
of this. And so what I would say is all
00:51:19
of this fundraising gobbledegook has
00:51:21
actually created this breathlessness
00:51:23
that is not useful and isn't helping.
00:51:27
And I would say there needs to be a lot
00:51:28
more seriousness by these folks to
00:51:30
actually run this business thoughtfully.
00:51:32
You can't be a dilotant. You can't
00:51:34
flip-flop. You can't pressure test AB
00:51:37
test this kind of messaging in public.
00:51:40
But I understand why you're doing it
00:51:41
because the stakes are so high. You're
00:51:43
playing this enormous poker game. But I
00:51:45
think we need to do a better job of
00:51:47
explaining all this to people because
00:51:48
right now my end of this is look at this
00:51:50
chart. This is now the result of those
00:51:53
three messages. Here is where AI is. It
00:51:55
is slightly above the Democratic party
00:51:58
and an autocratic state. That's where AI
00:52:01
is. ICE is more popular than AI.
00:52:05
So
00:52:06
>> not very popular.
00:52:08
>> So to me, this is really at the crux of
00:52:11
this where we are not really being
00:52:13
honest. It would be much better if we
00:52:16
said soberly, there's a lot of
00:52:17
experimenting. This revenue is great,
00:52:19
but we don't really know what's real.
00:52:21
We're going to try to figure it out.
00:52:22
We're going to work methodically.
00:52:23
There's a lot of regulated industries.
00:52:25
We're going to work within those. We're
00:52:26
not going to flout the law and the
00:52:28
rules. Lensure will still mean
00:52:30
something. That's a way better,
00:52:32
thoughtful, mature message.
00:52:36
And what do you think?
00:52:37
>> And rant. Great rant. Uh Brad, does the
00:52:41
industry have a PR problem? Obviously,
00:52:43
these recent surveys and especially
00:52:46
comparing them to China where people see
00:52:49
AI as abundance and like this incredible
00:52:51
new technology they want to embrace.
00:52:52
Here, people are scared. People are
00:52:54
scared they're going to lose their job.
00:52:56
People are scared about wealth u
00:52:58
disparity. The rich get richer, the poor
00:53:00
get poorer. There's a lot of fear here
00:53:02
in the United States. What can our
00:53:05
industry do to turn this around in terms
00:53:08
of communication from the big companies?
00:53:11
They don't seem to be communicating in
00:53:13
any coordinated fashion and and they
00:53:16
obviously are scaring the [ __ ] out of
00:53:17
the public.
00:53:19
>> Yeah. No, listen, I think it's uh I
00:53:20
think it's a fair rant and a fair point.
00:53:22
At the start of the industrial
00:53:24
revolution at the start, you know, in
00:53:26
the in the late 1800s, we had similar
00:53:29
social responses to innovations that
00:53:31
were occurring. We in fact had some
00:53:33
violent clashes. We had demonstrations
00:53:36
in the street. We had the entire robber
00:53:38
baron movement you know so class warfare
00:53:40
and worse is you know has come with
00:53:43
other re you know kind of industrial
00:53:46
changes of this magnitude so it doesn't
00:53:48
surprise me that we have a lot of
00:53:50
anxiety by people that they may lose
00:53:52
their job and I think there are people
00:53:54
out there who are kind of forecasting
00:53:56
into the future in ways that are scary
00:53:59
to you know the average person who's
00:54:01
who's listening to this and I don't
00:54:02
think that's particularly helpful so
00:54:04
could we do a better job messaging no
00:54:06
doubt about it. But if I just rewind to
00:54:09
kind of where we started, I actually
00:54:11
think the industry is um you know this
00:54:15
is going to be a pivotal year for the
00:54:16
industry to demonstrate right how this
00:54:19
is really beneficial for humanity. I
00:54:22
think we're going to be able to
00:54:23
demonstrate that it's very beneficial
00:54:25
from a health care perspective, from a
00:54:27
drug discovery perspective, from an
00:54:29
education perspective, etc. But we need
00:54:31
to have a coordinated effort because
00:54:32
Chimas right is deeply unpopular in the
00:54:35
country. I happen to be on the
00:54:37
optimistic side of this. 70% of the jobs
00:54:40
that exist in the United States today
00:54:42
did not exist 40 years ago. Right.
00:54:44
Right. We've gone through the digital
00:54:47
disruption that put a lot of people out
00:54:49
of work. But the abundance and the and
00:54:52
the recreation of new jobs, right,
00:54:54
expanded the pie for for largely
00:54:57
everyone. I think that will be the case
00:54:58
here. If you listen to Daario, he says
00:55:00
the concern is that the disruption
00:55:02
occurs at a faster and and and and
00:55:04
bigger rate and so that we can't keep up
00:55:06
with kind of that replacement. I think
00:55:08
that's another fine point. But if we
00:55:10
just go back to where we started the
00:55:11
conversation, which was are these good
00:55:14
investments, right? Do
00:55:16
>> That's not the conversation. No, of
00:55:18
course they're good investments. Of
00:55:19
course you're going to make money.
00:55:20
>> No, no, no. That's not what it's about.
00:55:22
>> That's not it.
00:55:23
>> He asked the question. You made the
00:55:24
argument. Jason asked the question,
00:55:27
right? Are these companies simply
00:55:29
selling do tokens at a loss? Right? And
00:55:32
we moved into this.
00:55:33
>> No, no, no. They're they're selling at a
00:55:35
profit. I'm buying them and losing
00:55:36
money.
00:55:37
>> Right.
00:55:38
>> In the 1849 gold rush,
00:55:40
>> Anthropic and OpenAI and all of these
00:55:42
model makers are selling the pick and
00:55:44
shovel in the gold rush.
00:55:46
>> I am buying it and I'm trying to pan for
00:55:48
gold. But as with the gold rush, most of
00:55:51
these companies will go out of business.
00:55:53
And all I'm saying is if we are really
00:55:55
circumspect and honest, there is still
00:55:58
way more to figure out than has been
00:56:00
figured out. This is not a solved
00:56:02
problem. And I think we it would behoove
00:56:04
everybody to just tell the truth about
00:56:05
this. It would be way better to be
00:56:08
honest. This is not figured out.
00:56:10
>> I would say I think the data the cards
00:56:11
that are being turned on the table move
00:56:13
me in the exact opposite direction.
00:56:15
>> Okay, let me get Sachs involved and then
00:56:17
I'll give my take. Sax, you have any
00:56:18
thoughts here? Well, I I have some
00:56:20
thoughts on the question you asked about
00:56:22
is the industry doing a good job with
00:56:25
PR. I think the answer is clearly no. I
00:56:27
think they are scaring the be Jesus out
00:56:29
of the public and that's why the
00:56:30
popularity is right down there with I
00:56:33
don't know what was it Iran. I mean it's
00:56:35
pretty I think we're a little more
00:56:37
popular than than Iran. But um but look,
00:56:40
>> Iran Iran's had 100 years to [ __ ] it up.
00:56:42
So we've only had two.
00:56:43
>> Yeah. Or 47 anyway.
00:56:45
>> Yeah. Iran's had 47 years to We've only
00:56:47
had two. But it is very much a
00:56:49
US-specific problem. If you look at
00:56:52
data, sentiment data across countries.
00:56:55
What you'll see is that other countries
00:56:56
are much more optimistic about AI than
00:56:59
the US. I think the Stanford did a study
00:57:02
on AI optimism. They simply asked the
00:57:04
question, do you think AI is going to be
00:57:06
more beneficial than harmful? Something
00:57:08
like 80% of people in China said yes. In
00:57:10
the US, it was in the 30s and it might
00:57:12
be even lower now. And it's not just
00:57:14
China and the US. you see across Asian
00:57:16
countries, they tend to be pretty
00:57:17
optimistic and then the US and Western
00:57:20
Europe tend to be pretty pessimistic
00:57:21
about it. I think that's less about the
00:57:24
reality of AI and more about our media
00:57:26
environment and who influences it.
00:57:28
Obviously, you have the influence of
00:57:30
Hollywood. It's created a lot of
00:57:31
dystopian films about AI. You've got the
00:57:35
fact that like we talked about these
00:57:36
CEOs are doing a horrible job and they
00:57:39
keep talking about putting everyone out
00:57:40
of business. I mean, this has I think
00:57:42
been not accidental. Well, I would say
00:57:44
some of these CEOs are speaking this way
00:57:47
because they're not very good at comms.
00:57:49
I think others are actually doing it
00:57:51
because they see a strategy there. You
00:57:54
know, they're going for a regulatory
00:57:55
capture.
00:57:56
>> Such a good point. It's delusions of
00:57:58
grandeur plus they're positioning their
00:58:00
companies. Yeah.
00:58:01
>> It could be for financing reasons like
00:58:03
you've mentioned Schmoth. like they want
00:58:04
to tap the stuff for fundraising, but
00:58:06
also I think that some of it is to
00:58:10
create a regulatory backlash that they
00:58:13
can then control. You know, create a
00:58:14
licensing scheme or permissioning
00:58:16
scheme. And that's a big part of it,
00:58:19
too. And then I think you just have the
00:58:21
fact that in our media environment, the
00:58:24
scare stories are the ones that get a
00:58:26
lot more attention than the heartwarming
00:58:29
stories. You know, if it if it bleeds,
00:58:31
it leads type thing. So you can just see
00:58:33
how unpopular it is for all of these
00:58:34
reasons.
00:58:35
>> You know, New York is about to outlaw
00:58:36
medical and legal advice from AI
00:58:39
chatbots, which by the way, that's
00:58:40
probably the the most obviously
00:58:43
>> valuable and highest ROI thing for for a
00:58:47
consumer
00:58:49
people the worst. Hurts poorest people
00:58:50
the worst. But do you understand like if
00:58:52
you're say a professional association
00:58:55
that sees it as your job to protect your
00:58:57
industry from disruption, you might
00:59:00
actually want to spread FUD about AI in
00:59:03
order to then seek those protections
00:59:05
through your state legislature.
00:59:07
>> Well, if you have companies that are
00:59:09
fanning those flames and those companies
00:59:11
are the ones in the industry, it's
00:59:12
making your job even easier.
00:59:14
>> Well, just think about the poorest
00:59:15
person. they they can't afford a lawyer
00:59:17
and they want to do their own research
00:59:19
and they research the legal stuff to in
00:59:22
order to fight an eviction or there are
00:59:24
poor people who don't have a primary
00:59:25
care doctor. They're not insured and
00:59:27
they find a way to deal with, you know,
00:59:30
some medical issue they're havinging.
00:59:31
This is the greatest thing. It's the le
00:59:33
it's a it levels the playing field for
00:59:36
poor people or people without resources.
00:59:39
So, this is the craziest legislation
00:59:41
ever. New York City is a disgrace. I
00:59:43
give them out of the week. No, the pro
00:59:44
the problem is very specifically that
00:59:47
these people that rely on these models
00:59:50
to make a healthc care diagnosis or get
00:59:52
a legal opinion to help improve their
00:59:54
lives, the makers of those tools are
00:59:57
telling everybody that they're about to
00:59:59
bring death and destruction upon the
01:00:01
economy and the world. So then the
01:00:03
lawyers and the doctors are like, "Well,
01:00:05
then maybe we should slow this down."
01:00:06
and they tell their lobbyists who then
01:00:08
go to New York and then tell the New
01:00:10
York legislators, hey, uh, these guys
01:00:13
are like trying to wreak havoc and then
01:00:15
they're like, oh yeah, well then maybe
01:00:17
we should shut it down. That is the loop
01:00:19
that's happening.
01:00:20
>> That was the Bernie Sanders moment this
01:00:22
week where he said we ought to have a
01:00:24
moratorum on all data centers being
01:00:26
built in the United States because AI is
01:00:28
dangerous. That was his message and that
01:00:30
is what he's pushing. And actually that
01:00:32
brings me to another point which is if
01:00:34
you look closely at Bernie Sanders
01:00:36
messaging, one of the talking points he
01:00:37
used was literally verbatim from Future
01:00:39
of Life Institute. Uh I think it was
01:00:41
something about how AI is less regulated
01:00:43
than a sandwich shop, which is just not
01:00:45
true. But future of life is one of these
01:00:47
EA funded doomer think tanks and they've
01:00:50
got something like a billion dollars.
01:00:52
Vitalic Butering donated $600 million of
01:00:54
dog coins to to the Anyway, you've got I
01:00:58
mean this is one really weird sort of
01:01:00
quirk about our media environment is we
01:01:02
have these EA funded think tanks with
01:01:05
literally billions of dollars. You know,
01:01:07
it's guys like Dustin Moscowitz and
01:01:10
>> they're the Dels, they're the Democratic
01:01:12
Socialist kind of arm. It's so weird
01:01:13
that like New York is now taking the
01:01:15
crown of most [ __ ] state from
01:01:17
California. I don't know how this
01:01:18
happened, but they just seem to be
01:01:20
making every mistake possible. Yeah, but
01:01:21
but hold on, let me just finish my point
01:01:22
because you've got these, let's call
01:01:24
them doomer think tanks funded by these
01:01:25
EA billionaires. They have literally
01:01:27
billions of dollars. You can influence a
01:01:29
lot of public discourse with that a lot.
01:01:32
>> Sure.
01:01:32
>> And they are behind a lot of the nimi
01:01:35
stuff around data centers. They've been
01:01:36
spreading a lot of the FUD around data
01:01:39
centers increasing your electricity
01:01:40
prices, which again I think is a solved
01:01:42
problem now because the users, the AI
01:01:45
companies have agreed to pay for the
01:01:47
incremental costs and stand up their own
01:01:48
power generation. They've been spreading
01:01:50
a lot of stories about water usage,
01:01:52
which is just totally made up. I mean,
01:01:54
the modern data centers recirculate
01:01:56
water, so they don't use up water. But
01:01:58
again, you've kind of got these doomer
01:02:00
groups who are just trying to stop AI
01:02:02
however they can.
01:02:04
>> And they're extremely well funded and
01:02:05
they're having a big impact. And I think
01:02:06
actually this is one of the reasons why
01:02:09
you're seeing in the US again that AI I
01:02:11
think is it's basically the least
01:02:13
popular thing they can pull for except
01:02:14
for the Democratic party and Iran. And
01:02:16
by the way, FLI, the Future Life
01:02:18
Institute, they also fund journalism
01:02:20
fellowships and endowments at
01:02:22
publications. So,
01:02:24
>> who probably writes negatively about AI?
01:02:26
>> Yeah,
01:02:26
>> exactly.
01:02:27
>> Here, Sax, look at this chart. It goes
01:02:29
back now a little bit earlier than 23,
01:02:31
but I had the data accurately from 23,
01:02:34
so we're going into the fourth year.
01:02:36
About 40% of all protested data centers
01:02:39
in America get cancelled.
01:02:42
And so, in 2023, this was a non-issue.
01:02:44
There were a few data centers that were
01:02:46
protested and a few of those 40% of them
01:02:49
I think it was literally like one or two
01:02:51
got cancelled but then starting in 24
01:02:54
when you had this divergence of
01:02:55
messaging or this chaotic slipshot
01:02:57
messaging and it was just a fever pitch
01:03:00
to raise money what you started to see
01:03:02
was this fermenting of negative
01:03:05
perspective by individual people on the
01:03:07
ground and so in 2024 about 40% of all
01:03:11
protested data centers were cancelled
01:03:13
still a small number, you could ignore
01:03:15
it. But last year was when the bottom
01:03:17
fell out. We had about 25 data centers
01:03:20
canled, about five gigawatts that got
01:03:24
cancelled. If you use Sarah Frier's
01:03:26
number, that's $50 billion a year of
01:03:29
revenue,
01:03:31
which is off the table because of what
01:03:33
happened in 25. Now, that has
01:03:35
implications to everybody. Look at the
01:03:37
amount of taxes that that would actually
01:03:39
raise for federal and local and state
01:03:41
governments. All gone. Vanished. in 26
01:03:45
just at the end of February there are
01:03:47
about a 100 data centers being protested
01:03:50
which if you flow that through will mean
01:03:53
about 40 will get cancelled and that
01:03:57
number right now is about 7 gawatt so
01:03:59
another $70 billion a year of revenue so
01:04:03
just last year and this year we've taken
01:04:05
off the table $120 billion of [ __ ]
01:04:08
revenue per year this is a wakeup call
01:04:11
that this messaging is These people are
01:04:14
not doing what is right on behalf of a
01:04:16
very nent and critical industry for
01:04:18
America. There's only so much that Sachs
01:04:21
can do, the White House can do. All
01:04:23
these other people are kind of at the
01:04:25
periphery. But if the people that are on
01:04:26
the ground don't get their [ __ ]
01:04:28
together, this is a national disaster.
01:04:30
Just to give people a uh sense of where
01:04:33
this is happening, almost all of these
01:04:35
cancellations are Virginia and Indiana
01:04:37
according to some just cursory research
01:04:39
here. And there have been zero
01:04:41
cancellations due to local opposition
01:04:43
here in the great state of Texas where
01:04:45
we have over 150 gawatts of data
01:04:48
capacity requests. So if you want to do
01:04:50
this, come to Texas, talk to Abbott,
01:04:53
talk to Ted Cruz. You just CC them on
01:04:54
your tweet and they'll have you to the
01:04:56
poker game and they will greenlight it.
01:04:58
Brad, before we move on, want to get
01:05:00
your opinion on open source and how
01:05:03
powerful it is and how powerful Apple
01:05:06
silicon is getting. I'm not sure how you
01:05:08
factored this in at alimter into your
01:05:11
thinking, but this seems to me to be a
01:05:13
massive headwind against the two big
01:05:16
bets you have
01:05:18
all of these open source models. We
01:05:20
started running them picking up about
01:05:21
85% of our tokens right now and every
01:05:24
startup I know is saying we are standing
01:05:27
up our local models and we only use the
01:05:30
top models, the paid ones uh when we
01:05:33
have jobs we can't do. So I I'm just
01:05:34
curious your thoughts on that and you
01:05:36
can add to that the auto research
01:05:38
project from Karpathy that came out this
01:05:41
weekend for people who don't know we now
01:05:43
have a group of tinkerers who are
01:05:45
setting up their open clause and now
01:05:47
setting up large language models and now
01:05:48
trying to train them with this auto
01:05:50
research tool. This seems like a
01:05:52
parallel track that could be material.
01:05:54
I'm just curious if you're monitoring it
01:05:56
at all.
01:05:56
>> I mean first I would say that uh I am
01:06:00
very enthusiastic for open source.
01:06:02
>> Okay. We see it in widespread use
01:06:04
everywhere. But here's the interesting
01:06:06
thing, you know, for the advanced
01:06:08
companies, they're doing some like
01:06:09
planning, you know, with the frontier
01:06:12
labs and then they're kind of doing the
01:06:13
execution, if you will, with the open
01:06:15
source models. So they're running an
01:06:16
ensemble of model strategy. But here's
01:06:19
what I think's more impressive. We have
01:06:22
incredible open- source models nearly on
01:06:25
the frontier. And notwithstanding that,
01:06:28
we're seeing companies like Anthropic
01:06:31
add five or six billion dollars of
01:06:33
revenue in a single month, which is
01:06:35
extraordinary. We've never seen anything
01:06:37
like it in technology. And that's in the
01:06:40
face, Jason, of all this open- source
01:06:42
usage. So, what does it tell me? It
01:06:44
tells me that the TAM
01:06:46
>> is dramatically bigger than any of us
01:06:48
think that it is. And that, you know,
01:06:51
when we when we look back on this
01:06:53
period,
01:06:54
>> you know, that will be the big takeaway.
01:06:55
It's a takeaway with Uber, the takeaway
01:06:57
with Google, the takeaway with Amazon.
01:06:59
The TAM was way bigger. We've crossed an
01:07:01
important threshold. Open source will be
01:07:03
a part of it, but clearly the front
01:07:05
Frontier Labs can do well even in the
01:07:07
face of it. All right, a little
01:07:09
housekeeping. The All-In Summit is
01:07:11
coming and liquidity is uh I think it's
01:07:15
sold out. We're about to sell out. We
01:07:16
might have a couple tickets left. You
01:07:18
can find both events at allin.com/events
01:07:21
and all-in summit tickets if you want to
01:07:23
get there quickly before they sell out
01:07:25
September 13th, 14th, and 15th in Los
01:07:27
Angeles. Again,
01:07:30
uh for all the all-in listeners, we're
01:07:32
launching a survey today. This is super
01:07:34
important to take the all-in survey. If
01:07:36
you made it to this point in the
01:07:37
podcast, if you are an early true
01:07:40
believer in the pod, if you're one of
01:07:42
the uh all-in stands, we need you to
01:07:44
fill out this survey.
01:07:45
Allin.com/servey.allin.com
01:07:46
allin.com/servey.
01:07:48
All right, let's uh wrap up with this
01:07:50
final story that went viral. The
01:07:52
millionaire tax has hit Washington
01:07:54
State. Howard Schultz, CEO uh longtime
01:07:57
CEO of Starbucks
01:08:00
has bailed and he's gone to Miami.
01:08:02
>> Surfside. He's in Surfside. He He bought
01:08:05
a He bought a condo in Surfside.
01:08:07
>> He uh he pulled a JCO.
01:08:08
>> He pulled a JCL.
01:08:10
>> I think you mean a sexy boo. Well, yeah,
01:08:12
but I was never I was never a Starbucks
01:08:14
liberal before I left the state of
01:08:16
California.
01:08:16
>> Listen, I I don't know how many times I
01:08:18
have to make this correction. I am a
01:08:19
moderate. I literally voted four
01:08:21
elections in a row for Republicans.
01:08:23
People have asked me for the receipts.
01:08:24
Pitakei, Giuliani, um, and Bloomberg. I
01:08:29
literally for almost a decade voted
01:08:32
exclusively Republican. Washington's
01:08:34
millionaire tax passed this week. Here's
01:08:36
what the tax is. People making more than
01:08:38
$1 million a year will pay an extra 9.9%
01:08:42
in tax starting in 2029.
01:08:45
The budget center estimates the tax will
01:08:47
impact 30,000 households, bring in
01:08:49
another 4 billion uh for the state's
01:08:51
general fund. Um the funds are supposed
01:08:54
to go towards public schools, higher
01:08:55
education, healthcare. In a huge
01:08:57
coincidence, on the same day the new tax
01:08:59
was passed, Howard Schultz, the
01:09:01
billionaire Starbucks founder, um
01:09:04
>> a huge coincidence, did you say?
01:09:06
>> Yeah. just
01:09:08
just unrelated stories.
01:09:10
>> Unrelated an unrelated story. Yeah.
01:09:11
Unrelated news.
01:09:12
>> He will be leaving Seattle after a 44
01:09:14
year run because he found out about
01:09:17
these incredible Cuban.
01:09:18
>> There was there was an opportunity.
01:09:20
There was an opportunity to buy a $44
01:09:22
million condo in Surfside. He couldn't
01:09:23
pass it up. It just happened to be on
01:09:25
the same day that they passed a
01:09:26
millionaire tax.
01:09:28
>> He he had the uh Cuban sandwich at Liss
01:09:30
Sanguich and uh he he fell in love.
01:09:34
Schultz has been getting crushed after
01:09:36
saying when he ran for president that he
01:09:38
would be willing to pay more taxes. Uh
01:09:40
Bezos obviously left back in November of
01:09:42
2023
01:09:44
>> and people speculated maybe the 7%
01:09:46
capital gains tax um would um have
01:09:50
influenced that. Who knows? So I guess
01:09:53
Shimoth what is the endgame here?
01:09:56
Because for these local politicians,
01:10:00
they must have learned the lesson that
01:10:04
people of means can move. They have the
01:10:08
ability to buy new homes, put their old
01:10:11
homes on the market. They're very mobile
01:10:14
and they could even leave the United
01:10:16
States and go to Singapore or Dubai or
01:10:18
other locations in the world. Why are
01:10:21
they still enacting these? and will they
01:10:22
continue to sell enactes until we get to
01:10:24
60 70% tax rates and we just lose all of
01:10:28
the creators and this becomes an
01:10:30
Randian.
01:10:30
>> I think that state politicians on the
01:10:33
west coast are very ineffective and not
01:10:36
not very smart. Nick, there's a there's
01:10:39
a tweet that was published I think maybe
01:10:41
it was a infographic that showed net
01:10:43
migration rates of every single state
01:10:47
for 2025. Washington is a few months
01:10:50
behind California in trying to enact
01:10:54
these stupid taxes. And the reason
01:10:58
they're stupid is these kinds of things
01:11:00
don't work at the state level. And we
01:11:02
know what it's already done in
01:11:03
California because the Hoover
01:11:04
Institution just published something
01:11:06
this morning and it's a complete
01:11:08
indictment of of what the billionaire
01:11:11
tax was trying to do. And by the way,
01:11:12
this billionaire tax is only polling
01:11:15
right now 25% of the votes it needs. So
01:11:18
maybe it'll find a way to get on the
01:11:20
ballot and then even then it'll have an
01:11:22
uphill climb to get voted in. But look
01:11:24
at the destruction that it has done in
01:11:26
California by just announcing it. The
01:11:29
Hoover Institution basically ran this
01:11:31
Monte Carlo simulation. They ran 100,000
01:11:34
runs and in 71% of those runs, it comes
01:11:37
out with a negative NPV.
01:11:40
And if you expected value it out, it's
01:11:43
about a $25 billion hole. They also
01:11:46
found that they overcounted the number
01:11:48
of billionaires in California. So that
01:11:50
number was wrong. They underounted the
01:11:53
amount of revenue that they pay. So that
01:11:55
was wrong. And they overcounted the
01:11:57
estimate of how much money that they
01:11:58
would make. So that
01:11:59
>> they're not good at math. They're not
01:12:00
good at math.
01:12:01
>> So when you add it all up, they thought
01:12:02
they were going to make a hundred.
01:12:04
They're actually going to make 40. The
01:12:06
people that left pay, you know, $3 to5
01:12:08
billion
01:12:10
a year of taxes. It's going to create a
01:12:13
$25 billion hole. You're going to have
01:12:15
the middle class that's now going to
01:12:16
have to foot this because this is net
01:12:18
revenue that's not going to come into
01:12:20
the budget. That's about 2500 per middle
01:12:24
class household. There's about 10
01:12:25
million in California. So that's what's
01:12:28
happened just by making the threat.
01:12:31
>> Washington had a 23-hour debate and
01:12:33
passed the law. So I suspect when you
01:12:36
look back on this in 18 or 24 months,
01:12:39
it'll be as bad or worse than
01:12:41
California. These things don't make
01:12:44
sense. The reason they don't make sense
01:12:46
is that you are putting good money after
01:12:48
bad. We all know that money that goes to
01:12:51
the state governments are wasted. We
01:12:53
just don't know how much. And so when
01:12:55
you keep asking more, eventually the
01:12:58
smart people say enough's enough. I'm
01:12:59
out of here. We might find out how much.
01:13:01
I think uh Bari Weiss is on the case. I
01:13:03
don't know if you saw her do her CBS
01:13:04
report this week. She's she's going hard
01:13:07
for fraud. And until you get fraud out
01:13:10
of the system, I don't think you have
01:13:11
the moral high ground to raise taxes. I
01:13:14
think that should be the message that
01:13:15
all Americans
01:13:16
>> send to politicians.
01:13:18
>> That should be your campaign promise
01:13:19
when you run whatever whatever you do.
01:13:21
>> Hi, I'm Jason Caliganos and I will get
01:13:22
rid of fraud and lower your taxes.
01:13:25
Look, you may have seen an even more
01:13:27
severe attack was proposed at the
01:13:28
federal level where Bernie Sanders and
01:13:30
and I think Roana came out with their
01:13:34
version of a national wealth tax where
01:13:36
it wasn't just 5% once like in
01:13:38
California, it was 5% per year,
01:13:40
>> per year.
01:13:41
>> So, in other words, in roughly 20 years,
01:13:42
the federal government's just going to
01:13:44
take all of your money. I mean, that's
01:13:45
it. Look, this is socialism. This is
01:13:47
another way to get to the same end
01:13:48
point, which is the government owns
01:13:50
everything.
01:13:51
>> They seize. Well, I mean, the seizure
01:13:53
part of it, I think, is the nuance point
01:13:56
we have to get across, which is if you
01:13:57
earned it and paid your taxes already,
01:14:00
>> and then the state can just decide, you
01:14:02
know what, we didn't take enough 10
01:14:04
years ago, we need to go seize that.
01:14:05
Hey, when you sold Yammer, we didn't
01:14:08
take enough. We need to take it now.
01:14:09
>> Raise your hand if you believe the
01:14:11
things you own are better off being
01:14:14
owned by Bernie Sanders and Roana. Raise
01:14:16
your hand if that's what you believe. I
01:14:17
mean, you'd have to be an idiot to
01:14:19
believe that. Well, I'll tell you the
01:14:21
last time we saw these proposals of
01:14:24
asset seizures was during the guilded
01:14:26
age, you know, 1870 to 1920.
01:14:30
>> You know, then it was Carnegi and
01:14:31
Rockefeller. You know what's
01:14:33
interesting? I went back and looked at
01:14:34
it. There were actually like it was
01:14:36
actual warfare. We had hundreds of
01:14:38
people killed in clashes um you know
01:14:41
during the great railroad strike, the
01:14:42
Pullman strike, etc. Um and it was all
01:14:45
over this. And so I think shame on the
01:14:47
politicians that are fanning the flames
01:14:49
of class warfare. Um we all need to
01:14:52
bring the temperature down. There are
01:14:54
fair debates on whether states have
01:14:57
enough resources to fulfill their
01:14:59
obligations to the citizens. There are
01:15:00
fair debates about fraud. All has to be
01:15:03
taken on. But I think it's interesting
01:15:04
in the state of California, right? I
01:15:06
think the teachers union is against the
01:15:08
the billionaire's tax because they know
01:15:10
it's going to lead to less dollars for
01:15:12
education for the state of California.
01:15:13
and Matt Mayan who's running for
01:15:15
governor is against the tax. The current
01:15:16
sitting governor, Democrat, against the
01:15:18
tax, right? They all need to step up and
01:15:21
explain not just that they're against
01:15:23
the tax, but we you're either on the
01:15:26
side of business and entrepreneurs and
01:15:28
creativity and moving the state forward
01:15:31
and growing the economy or you're
01:15:32
against it. And that's what's at stake
01:15:34
here. And fortunately, the outcome of
01:15:36
the battle uh you know during the
01:15:38
guilded age was that America didn't
01:15:41
abandon entrepreneurialism. We didn't
01:15:43
abandon capitalism like Europe did and
01:15:46
now we leaned into it. Now we played it
01:15:47
out.
01:15:48
>> One point on that. So you mentioned that
01:15:50
some of the unions in California are
01:15:52
opposed to this this asset seizure tax.
01:15:55
That's only because they weren't cut in
01:15:57
on it,
01:15:57
>> right?
01:15:58
>> So there's already rumors that the
01:16:00
California Teachers Association is
01:16:02
working on their own version.
01:16:03
>> Oh boy.
01:16:05
>> Billionaire asset seizure
01:16:07
>> for not this election cycle, but for the
01:16:08
next one. Yeah.
01:16:09
>> And next time they're going to have
01:16:11
their ducks in a row and they're going
01:16:12
to have all the pigs at the trough and
01:16:14
all the unions are going to get together
01:16:16
because the SEIU
01:16:18
uh they did this on their own. So again,
01:16:21
they didn't allow all the other groups
01:16:23
to wet their beak. So I think that
01:16:25
unfortunately I think that's going to be
01:16:27
corrected. If this one doesn't pass,
01:16:29
they'll correct it for 28 and it's more
01:16:31
likely to pass because they're all going
01:16:33
to do it. And the other thing is that I
01:16:36
think that by 28 this national wealth
01:16:39
tax will just be a standard part of the
01:16:41
Democratic platform.
01:16:43
>> Table stakes. Yeah.
01:16:44
>> Yeah. It's table stakes. I think that
01:16:45
you know the Bernie Sanders Roana
01:16:46
position will be the position of the
01:16:48
Democratic party. And you even see Gavin
01:16:50
Newsome creating wiggle room for himself
01:16:52
to embrace his position. If you look
01:16:54
closely at his statements formally
01:16:57
opposing the the BTA in California, what
01:17:01
he says is that a state can't do this by
01:17:02
themselves because they got 49 other
01:17:04
states to compete with.
01:17:06
>> So what he's saying is that, you know,
01:17:08
look, California is operating in a
01:17:10
comparative environment. One state can't
01:17:12
do it. And then he's leaving the other
01:17:15
part elliptical, which is well the the
01:17:17
federal government needs to do this. And
01:17:19
I think that you can expect him to
01:17:21
embrace that position by 2028. There is
01:17:24
a way out here from this socialist
01:17:26
movement. It's a it's very simple if you
01:17:29
think from first principles. What does
01:17:30
an American want? What does an American
01:17:31
family want? What what do mothers and
01:17:33
fathers in this country want? They want
01:17:35
to educate their kids. They want to be
01:17:37
able to own a nice home. They want to
01:17:38
have decent healthare. They want to have
01:17:40
healthy food. It's a very small subset
01:17:42
of issues. And AI is uniquely positioned
01:17:46
to solve a lot of these problems. and
01:17:48
entrepreneurs can come in and take these
01:17:50
highly regulated industries if we're
01:17:52
allowed to participate in them.
01:17:55
Education, we have to break this
01:17:57
accreditation, you know, uh, cartel. And
01:18:00
then housing, we have to break these
01:18:02
regulations like the great state of
01:18:03
Texas, Nevada, and Florida have. And
01:18:06
then you know when it comes to healthare
01:18:08
this is where AI could have a tremendous
01:18:10
impact and entrepreneurs could have a
01:18:11
tremendous impact in lowering the cost
01:18:13
of healthare and letting people solve
01:18:15
for that with their own um you know
01:18:18
healthcareled you know self-led
01:18:19
healthcare. These are the problems. If
01:18:22
you solve for these problems people's
01:18:24
homes, people's health, people's
01:18:25
education of their kids, we're going to
01:18:27
solve these problems and we don't need
01:18:29
to go to socialism and seize people's
01:18:31
assets. That's what entrepreneurs should
01:18:33
be doing. That's what entrepreneurs
01:18:34
should be working on and that's what
01:18:36
where the government can help. That's
01:18:37
where Trump's uniquely qualified. He is
01:18:39
the regulatory breaker. He got nuclear
01:18:42
back on the agenda. No other president
01:18:44
had gotten nuclear back on the agenda
01:18:46
for America. He can get housing back on
01:18:48
the agenda. We have to break those
01:18:50
things and not start foreign wars. My
01:18:52
earlier point and start creating houses
01:18:54
for Americans. Another amazing episode
01:18:56
of the All-In podcast. Gentlemen,
01:18:59
>> you always give yourself the last word,
01:19:01
Jal. Well, I go last. If you'd like, you
01:19:03
can moderate and I'll go first. But I
01:19:05
always go last, but you can you I'm you
01:19:08
could give me a round of applause or you
01:19:10
can throw a tomato. Go ahead, Sax. It
01:19:11
doesn't matter.
01:19:12
>> It's all good. It's all good. Go for it.
01:19:13
>> All right, everybody. Another amazing
01:19:15
episode of the All-In podcast. Thank
01:19:18
you, Bestie Brad, for joining us. Thank
01:19:20
you, Bestie Brad. Love you, boys.
01:19:22
Byebye.
01:19:24
>> We'll let your winners ride.
01:19:31
And it said, "We open sourced it to the
01:19:33
fans and they've just gone crazy with
01:19:34
it." Love you, Queen of
01:19:44
>> Besties are
01:19:47
my dog taking notice your driveways.
01:19:52
>> Oh man, my habitasher will meet up.
01:19:54
>> We should all just get a room and just
01:19:56
have one big huge orgy cuz they're all
01:19:57
just useless. It's like this like sexual
01:19:59
tension that we just need to release
01:20:01
somehow.
01:20:05
>> Your feet.
01:20:08
>> We need to get merch.
01:20:09
>> I'm going all in.
01:20:17
I'm going all in.
