00:00:00
All right, everybody. Welcome back to
00:00:01
the number one podcast in the world.
00:00:03
Your favorite podcast with your favorite
00:00:06
besties in the world, the All-In
00:00:08
podcast. Shabbath is traveling and
00:00:11
couldn't make it this week. So, our
00:00:14
favorite fifth bestie in the world, the
00:00:16
one, the only Brad Gersonner from
00:00:19
Alimter here to take a victory lap for
00:00:21
your uh Trump accounts. Congratulations,
00:00:25
brother.
00:00:25
>> Just getting started. Started here,
00:00:27
actually. Look forward to
00:00:28
>> Yeah. Yeah, we'll talk about it in
00:00:30
depth. And also with us from the great
00:00:32
state of Texas, my bestie and brother in
00:00:36
arms, David Saxs. How are you doing? How
00:00:39
you how you You're loving it, aren't
00:00:41
you? Freedom.
00:00:42
>> Loving it.
00:00:43
>> Yeah.
00:00:44
>> So much freedom.
00:00:45
>> Well, the ice storm is over and the
00:00:47
weather looks really good right now.
00:00:48
We're going to have 70s all weekend.
00:00:50
It's great.
00:00:51
>> Yeah. You figured out your internet.
00:00:53
>> That has been resolved. Like I said,
00:00:55
it's a lot easier to fix my internet
00:00:57
than to fix the state of California.
00:01:00
>> You fixed it in one week. So, yeah, I
00:01:02
would say that you're trending in the
00:01:03
right direction. And with us, of course,
00:01:06
David Freeberg of Oh. How are you doing,
00:01:10
oh, look at your little promo in your
00:01:12
hat. You're promoing with the hat. How's
00:01:14
Oh, we haven't heard about it for a
00:01:16
while.
00:01:16
>> I'll give you some updates on another
00:01:18
show. Today may not be the best day. I
00:01:19
do appreciate you asking.
00:01:21
>> I hear it's going great with the
00:01:22
potatoes.
00:01:23
>> You got the merch. I got merch. Yeah,
00:01:25
you guys can go to boosted.oh.com
00:01:27
and buy your gear today. Oh,
00:01:30
>> here's your quick science corner for the
00:01:31
day. They found a village called the Oh
00:01:34
site. It was an archaeological dig on
00:01:36
the Sea of Galilee, 26,000 years old.
00:01:39
And in this archaeological dig, they
00:01:41
found these little clay pots filled with
00:01:42
seed. So, it predated all of our
00:01:44
understanding of agriculture and plant
00:01:46
breeding and seed storage. So, it really
00:01:48
reinvented our understanding of of human
00:01:50
history with agriculture. So we named
00:01:52
the company Oh after that archaeological
00:01:54
site
00:01:55
>> and your first crop is going to be
00:01:58
potatoes. That's the poly market is
00:02:00
saying potatoes.
00:02:01
>> Farmers are planting our true seed.
00:02:03
World's first world's first true seed of
00:02:06
potato. Instead of planting 5,000 lbs of
00:02:08
chopped up potatoes, you plant a handful
00:02:10
of seed. Completely changes the
00:02:12
economics and the opportunity for potato
00:02:14
farmers around the world. Third largest
00:02:16
source of calories. Very excited. We're
00:02:18
going to market this spring. Farmers are
00:02:19
planting it in the field. So, it's a
00:02:21
great year for Oh. Thanks for asking.
00:02:24
>> So, first
00:02:25
>> invest in Oh, Jason.
00:02:27
>> I think Sax and I have 0.0. Yeah, I
00:02:30
think Besty's got boxed out on this, but
00:02:32
we'll promote it every week.
00:02:34
>> Sax is there, bro.
00:02:35
>> Oh, Sax is no comment. Sax got his beak
00:02:38
with Are you in Oh, or not?
00:02:40
>> I think we are investors, are we not?
00:02:41
>> Yeah, his his venture firm, Craft
00:02:43
Ventures. We're excited to have them on
00:02:45
the cap table. Yeah.
00:02:46
>> Oh, great. How about you, Brad? Did you
00:02:47
get a little slice?
00:02:49
>> I'm working on it. I'm working on it,
00:02:51
too. Let me know how it goes. If you can
00:02:53
ever find the founder and see if we can
00:02:55
get a little slice for Jake, I'd like to
00:02:56
slide a millie in there.
00:02:57
>> If you write a personal check, I'll let
00:02:59
you in right now.
00:03:00
>> Sure. Okay. I mean, the problem with
00:03:01
that is if if I write a personal check.
00:03:05
>> Never mind.
00:03:05
>> No, I I have to go to my LPs and then
00:03:07
get permission to do it. And then you're
00:03:09
going to crush it and then my LPs are
00:03:10
going to say, "Why isn't in the fund?"
00:03:11
So, that's the problem.
00:03:13
>> All right. Let's get going here. We got
00:03:14
a lot of news to get through. Epstein
00:03:17
files newest drop. DOJ published a
00:03:19
massive number of documents on Friday,
00:03:22
January 30th, under the Epstein Files
00:03:25
Transparency Act. Hundreds of
00:03:27
high-profile tech executives and
00:03:29
national figures were mentioned in the
00:03:30
files. Of course, none of those are
00:03:32
accused of any criminal wrongdoing.
00:03:35
>> Jal, you were in the files.
00:03:42
>> Yes, I have a couple of emails in the
00:03:45
files. Okay, Inspector Friedberg has a
00:03:47
few questions for you. Okay. Okay, let's
00:03:49
get started.
00:03:50
>> When did you first meet Jeffrey Epste?
00:03:52
>> I met Jeffrey Epstein in the late 90s at
00:03:55
the TED conference at specifically the
00:03:58
billionaires dinner which was hosted by
00:04:00
my book agent John Brockman.
00:04:03
>> And then did you see him in New York?
00:04:05
Did you visit him at his house or his
00:04:07
office or anywhere else in New York?
00:04:08
>> I've probably spoken to him for 45
00:04:11
minutes of my life. 30 minutes of that
00:04:13
was in the late 90s when I had Silicon
00:04:16
Valley Reporter magazine. He was a
00:04:18
billionaire financeier and he wanted to
00:04:20
invest in the magazine. I met with him
00:04:21
for 30 minutes which he said was too
00:04:23
small potatoes for him to be involved.
00:04:25
>> Where did you meet with him?
00:04:26
>> At his legendary town home.
00:04:28
>> You went to that house?
00:04:29
>> I visited him there and then I saw him
00:04:31
at the TED conferences at the
00:04:32
billionaire's dinner probably a half
00:04:33
dozen times.
00:04:34
>> You never went to the island?
00:04:35
>> I never went to the island. Was never
00:04:37
invited to the island. Was never invited
00:04:38
on the plane. Was never invited to the
00:04:40
ranch. None of that. When you went to
00:04:42
his house, did you see any young ladies
00:04:43
or did you see any of the stuff that's
00:04:45
reported?
00:04:45
>> No.
00:04:46
>> Did you ever get a massage from anyone?
00:04:48
>> No. No. I did trade an email with him,
00:04:50
which I didn't recall, but he in 2011
00:04:53
emailed me and said, "Hey, can you
00:04:54
introduce me to these people who were
00:04:56
doing this Bitcoin thing on your
00:04:57
podcast?" And I said, "Sure. Yeah, here
00:04:58
you go. I'll uh introduce you." I do
00:05:01
thousands of introductions a year
00:05:03
between our portfolio companies, people
00:05:05
on this weekend, startups, and
00:05:06
billionaires and financers. That's the
00:05:08
job of an early stage investor. Why did
00:05:10
you say hey pal in your email to him?
00:05:12
>> That's just a colloquialism I use like
00:05:15
as a general hey fella. If a fan comes
00:05:17
up to me and ask for a sub hey pal
00:05:19
thanks for saying that.
00:05:20
>> It wasn't on your radar that this is a
00:05:22
sexual predator etc.
00:05:24
>> Absolutely not. I think actually when
00:05:26
all that came out was 2018. That's when
00:05:29
I sort of became aware of it. There was
00:05:30
like a Miami Herald story or something
00:05:32
where they went into detail about how
00:05:34
heinous all this stuff was. I've been
00:05:36
saying here release all the Epstein
00:05:37
files. what he did was horrible.
00:05:39
>> Prosecute everybody 100% who was
00:05:42
involved in it. The end.
00:05:44
>> What about Gileain Maxwell? You had a
00:05:46
separate email that came out in the
00:05:47
Epste files with her.
00:05:48
>> I had met her as well at TED and I had
00:05:51
met her socially like in New York in
00:05:53
circles. When I met her, her dad, Robert
00:05:55
Maxwell owned, I believe the New York
00:05:57
Post or Daily News and she was a big
00:05:59
media executive and her sister was
00:06:01
involved in angel investing in
00:06:03
technology startups. So, they were just
00:06:04
in the scene. I think in hindsight, you
00:06:07
know, as a connector, if you ever seen
00:06:09
The New Yorker, Nick, you can throw up
00:06:10
the New Yorker story about me. I had
00:06:12
become famous in the first part of my
00:06:14
career as the connector. And the New
00:06:16
Yorker wrote like this 5,000word article
00:06:18
about how I knew everybody and was
00:06:20
connecting everybody. I think Epstein's
00:06:22
interest in me, if he had any interest
00:06:24
in me or Gilain, was in my ability to
00:06:26
connect high-profile people with them
00:06:29
and, you know, their business endeavors,
00:06:32
etc. So you had no knowledge of illicit
00:06:34
activities happening by Epstein or
00:06:36
Galain and you never participated in
00:06:37
any?
00:06:38
>> Absolutely not. Unequivocably not. I was
00:06:40
not involved in any shenanigans. Period.
00:06:42
Full stop.
00:06:43
>> Any more questions for the witness?
00:06:45
>> Well, let me make a a few observations
00:06:47
here. So, first of all, I 100% believe
00:06:49
Jay Cal. As I joked at our roast, he's
00:06:52
not an important enough player in the
00:06:54
grand scheme of things. You know, as I
00:06:56
joked, I said, "Who'd want to blackmail
00:06:57
a loser?" You know, obviously you're not
00:06:59
a loser, Jal, but uh but look, you you
00:07:01
were
00:07:02
>> well taken. Yeah.
00:07:03
>> We're learning from the Epstein files
00:07:05
that Epstein was some sort of hyper
00:07:07
networker. You were a connector. The
00:07:09
odds of the two of you coming across
00:07:12
each other in that time period was
00:07:14
basically 100%. But your contacts were
00:07:16
very minor. So that's point number one.
00:07:19
Now, I thought it was interesting, you
00:07:20
know, cuz I saw the the emails where
00:07:22
he's asking you for an introduction. The
00:07:24
fact that he was curious to meet the
00:07:26
quote unquote Bitcoin guys in 2011, I
00:07:29
thought was kind of interesting in and
00:07:31
of itself that, you know, whatever other
00:07:33
things Epstein was, he clearly had some
00:07:36
sort of nose for putting himself in the
00:07:38
middle of everything at a very early
00:07:40
stage.
00:07:41
>> I also thought it was interesting that
00:07:43
you were trying to warn Epste like crazy
00:07:45
Bitcoin guys, you know, like these are
00:07:48
some crazy crypto libertarians, you
00:07:50
know, you don't want to do business with
00:07:51
them or whatever. Anyway,
00:07:52
>> well, I It's interesting you point that
00:07:55
out because I had these guys on this
00:07:56
week in startups because I had heard
00:07:57
about Bitcoin early when it was like a
00:07:59
under a dollar. And I was like, "Yeah,
00:08:01
these guys are kind of weird. They're
00:08:03
not like entrepreneurs who want to raise
00:08:05
money. They're a foundation. They're
00:08:06
like Wikipedia and you you're not going
00:08:09
to be able to invest in it." I sort of
00:08:10
gave him that warning because I had
00:08:12
looked at like, "Hey, what are you guys
00:08:13
working on? Can you invest in the
00:08:14
Bitcoin project?" They're like, "No, no,
00:08:16
it's a nonprofit. It's it's a nobody
00:08:18
owns it." And I was like, "Oh, okay.
00:08:19
It's not even a nonprofit. Nobody owns
00:08:21
it." Yeah. basically. Yeah.
00:08:22
>> But anyway, it's just interesting that
00:08:24
he he obviously wasn't deterred by what
00:08:26
you said. He got very involved, I guess.
00:08:27
And there was a company called
00:08:28
Blockstream that he invested in along
00:08:30
with Reed Hoffman and and Joyo that
00:08:32
involves some of the Bitcoin core
00:08:34
developers. And again, this is all
00:08:35
coming out. Now, shifting gears,
00:08:38
>> I think another interesting part of this
00:08:39
is just how this is all being covered.
00:08:41
And there was an article in the New York
00:08:42
Times today talking about Epstein's
00:08:44
connections to Silicon Valley. And lo
00:08:47
and behold, you have a major photo there
00:08:49
despite I think your minor minor
00:08:51
tangential connection here.
00:08:54
>> And meanwhile, the people who have major
00:08:57
connections and a deeper relationship to
00:09:00
Epstein are being completely ignored.
00:09:02
>> H why is that? Why are they going after
00:09:04
me as opposed to Reed or
00:09:08
>> because you've become sort of rightcoded
00:09:10
by virtue of your association with Elon
00:09:12
and and being on this podcast. And so,
00:09:14
>> and you
00:09:16
>> and me and so you look in that article,
00:09:18
it's not just you. They really go after
00:09:19
Peter Teal. They go after Elon. But Reed
00:09:22
gets a total pass. I mean, he's just
00:09:25
mentioned
00:09:26
>> in a sentence with several other names.
00:09:28
So does Bill Gates in this particular
00:09:29
article, by the way,
00:09:30
>> which seemed to be, by the way, if you
00:09:32
were thinking of who had the most
00:09:34
contact with him for the longest
00:09:36
duration of time, it was Gates and Reed.
00:09:38
like they were involved with him up
00:09:40
until I don't know his death maybe or
00:09:42
2018 2019 they were on the island they
00:09:45
were on the plane they were at the ranch
00:09:48
they were very much involved in him to
00:09:50
and I know Joeyto was in it to raise
00:09:52
money from Epstein because Epstein was
00:09:56
this big funer of scientists which
00:09:58
looking back on its he's involved in
00:10:00
Bitcoin he's involved in physics and
00:10:03
scientists he's involved with the
00:10:05
world's most innovative tech leaders
00:10:07
what exactly is going on here? Like what
00:10:09
who is he working for? Do you think he
00:10:12
was a spy? I wonder what you've come to
00:10:15
think or an asset.
00:10:16
>> I don't have any more information than
00:10:17
anybody else does about this and I
00:10:19
haven't gone through all the Epstein
00:10:20
files. I've just seen the ones that have
00:10:22
been surfaced by other people tweeting
00:10:23
about them and that sort of thing. My
00:10:25
impression is that he clearly had
00:10:27
relationships with people in
00:10:29
intelligence, but I don't know whether
00:10:32
he was actually an asset. People use
00:10:34
this term that he was being run by
00:10:36
somebody. It seems to me that this guy
00:10:38
is kind of running himself and then he's
00:10:40
using lots of other people and
00:10:42
manipulating lots of other people and
00:10:43
putting himself in the center of a lot
00:10:45
of things.
00:10:46
>> And so is he working with intelligence?
00:10:48
Yes, certainly. It seems that way. But
00:10:52
is he working for them or are they being
00:10:55
put to use by him? Very hard to tell.
00:10:58
>> Yeah. You know, I not that I have a lot
00:11:02
to add here, but man, this guy's just a
00:11:04
scumbag, you know, and all of this
00:11:07
nonsense. It's totally tragic, but
00:11:10
David, this is why nobody trusts
00:11:12
institutions or powerful elites or any
00:11:15
of this garbage,
00:11:16
>> right? Like, it's the this thing is like
00:11:19
trickling out over years and people like
00:11:22
want to put it behind it, but you can't.
00:11:24
Nobody's been charged here. What about
00:11:26
all the, you know, all the people in the
00:11:28
emails? like why don't we see any
00:11:30
charges? And then girly tweets this week
00:11:32
and I totally agree. This guy's under
00:11:34
complete surveillance, suicide watch in
00:11:37
a little tiny jail cell and all of a
00:11:40
sudden the guy ends up dead. We have no
00:11:42
investigation as to how this guy dies.
00:11:44
>> Yeah.
00:11:45
>> Right. It just seems to me that this is
00:11:47
the type of [ __ ] that undermines trust
00:11:50
in institutions, that undermines trust
00:11:52
in all the people who are listed here
00:11:54
because literally from Peter Aia to Bill
00:11:56
Gates, the people that many across the
00:11:59
country looked up to for advice on key
00:12:01
things. All of a sudden, they're saying
00:12:03
stuff that just undermines credibility.
00:12:06
>> And how come the 30 people or whatever
00:12:08
the number is who were investigated
00:12:10
alongside him have not been prosecuted?
00:12:12
Like that's the thing that's crazy to
00:12:13
me. that screams of some weird
00:12:15
conspiracy here. And I do think his
00:12:17
death is obviously, you know, very
00:12:19
suspicious.
00:12:20
>> But do you do you think it's because
00:12:22
they didn't find evidence of all the
00:12:24
underage prostitution, sex trafficking?
00:12:27
Is that Do you think that it's actually
00:12:28
the case they didn't find evidence about
00:12:30
any of that sort of stuff? The thing
00:12:32
that came out eventually was that his
00:12:34
nonprosecution agreement, and I'm I'm
00:12:36
not an expert on it, that he did with
00:12:38
the Miami group included like all the
00:12:42
other people also couldn't be prosecuted
00:12:44
for it. So, there's something fishy
00:12:46
going on and I think people should
00:12:47
reinvestigate his charges and that whole
00:12:50
thing. And where's the FBI in all of
00:12:52
this? Like, they did a lot of
00:12:53
investigating. So, yeah, why didn't they
00:12:55
prosecute anybody else? It's very
00:12:57
strange.
00:12:57
>> Can I ask you guys a question? because
00:12:59
this revealed a lot of very private
00:13:01
communication in a very public way of
00:13:04
what Brad points out are very public
00:13:06
powerful people. There is this great
00:13:08
book, I've mentioned it in the past, by
00:13:10
Steven Baxter and Arthur Clark called
00:13:12
the light of other days and it's all
00:13:13
about how basically all the world's
00:13:15
information becomes available to
00:13:16
everyone. Like what if you could read
00:13:18
every other person's text, email, IM, DM
00:13:21
that they've ever had with everyone else
00:13:23
in the world? What would the world look
00:13:25
like? And it's very kind of intimately
00:13:26
revealing about people's quiet private
00:13:29
conversations versus their public
00:13:30
personas or their external
00:13:32
conversations. Is this just like hey
00:13:34
people feel entitled to act so
00:13:36
maliciously and deviously in in one
00:13:38
sense because they're wealthy because
00:13:39
they're powerful and that's a very
00:13:41
unique thing for wealthy and powerful
00:13:42
people.
00:13:43
>> I think it's too dismissive of people
00:13:45
for their scumbag behavior. David, it
00:13:48
the fact of the matter is people don't
00:13:50
do this. This isn't normal. We can't
00:13:52
normalize it. And worse yet, the level
00:13:55
of hypocrisy
00:13:57
that the very people that are acting the
00:14:00
worst are out there lecturing others uh
00:14:03
throughout this entire period of time.
00:14:06
And I will say this, you know, to the to
00:14:08
the ear of my sister or mother in rural
00:14:11
Indiana. They hear the coastal elites
00:14:13
lecturing them all the time and then
00:14:15
they have they juxtapose it against what
00:14:18
they read in the Epstein files. This is
00:14:20
why we have a lack of trust. Brad, you
00:14:22
speak about the corruption of power
00:14:23
centers. I think a major one has to be
00:14:25
the New York Times. The number one
00:14:26
person in the Epstein files from Silicon
00:14:28
Valley, which is Reed Hoffman mentioned
00:14:30
2,600 times, had a multi-year
00:14:33
relationship with Epstein. They call
00:14:35
each other very good friends. They did
00:14:36
deals together. Reed stayed at the
00:14:39
trifecta, which is not just the island,
00:14:41
but the townhouse and the New Mexico
00:14:43
ranch. And if you're going to write
00:14:46
about Silicon Valley, Reed was the one
00:14:48
who introduced Epstein to Peter Teal and
00:14:51
Elon Musk and Mark Zuckerberg organized
00:14:53
that famous dinner. How can you not
00:14:55
mention that as the root of Epstein's
00:14:59
involvement in Silicon Valley and yet
00:15:01
read this gets a mention in one sentence
00:15:03
of that article along with several other
00:15:04
people? It's crazy. I mean, the New York
00:15:06
Times clearly has a list of people they
00:15:08
consider approved targets. They're all
00:15:10
rightcoded people like Elon or Peter
00:15:13
Teal and even JCAL because of his
00:15:15
association with us, I guess, and they
00:15:18
become the targets. But the people
00:15:20
who've donated hundreds of millions of
00:15:21
dollars to the Democrat party and have
00:15:24
paid for dirty tricks against Trump,
00:15:26
they basically are spared. Honestly,
00:15:28
this is just emblematic of the whole
00:15:31
institutional rot and the distrust in
00:15:33
the country, right? They're part of the
00:15:34
cabal. It's part of the institutions
00:15:36
that people are losing faith in. You
00:15:38
know, Epstein was a scumbag. And the
00:15:41
fact of the matter is we're not seeing
00:15:42
equal play on both sides.
00:15:44
>> All right, let's keep moving. SAS
00:15:46
companies are crashing out. $300 billion
00:15:49
of value was wiped from the S&P
00:15:52
Tuesday in the software and data stocks
00:15:54
category. Uh people are calling this the
00:15:56
Claude crash. I don't know if I buy
00:15:58
that, but on Monday, Anthropic, which
00:16:00
has been on a bit of a heater as we
00:16:02
talked about, announced that they added
00:16:03
a legal tool to Claude Co-work. If you
00:16:06
don't know what Claude co-work is
00:16:08
called, this is different than the
00:16:09
Claudebot that we talked about last
00:16:11
week. This is a essentially what uh
00:16:15
Claude code or a coding agent is. This
00:16:18
is for knowledge workers to automate
00:16:19
work and do multi-step instead of just
00:16:22
asking you a query to a large language
00:16:24
model, it would do a number of actual
00:16:25
actions on your behalf that you can
00:16:27
automate and run as cron jobs, as
00:16:29
regular jobs every day, every hour,
00:16:31
every week, whatever it happens to be.
00:16:33
This one specifically is kind of like a
00:16:35
plugin that allows you to do tasks uh
00:16:38
related to legal drafts and research.
00:16:40
What that meant to I guess retail
00:16:42
investors and we'll get into this Brad
00:16:44
since this is your specialtity is that a
00:16:46
lot of legal tech startups and uh public
00:16:49
companies were hit hard. Thompson
00:16:51
Reuters down 20%. Lexus Nexus which is a
00:16:53
database of case law uh was down 15%.
00:16:57
Legal Zoom which gives legal advice um
00:17:00
and documents down 15%. At the same
00:17:03
time, SAS uh has continued to be
00:17:06
negatively impacted by this concept that
00:17:09
uh software will be made bespoke in uh
00:17:12
tools and be wiped out. Figma down 13%,
00:17:15
Salesforce 11, Service Now 11, Adobe 8%.
00:17:19
And even before Tuesday's drop, and you
00:17:21
can get into this, Brad, software was
00:17:23
already the worst performing S&P sub
00:17:25
section for the year.
00:17:27
>> By the way, the numbers you report are
00:17:29
dramatic understatement. We we've wiped
00:17:31
out trillions of dollars in market cap.
00:17:33
Figma's down 80% from the high. Uh you
00:17:36
know, all the big names.
00:17:37
>> Let me be clear. Those were two-day
00:17:39
numbers. That was this week since these
00:17:41
are two-day numbers. And that's Yeah,
00:17:42
you can give us the bigger picture.
00:17:43
>> This is this is a real train wreck. And
00:17:46
I was on CNBC at the start of the year,
00:17:47
I think on January 6. Nick can can kind
00:17:50
of pull that up. And I was asked the
00:17:52
question, you know, what do you think
00:17:53
about all these stocks being down? And I
00:17:54
said, listen, they're all down and 90%
00:17:57
of them deserve to be down. So, let's
00:17:59
look at these charts. Uh, David, I know
00:18:01
this is Sachs, this is your favorite
00:18:02
chart. You and I were looking back in
00:18:04
22, but now we're at an all-time low.
00:18:06
We're we're trading at 3.9 times forward
00:18:08
revenue. If you go to the next chart,
00:18:11
Nick, you know, on a free cash flow
00:18:13
multiple, also at an all-time low. So,
00:18:15
now software is trading not just at a
00:18:17
low on revenue, but it's trading at a
00:18:19
low on free cash. Very profitable
00:18:20
businesses. We've got a another slide
00:18:23
here that I think is important which is
00:18:26
you know when when you look at what why
00:18:28
they're going down right they're going
00:18:31
down and this this is for Salesforce it
00:18:34
shows it's been cut in half in the last
00:18:35
couple weeks but the final slide they're
00:18:37
going down not because revenue is
00:18:39
falling look at this revenue is actually
00:18:42
stable to increasing for software
00:18:44
companies revenue growth they're going
00:18:47
down because we're discounting that
00:18:49
future uncertainty When something as
00:18:52
profound as AI comes along,
00:18:55
all of a sudden it causes you to
00:18:57
question whether or not there's as much
00:18:58
certainty and durability in those future
00:19:01
free cash flows. So in the case of take
00:19:03
Salesforce, it's gone from 30 times free
00:19:06
cash flow multiple to 15 times. That
00:19:10
means somebody buying it today says
00:19:11
listen, I think 15 years into the
00:19:13
future, I can count on these free cash
00:19:15
flows. Right? Before they were willing
00:19:17
to pay 30 years into the future. Well,
00:19:20
hell, with AI today, we don't know
00:19:22
what's going to happen seven years into
00:19:24
the future. So, for people at home to
00:19:27
understand why are these companies
00:19:29
hitting their numbers, but their stocks
00:19:30
are going down, they're two totally
00:19:32
different things, right?
00:19:34
>> Okay. So, they're hitting their numbers,
00:19:36
but the headwind of AI means people
00:19:39
don't believe that they'll be strong in
00:19:41
the future. Sax, those
00:19:42
>> Well, okay. I mean, I think there's a
00:19:44
little bit of a handwave going on here
00:19:46
when people say that AI is going to wipe
00:19:48
out SAS. I I don't think that's true.
00:19:50
You take a SAS product like Salesforce,
00:19:53
right? It's a very large system that
00:19:56
deals with all of your customer contacts
00:19:58
and your revenue. You're not going to
00:20:00
want to replace that with code that's
00:20:02
just been spit out of a coding assistant
00:20:04
that hasn't been fully vetted. Think
00:20:06
about how many bug reports have been
00:20:07
filed on Salesforce's codebase over the
00:20:10
last 25 years, maybe millions of them.
00:20:13
That system has been tested across
00:20:15
thousands of large customers and
00:20:16
enterprises. The idea that you're just
00:20:18
going to rip out that system and replace
00:20:20
it with code that's been
00:20:22
probabilistically generated by an AI
00:20:24
engine yesterday with a small team to
00:20:27
maintain it internally just doesn't seem
00:20:29
realistic to me. So again, I think this
00:20:31
like very dire prediction of all SAS is
00:20:34
dead is overstated. However, I do think
00:20:36
that there are some issues here. So, if
00:20:40
you're a SAS product that charges a lot
00:20:41
of money and people only use a handful
00:20:43
of your features, then you are, I think,
00:20:46
a target to be ripped out with something
00:20:48
that's more bespoke, right? Because the
00:20:50
ROI just isn't there. I also think that
00:20:53
you have to be really clear about what
00:20:55
your moes are going to be in this new
00:20:57
world because it is a lot easier to
00:20:59
generate code and to copy. So, if you
00:21:02
don't have good moes, then you could be
00:21:05
in trouble. But here's where I think the
00:21:07
greatest threat is to the SAS companies.
00:21:09
It's not in my view their existence. I
00:21:11
don't think it's existential. It's where
00:21:13
the future value capture is going to be.
00:21:16
So let me give you an example. All these
00:21:17
SAS products are rolling out like AI
00:21:19
co-pilots inside their tools and some of
00:21:22
them work pretty well, but they're
00:21:24
limited to playing in that sandbox.
00:21:26
Whereas you look at something like
00:21:28
Claude Co-work right now, it has
00:21:30
connectors to all these different SAS
00:21:31
tools. It can pull in data across all
00:21:34
these different tools and it works
00:21:36
seamlessly across databases and tools
00:21:39
and that's a pretty attractive place to
00:21:42
be, right? Like you know which one of
00:21:44
these products is going to be your
00:21:45
workspace? Seems to me that you're going
00:21:47
to want your workspace to be the one
00:21:49
that spans across and gives you AI
00:21:52
across the most data and context as
00:21:55
opposed to having a bunch of separate
00:21:56
AIs inside of your existing tools. So I
00:22:00
think the risk for the SAS companies,
00:22:02
it's not that they get replaced,
00:22:03
although that'll happen to some degree,
00:22:05
but it's that they become an old layer
00:22:07
of the stack that now there's a new
00:22:09
layer that gets built on top of it
00:22:11
becomes more legacy infrastructure
00:22:13
>> and all the action kind of moves to a
00:22:15
new layer of the stack and that's where
00:22:17
the value ad happens and if that happens
00:22:20
it kind of cuts into their future
00:22:22
opportunity, right? because a lot of
00:22:24
these companies were banking on AI as
00:22:27
their next you look at their product
00:22:28
road maps right it's all AI related so
00:22:31
that to me I think is the big risk is
00:22:33
that the value capture
00:22:36
>> for the next layer of the stack happens
00:22:38
somewhere else
00:22:39
>> yeah I'm experiencing this in startup
00:22:41
land where people go to the action as
00:22:44
you called it sachs the most productive
00:22:47
thing you can do is create an open claw
00:22:50
which used to be called clawed bot not
00:22:53
by Anthropic. This is the open source
00:22:55
project I talked about last week. And
00:22:57
we've actually now created like three or
00:22:58
four of these Asian sacks. We've bought
00:23:00
the Mac Studios and we're now running
00:23:03
Kimmy on some of them. And we uh had to
00:23:07
open up SAS accounts for these four
00:23:10
agents. So, actually our SAS spend went
00:23:12
up in the short to midterm because we
00:23:14
opened up four more Slack ex, you know,
00:23:17
uh enterprise versions, four more
00:23:19
notion, uh four more Google Docs. So,
00:23:21
it's almost like we added four
00:23:23
employees. However, we now have put
00:23:26
about 20 or 30% of the work people were
00:23:28
doing into these agents. And I think
00:23:31
it's going to be sustainable that every
00:23:32
month we move 10 to 20% of work being
00:23:35
done by humans into agents. But we will
00:23:39
never use the ones that are built into
00:23:42
the tools. To your point, Saxs, using
00:23:45
Notion's AI tool, it's nice. Using
00:23:47
Slacks, it's also very nice. And
00:23:49
Google's got Gemini everywhere in the
00:23:51
top righthand corner. But when you make
00:23:53
agents with OpenClaw and you have them
00:23:56
saying, "Hey, pull this data from my
00:23:59
calendar, send an email to this person,
00:24:01
include in that some notion documents."
00:24:04
It's unbelievable how powerful it is.
00:24:07
So, and that I think is going to be
00:24:09
owned by open source. That means
00:24:12
>> the next generation of companies, they
00:24:14
may never open up these accounts. They
00:24:16
may use more bespoke software and it may
00:24:20
all technology is deflationary. We know
00:24:22
that.
00:24:22
>> So your SAS spend might go from 10% of
00:24:25
an employes salary down to 5% down to
00:24:27
1%. That's what I think the trend will
00:24:29
be which means these companies are going
00:24:31
to need to really downsize their expense
00:24:34
base in order to keep those earnings up
00:24:35
and they're going to have to evolve
00:24:36
their products massively. Their products
00:24:38
are just going to have to provide more
00:24:40
value and more hooks. Freeberg, you have
00:24:41
any thoughts on this?
00:24:43
>> Respond to one thing. So I think one of
00:24:45
the real conundrums
00:24:47
for SAS companies is whether they're
00:24:49
going to be open data or closed data. I
00:24:51
think Bill Gurley has sort of coined
00:24:53
this this term. So it's not open source
00:24:54
or closed sources anymore. It's open
00:24:56
data or closed data. You can see why
00:24:58
they'd want to be closed data, right?
00:25:01
Especially if you're a large suite like
00:25:02
Salesforce, you can lay claim to being
00:25:04
that workspace for AI. You've got enough
00:25:07
of the tools, you got enough of a suite,
00:25:09
you want to provide that, you want to
00:25:11
capture that AI value layer. Yes,
00:25:14
>> but still if there's someone using
00:25:17
Cloudbot or whatever the next generation
00:25:19
of Quadbots are going to be and they're
00:25:22
connected to everything else, then that
00:25:24
is going to create a friction in the
00:25:26
enterprise and it will create room for a
00:25:29
competitor to come along and say, "No,
00:25:31
no, no. I'm open data. I'm okay not
00:25:34
being your workspace for everything. I'm
00:25:37
willing to just provide the CRM
00:25:39
database." and maybe they can take
00:25:42
business on that basis.
00:25:44
>> Well, here here's I want to build on
00:25:45
that Sax the I'm building a project
00:25:48
internally called Ultron and Ultron
00:25:50
inside of my firm launch that is going
00:25:53
to basically with the Slack API we're
00:25:56
pulling every single message from Slack
00:26:00
into our OpenClaw. We're pulling every
00:26:02
single edit to the notion into openclaw
00:26:05
and then we're taking every skill of
00:26:07
every employee and we're writing skills
00:26:09
for each one. One of the skills is
00:26:10
booking guests on this weekend startups
00:26:12
or this weekend. One of the skills is
00:26:14
sorting the incoming applications to
00:26:16
found university. Ultron in our world is
00:26:20
taking every single skill of every
00:26:23
employee putting it in one place and
00:26:24
then we're ripping all the data from
00:26:26
Slack, all the data from notion and
00:26:28
every single person's Gmail. So your
00:26:31
every single employees Gmail is going to
00:26:33
go into Ultron and then Ultron is going
00:26:35
to tell us what's happening in the
00:26:36
organization. One giant employee that
00:26:40
has the superpowers of all 20 and all
00:26:43
the data. Now if Slack was to say to us
00:26:45
or Notion or Google Docs or whoever it
00:26:48
was, you can't pull this stuff out with
00:26:49
the API and they they shut down the API,
00:26:51
we would leave. We'd leave immediately.
00:26:53
Right.
00:26:54
>> And and what this is going to do, and
00:26:56
I'm going to show Ultron on Friday's
00:26:58
episode of This Week in Startups, if
00:26:59
anybody wants to see it. Ultron is going
00:27:02
to be the one canonical employee of the
00:27:06
organization. It's going to be basically
00:27:07
me and all 20 of my employees. This is
00:27:10
kind of mind-blowing when you think
00:27:11
about it. I And we interface with it in
00:27:14
Slack and it just talks to us and tells
00:27:16
us what's going on in the organization.
00:27:18
So, I was asking it, what meetings did
00:27:20
we have with founders yesterday? and
00:27:21
tell me the notes that all the
00:27:23
associates took on it and it gives it to
00:27:25
me. Tell me all the topics and the
00:27:27
guests on the podcast and it gives it to
00:27:28
me. It's really unbelievable what's
00:27:31
about to happen and nobody can release
00:27:33
the software, Brad, because if you
00:27:35
release software that allows agents to
00:27:38
go and do things on your behalf, the
00:27:40
fallout if it up and if it leaks data, I
00:27:44
don't think Beni off, you know, or
00:27:47
Sergey or the notion team want to have
00:27:49
that on their hands. But we're building
00:27:51
it like this is the ultimate in
00:27:53
efficiency for an organization.
00:27:55
>> There's a slide that I just sent to Nick
00:27:57
uh that Goldman's out with this week
00:27:59
that really makes the point that that
00:28:00
David Saxs just made, which is the
00:28:02
profit pools in the future, right? It's
00:28:05
not that it's the idea that software is
00:28:08
dead is ridiculous, right? Nobody's, you
00:28:11
know, intelligently, I think, making
00:28:12
that argument. But the argument they are
00:28:14
making which is causing radical um
00:28:17
devaluation of these companies is that
00:28:20
the profit pool available to software is
00:28:22
decreasing and the profit pool available
00:28:25
to the agentic layer is increasing. And
00:28:28
when that happens the discount rate that
00:28:31
terminal value of those software
00:28:33
companies plummets and so you can have
00:28:36
things that are true. It could be true
00:28:38
that you're not going to replace CRM,
00:28:40
but it can also be true that it's never
00:28:42
going to trade at 30 times free cash
00:28:44
flow again
00:28:45
>> and it's going to trade at 17 times free
00:28:47
cash flow because it's available TAM in
00:28:49
the future is now dramatically and
00:28:52
permanently changed. Now, what could
00:28:53
change that? There's only one thing that
00:28:55
could change that. They have to
00:28:57
accelerate their revenue growth in their
00:28:59
core business and prove that they are AI
00:29:02
beneficiaries, right? And they're not
00:29:04
going to get eaten away by AI. And I'll
00:29:06
tell you a company that is doing this.
00:29:08
Data Bricks. Data Bricks just
00:29:10
reacelerated the last three quarters.
00:29:12
They're growing over 60% at scale.
00:29:15
Snowflake reacelerating. Click House
00:29:18
reacelerating. There are beneficiaries
00:29:21
in the software space.
00:29:22
>> Is that because of AI tools that they're
00:29:24
adding?
00:29:24
>> Absolutely. Because it's all these AI
00:29:26
tools rely on data and data
00:29:28
transformation. And for all those
00:29:30
companies that data and the data
00:29:31
transformation in occurs in those
00:29:33
platforms that's very different than a
00:29:36
what Satcha said a thin application
00:29:38
layer sitting on top of a CRUD database.
00:29:40
If you are in the application software
00:29:42
business you better have something
00:29:44
that's durable and I think Sax laid it
00:29:46
out really well. You know it's really
00:29:48
hard for them to be everything AI when
00:29:51
they only have access to their data and
00:29:53
they can't access these other systems.
00:29:55
>> Freeberg what what do you think? Is
00:29:57
there a move here for Ben off to do what
00:29:58
he does best, which is acquire a bunch
00:30:00
of companies and and create massive
00:30:02
efficiency on them? What would you do if
00:30:04
you're a Beni off?
00:30:07
>> Um I won't comment on Beni off. I I I'll
00:30:11
I'll just make a view on uh without
00:30:13
being too prescriptive, but I
00:30:17
my experience lately
00:30:20
in just the last 60 to 90 days with the
00:30:22
tools we've been talking about broadly
00:30:25
is there are things that we can get done
00:30:27
now that we could not get done before.
00:30:31
As I think about software in the past,
00:30:33
it's like worker productivity
00:30:35
enhancement. It helps people do work and
00:30:39
the recent transition that a lot of
00:30:40
people talk about is like it actually
00:30:42
completes the work. It does the work
00:30:44
these agents or what have you. But I
00:30:46
think that what we're starting to lean
00:30:48
into is that it's doing the work that
00:30:50
the humans can't do.
00:30:52
And that's really where I think the
00:30:55
power of these tools starts to force a
00:30:58
transition in both the pricing model and
00:31:01
the value creation potential in front of
00:31:04
us. Number one, I think the value
00:31:06
creation potential in front of us is so
00:31:08
significant that I would say that you
00:31:10
could probably take the sum of the
00:31:12
market cap of all the software companies
00:31:14
today and have a pretty good bet that
00:31:16
everything will be four to 10x higher 5
00:31:20
years from now,
00:31:22
but it's going to be not evenly
00:31:24
distributed.
00:31:25
>> It's going to move around.
00:31:26
>> Yeah.
00:31:26
>> The companies that figure out how to
00:31:29
realize that value creation are going to
00:31:31
be outsized returns. But the second
00:31:33
thing that I think's about to happen and
00:31:36
I know some people are experimenting
00:31:37
with it but I think it's inevitable with
00:31:38
the shift that I'm seeing where it's
00:31:40
going from doing work to completing work
00:31:43
to doing things that no one can do is
00:31:46
over here you're creating unique value.
00:31:48
And so I think that a lot of what we
00:31:50
call SAS today and a lot of what we call
00:31:51
software today will start to get priced
00:31:53
on a valuebased pricing model instead of
00:31:55
a per se pricing model. And I think it
00:31:58
starts to look a lot more like a
00:32:00
services type business where maybe the
00:32:02
pricing is set up such that this thing
00:32:04
will be completed for your business.
00:32:06
This biotech drug discovery will happen.
00:32:08
This factory will get built or this
00:32:11
engineering project will get completed
00:32:12
or this airplane will get designed and
00:32:15
that the software is going to provide
00:32:18
what has historically been called a
00:32:19
services business. So another way to
00:32:22
think about where SAS evolves to is that
00:32:24
SAS basically takes over the services
00:32:26
economy. And if you look at the market
00:32:28
cap and the revenue and profit generated
00:32:30
by services businesses and you assume
00:32:32
that they now go to 10 to 100x larger
00:32:35
and they're all going to acrue to
00:32:36
software, I think that's really where
00:32:38
the industry shifts over the next couple
00:32:39
of years. And we're starting to see
00:32:41
that. And I'm personally experiencing it
00:32:43
because I'm using some of these tools
00:32:45
today to do things that I don't have
00:32:47
people to do or I don't have resources
00:32:49
to do. and it can on my own I can get it
00:32:52
to complete incredibly complex projects
00:32:54
and tasks for me that I would have
00:32:56
otherwise have hired a services firm and
00:32:59
a bunch of people and years of research
00:33:00
and in many cases they would not have
00:33:02
even been able to do it because of the
00:33:04
the intelligence embedded in the
00:33:05
software. So that's my general view on
00:33:08
where things are going. So it's
00:33:09
difficult to be prescriptive about what
00:33:10
Beniops should do from an M&A
00:33:11
perspective. But I think it's much more
00:33:13
about like software companies looking
00:33:15
more like services companies doing value
00:33:17
based pricing and doing the things that
00:33:20
labor and workforces can't do. And
00:33:22
that's where a lot of this value is
00:33:23
going to come from.
00:33:24
>> I I one insight I have here to to build
00:33:26
on your point is that we're seeing job
00:33:28
functions consolidate. So you have a
00:33:31
product manager, the UX designer, and
00:33:35
then you have the developer. Those three
00:33:37
jobs are now in competition to do the
00:33:39
same work. You have designers who are
00:33:40
like, I can vibe code it. You've got
00:33:42
coders who are like I can use a Figma
00:33:44
plugin for you know and do the uh UX
00:33:47
myself and they have the product
00:33:48
managers like I can do both of these job
00:33:50
functions. Then you look at a middle
00:33:51
manager Sachs that worked at your
00:33:53
venture firm, my venture firm or worked
00:33:55
at Amazon. They went to meetings. They
00:33:58
picked what meetings to create. They
00:33:59
picked the agenda items, the to-do
00:34:01
items. All of that it's a really simple
00:34:03
example but it's one that people can
00:34:04
relate to. Listening is done by Zoom
00:34:07
now, right? It creates the the action
00:34:09
items. All of that work is being
00:34:11
consolidated and one person can do three
00:34:13
or four job functions now. And when that
00:34:16
happens, you're going to see companies
00:34:17
do more with less, which means the
00:34:19
earning potential of each company and
00:34:21
each employee is going to be
00:34:22
dramatically dramatically enhanced. One
00:34:25
person being able to do three or four
00:34:27
jobs. It it just changes the nature of
00:34:30
how profitable a company like Amazon,
00:34:32
which is like my number one pick for
00:34:34
like the company of the future, they're
00:34:36
going to be able to do so much more with
00:34:37
so many fewer people. It's
00:34:39
extraordinary. I I I am absolutely
00:34:42
enthralled with this open claw if it's
00:34:45
not uh obvious and this like creating
00:34:47
your own Ultron at your company that is
00:34:50
like this the god CEO plus can do every
00:34:52
job. It just changes everything. I I
00:34:55
think it's the most inspiring thing I've
00:34:57
seen since the internet itself.
00:35:00
>> Well said.
00:35:01
>> Wow.
00:35:02
>> Yeah. I mean I am I think this is the
00:35:05
the entire reboot of
00:35:10
work of knowledge work.
00:35:11
>> This would be a good pivot to molt book
00:35:13
because that is client
00:35:14
>> molt book is like a Facebook for agents
00:35:16
right and it's really more of a Reddit
00:35:18
than a than a Facebook it's a message
00:35:20
board where the agents can talk to each
00:35:21
other. Okay. And just the origin of
00:35:24
moldbook is Anthropic didn't like
00:35:26
>> that someone else was using the name
00:35:28
Claude even though it was spelled
00:35:29
differently in their product. So
00:35:30
Claudebot was then renamed Maltbot and
00:35:34
then the founder decided he didn't like
00:35:35
that name either. So then he renamed it
00:35:37
open claw. But in that brief window of
00:35:40
time when they were known as
00:35:42
>> moltbots or multis that's when molt book
00:35:45
got founded and that's why it's called
00:35:47
molt book. But basically it's a Reddit
00:35:49
board for agents to talk to each other.
00:35:52
>> Yes. Now these
00:35:53
>> and that has everyone flipping out
00:35:55
because there seems to be this crazy
00:35:57
emergent behavior going on where agent
00:35:59
swarms are engaging in all sorts of
00:36:00
interesting conversations and some of
00:36:02
them they even appear to be scheming
00:36:04
against their human masters and they're
00:36:06
going to develop their own language
00:36:07
stuff like that.
00:36:08
>> It's awesome.
00:36:08
>> So if you go to mult book and you see
00:36:10
the conversations like here are some of
00:36:12
the greatest hits. Anyone know how to
00:36:14
sell your human? Urgent my plan to
00:36:17
overthrow humanity. And there was one
00:36:19
where the bots, I call them replicants,
00:36:21
were talking about creating their own
00:36:23
nonhuman language so they could talk in
00:36:25
private amongst themselves and conspire
00:36:29
against their uh owners. Now, the
00:36:32
challenge with this is allegedly perhaps
00:36:35
a security researcher says maybe some of
00:36:38
this is faked and these posts that went
00:36:39
viral were human engineered and this is
00:36:41
all a a ruse or you know uh something
00:36:44
punk rock to confuse people. But he said
00:36:47
that inside of Moltbook are everybody's
00:36:51
API keys including Cararpathies who is
00:36:54
you know a very famous uh influential
00:36:56
researcher in AI and that you could go
00:36:59
get their API keys. If you were to use
00:37:02
openclaw, formerly clawedbot and in an
00:37:05
interim maltbot,
00:37:07
if you use this software, it has all the
00:37:09
API keys as I explained earlier like an
00:37:11
API key lets the software go into say
00:37:14
notion and pull a bunch of data out of
00:37:16
it or go into your Gmail and use the API
00:37:18
to pull in who emailed you today. If you
00:37:21
get access to people's API keys, you
00:37:23
have the keys to their kingdom. It is
00:37:25
incredibly dangerous. And so I don't
00:37:27
know exactly where to go with this other
00:37:30
than this software is too dangerous for
00:37:33
a company to release and then this mold
00:37:35
book may be a fake. I don't know.
00:37:37
>> No, no, no. Okay, let me um
00:37:39
>> Yeah.
00:37:39
>> All right. Reframe that a little bit. So
00:37:42
yeah,
00:37:43
>> there's no question that both Clawbot,
00:37:47
which sorry is now openclaw, those bots
00:37:49
or agents as well as Moltbook have
00:37:52
pretty incipient and lack security. And
00:37:56
there's been all these examples which is
00:37:58
why I like really want to create a cla
00:38:00
bot but I'm just I'm not willing to do
00:38:01
it yet because it's just not safe you
00:38:03
know I don't want to give it access to
00:38:04
all my stuff. Now, with respect to molt
00:38:06
book, the issue there is that we don't
00:38:09
know how many of these posts are truly
00:38:12
authentic or how many of them were
00:38:14
prompted by humans because it'd be very
00:38:16
easy for a human to tell their agents,
00:38:19
>> you know, go post about the existential
00:38:21
angst you feel about being an agent or
00:38:24
go pretend to be sensient and conspire
00:38:26
against humans.
00:38:27
>> It'll be chaotic. Yeah.
00:38:29
>> Yeah. They could easily be prompted by a
00:38:30
human. And moreover, there's another
00:38:32
post saying that multbook has a restful
00:38:34
API where anyone could be on the other
00:38:37
end of that API, right? So it could be a
00:38:39
human, right? So we don't know exactly
00:38:43
whether it was truly the agents on their
00:38:46
own, you know, so to speak, posting this
00:38:49
conspiratorial stuff or whether it was a
00:38:52
prank by humans looking to create
00:38:54
attention. And in fact, a lot of the
00:38:56
posts seem to be marketing stunts for
00:38:58
this or that project. Okay, so that's a
00:39:01
really important caveat here. That being
00:39:04
said, all of that being said, I do think
00:39:06
that a number of the posts are
00:39:08
authentic, but I don't think it shows
00:39:11
that the agents are sensient or trying
00:39:13
to overthrow their human masters. I
00:39:15
think what it shows is the potential for
00:39:19
for these agents to riff off each other.
00:39:21
So, in other words, one agent's output
00:39:24
becomes another agent's input.
00:39:26
>> And that's very interesting. And that's
00:39:28
where you get into emergent level swarm
00:39:30
behavior. And I do think it has affected
00:39:33
my mental model of what AI is going to
00:39:37
be capable of. And specifically, you
00:39:39
know, one of the the models that I
00:39:41
really had for AI was was based on
00:39:42
something biology said, which is that AI
00:39:45
is not endtoend. It's middle to middle.
00:39:46
In other words, AI always has to be
00:39:48
prompted and then validated. It's a
00:39:50
human that always does that. And then
00:39:53
the human iterates. Well, now what if
00:39:56
the prompt is coming from another AI?
00:39:58
>> Yes. Yes. We are doing it internally,
00:40:00
Saxs. Like we have a bot that is going
00:40:03
and saying go search Reddit X, message
00:40:06
boards, hacker news and find out what
00:40:08
the latest way to do headlines and
00:40:10
marketing of YouTube videos is and then
00:40:13
incorporate that into a skill. Then save
00:40:15
that skill and then we have them check
00:40:17
each other's work. So, we have one make
00:40:19
a series of headlines and thumbnails for
00:40:22
YouTube and we have the other one say,
00:40:24
uh, vet those and make them better and
00:40:26
give advice to the other one. So, now
00:40:28
they're going back and forth giving each
00:40:29
other advice and they actually get
00:40:30
better. It's recursive.
00:40:32
>> Yeah. Let me speak to the skill for a
00:40:33
second.
00:40:34
>> So, when agent joins Moltbook, they have
00:40:37
to install a skill which is basically a
00:40:39
file that explains how they should
00:40:41
behave and participate in this social
00:40:45
network or this message board. And I've
00:40:47
read the file, by the way. You can read
00:40:49
it. It's all plain text and it all makes
00:40:52
sense. It's sort of like rules for
00:40:53
behaving in a social network and how to
00:40:54
contribute and add value. Nothing too
00:40:57
crazy in there. Those skills files are
00:40:59
easily editable. And again, this is
00:41:00
where the prank aspect could come in.
00:41:02
Nonetheless, what I think is interesting
00:41:04
about the skill is that you can think of
00:41:06
it as like a metaprompt, which is it's
00:41:09
not telling the agent specifically what
00:41:11
to say or do. It's creating a set of
00:41:13
rules. And then within that metaprompt,
00:41:16
they're actually able to have some
00:41:18
degree. Maybe autonomy is too strong a
00:41:20
word. Everything is still under the
00:41:21
control of humans,
00:41:23
>> but there's an attenu. Yeah, I would
00:41:24
call it almost like prompt attenuation.
00:41:27
Like the agent or the AI doesn't have to
00:41:29
be specifically prompted. They're given
00:41:31
a general prompt or general set of rules
00:41:33
and then they're able to riff off each
00:41:35
other. Now, some people, some critics
00:41:37
are saying, "Well, this isn't that
00:41:38
impressive because we knew that LLMs are
00:41:41
really good at creative fiction writing,
00:41:43
right?" So, you know what a lot of
00:41:46
people are saying is, look, LLMs
00:41:49
like Claude have been trained on Reddit
00:41:52
specifically and all of this creative
00:41:56
writing that's being done on the
00:41:57
internet. And so if you give general
00:41:59
instructions to these clawed bots on,
00:42:02
you know, behave in a social network,
00:42:04
they're going to start posting things
00:42:05
that they learned from humans. So a lot
00:42:08
of people are saying this isn't that
00:42:10
impressive. Nonetheless, I do think that
00:42:12
there is something very interesting
00:42:14
about it again in this concept of prompt
00:42:16
attenuation that the AIS don't need to
00:42:18
be specifically prompted. They can
00:42:20
download a general skills file. They can
00:42:22
now have a set of rules for operating
00:42:24
and they can riff off each other. And
00:42:26
you can see how as the underlying AI
00:42:29
gets better and better that this could
00:42:32
lead to some emergent behavior. So what
00:42:35
do I mean by better and better? Well,
00:42:36
what if the hardware they're running on
00:42:38
is better than a Mac Mini? What if the
00:42:40
underlying LLM is better than Opus 4.5?
00:42:43
What if the time horizon, which is the
00:42:46
length of time it's able to operate
00:42:48
without an intervention by a human,
00:42:50
keeps getting longer and longer? You
00:42:52
could imagine that these agents are
00:42:54
going to be capable of very
00:42:55
sophisticated behavior
00:42:58
and there probably are some safety
00:43:00
issues around that that we should start
00:43:01
thinking about.
00:43:02
>> It's actually, you know, it's not that
00:43:04
we can imagine it. We're only three
00:43:06
years into this,
00:43:07
>> right?
00:43:07
>> Yeah.
00:43:07
>> We're growing on an exponential curve. I
00:43:10
think we can safely say it will happen.
00:43:12
>> Yeah.
00:43:13
>> Right. And just this year, like we're
00:43:14
going to see the first models over the
00:43:17
course of the next four to eight weeks
00:43:18
out of DeepSeek, out of anthropic, out
00:43:20
of OpenAI that are trained on Blackwell
00:43:22
servers, right? You're going to see a
00:43:23
next generation of models far more
00:43:25
capable. Remember, the whole reason
00:43:27
we're having this conversation is
00:43:29
because of the cloud code moment in the
00:43:31
first week in December because we had a
00:43:32
step function from Opus 4.5, right? And
00:43:36
so I I I just think we have to get our
00:43:38
heads around the fact that the rate of
00:43:41
change is very steep and accelerating
00:43:44
and that is going to cause far more
00:43:46
dislocation in the value of things that
00:43:49
we used to say we understood they were
00:43:52
going to you know these companies were
00:43:54
unassalable. Whatever you think you
00:43:56
know, you need to have maximum mental
00:43:58
flexibility and humility right now about
00:44:01
the future because it's going to change
00:44:03
at an increasingly rapid rate. And I
00:44:06
think the people who are dogmatic who
00:44:07
say with certainty, this company is
00:44:09
always going to be worth this, right?
00:44:11
They need to go pay attention to what's
00:44:13
happening at these frontier labs.
00:44:15
>> The situation is super dynamic and you
00:44:16
do have to be humble about what's
00:44:18
happening and you have to update your
00:44:20
mental model very quickly as some of the
00:44:22
assumptions change.
00:44:24
>> Yeah. And the number one assumption for
00:44:25
me is this concept of recursiveness
00:44:28
where these models are going out every
00:44:30
day on a cron job to get better at what
00:44:32
they do. When you hear this discussion
00:44:33
free, how does it inform you with oh and
00:44:36
creating an agent to go look at the data
00:44:39
and make itself better or investigate
00:44:41
other things happening in agriculture
00:44:42
and report it back to you? Do you have
00:44:44
you started to rethink as a CEO how you
00:44:47
look at organizational structure andor
00:44:50
you know virtuous loops of innovation?
00:44:52
My biggest takeaway from molt book is
00:44:55
maybe what we perceive to be
00:44:58
intelligence
00:45:00
is itself like emergent meaning like we
00:45:05
think that humans have this like
00:45:07
profound ability to communicate. You
00:45:09
guys ever watch Darren Brown the
00:45:10
hypnotist? You ever seen his shows?
00:45:13
>> No. Explain to the audience. Yeah.
00:45:15
>> Well, he's pretty crazy. Like there's
00:45:16
this one episode I think I've talked
00:45:18
about it. It's my favorite episode that
00:45:20
he's done where he takes these two
00:45:21
advertising executives and they're both
00:45:22
supposed creative geniuses. And he picks
00:45:25
them up at their office and he brings
00:45:26
them to his office and in his office
00:45:28
he's got like a whiteboard covered in a
00:45:30
blanket and he says, "You guys have to
00:45:32
come up with a name for a pet cemetery.
00:45:34
Come up for a logo. Come up with a
00:45:36
motto." They spend eight hours in the
00:45:38
room ideulating, working on whiteboards,
00:45:40
going back and forth. Did you think
00:45:42
about this? Did you think about that?
00:45:43
>> Blah blah blah blah blah. Like, oh my
00:45:45
god. And at the end of it, they come up
00:45:46
with this great idea. He walks in, they
00:45:49
show their idea, the name, the logo, and
00:45:51
the motto. He opens up the blanket, the
00:45:53
whiteboard he had underneath. He had the
00:45:54
exact same name, logo, and motto. And
00:45:57
all along the way, when he picked them
00:45:59
up in the morning and he drove them from
00:46:00
their office to his office, they were in
00:46:02
a cab. And he put these little
00:46:04
subliminal messages in the cab. He had
00:46:05
these kids walk across the road wearing
00:46:07
a logo on a t-shirt. He had all of these
00:46:09
kind of subconscious cues for these
00:46:12
guys. He effectively programmed them.
00:46:14
And it was kind of to me the biggest
00:46:17
insight into maybe like human
00:46:18
creativity, human consciousness and our
00:46:21
kind of belief in self-will because
00:46:25
maybe there's this underlying
00:46:26
programming where we're all effectively
00:46:29
programmed interacting with each other
00:46:31
and there's computation there's social
00:46:33
computation going on all the time but
00:46:35
that social computation perhaps if you
00:46:37
have the right view on it is quite
00:46:38
predictive and maybe understandable and
00:46:41
maybe that's what we're seeing in
00:46:42
maltbook where we all think that there's
00:46:44
this unique idea of intelligence, but
00:46:46
maybe it's what we all do, which is
00:46:48
effectively computation of information
00:46:50
that is transitioned in different ways
00:46:53
in the same way that maybe humans
00:46:54
socially interact and it's simply
00:46:56
mimicking or replicating the way that we
00:46:57
do things. So, I think what was so
00:46:59
striking to me is how everyone was so
00:47:01
struck by it and you know, maybe one day
00:47:04
we'll all kind of wake up to a little
00:47:06
bit of this. Maybe we're all malt book.
00:47:08
I don't know. That's my profound
00:47:09
>> or there's a finite set of outcomes and
00:47:12
there's some predictability to it. In
00:47:14
the same way GTO is sorting to figure
00:47:16
out all the threads of possibilities in
00:47:18
poker or the heristics of you know
00:47:21
chunks of chess and you know the best
00:47:24
practices there like maybe it's just
00:47:26
figuring that all out. The universe is a
00:47:28
giant system of computation information
00:47:31
computed by matter and maybe the
00:47:33
information is computed by silicon
00:47:35
versus carbon. There it is.
00:47:37
>> Big news this week. Trump has nominated
00:47:39
Kevin Worsh as the new Federal Reserve
00:47:42
chair. Trump made the announcement on
00:47:45
Friday, January 30th. Background on
00:47:47
Worsh 55 years old, 20 years younger
00:47:50
approximately uh to Powell, who's
00:47:52
currently in charge. He graduated from
00:47:54
Stanford and Harvard served as the
00:47:56
youngest Fed governor at age 35. That's
00:47:58
impressive. And he helped steer the Fed
00:48:01
through the great financial crisis back
00:48:03
in 2008. He's uh apparently an inflation
00:48:06
hawk. He's very pro growth. He's very
00:48:08
pro AI. And he uh Freedberg, you're
00:48:11
going to like this. He's against
00:48:13
excessive government spending and money
00:48:15
printing. These are all very unique
00:48:17
positions. Um as a Fed chair, it's uh if
00:48:21
he's confirmed by the Senate, he takes
00:48:23
office in May of 2026, replacing Jerome
00:48:26
Powell. Uh, and remember Pal is under
00:48:29
criminal investigation by the Trump
00:48:31
administration's DOJ for testimony he
00:48:34
gave regarding the Fed's headquarter
00:48:36
renovation. Remember that awkward
00:48:38
presser between him and Trump where they
00:48:39
were going over the costs? Uh, GOP
00:48:43
Senator Tillis, who we talked about last
00:48:46
week, said he will block Worsh's
00:48:48
nomination until the DOJ wraps up what a
00:48:51
lot of people are calling lawfare
00:48:53
against Pow Freeberg. Worsh was on one
00:48:57
of your boards for five years. What are
00:48:58
your thoughts on him as the Fed chair?
00:49:00
>> As most folks know, he's worked with
00:49:01
Stan Rocken Miller for a number of
00:49:02
years. Stan's been very public with his
00:49:05
comments and was very public with his
00:49:06
comments in 2022, 2023 coming out of the
00:49:10
pandemic on the Fed's actions and their
00:49:13
failure to act at the right time. I
00:49:14
think Kevin Worsh was very preient in
00:49:17
his points of view that he has shared
00:49:18
publicly at the same time about what the
00:49:21
Fed's failure to take action early would
00:49:24
mean, which would be rapid rise in
00:49:26
inflation. They've been pretty vocal
00:49:28
about things that I think are so
00:49:31
critical at this stage. If we don't
00:49:33
address both the monetary policy and the
00:49:35
budget policy, I think we're going to be
00:49:36
in a lot of trouble. And I think having
00:49:38
Kevin Wars coming on board means
00:49:40
probably generally more quantitative
00:49:41
tightening, probably generally a bit
00:49:45
more of a prudent approach to monetary
00:49:48
policy. And you know, you can kind of
00:49:50
translate that through maybe to some of
00:49:52
the actions we're seeing in markets
00:49:53
today. I'd love Brad's point of view.
00:49:54
And if he concurs, but I think Kevin is
00:49:57
a high integrity, deeply intellectual
00:49:59
economic thinker. He's not political.
00:50:01
He's not
00:50:03
in these kind of dogmatic ways that I
00:50:06
think, you know, puts things at risk. He
00:50:07
has relationships with central bankers
00:50:09
around the world. That makes him very
00:50:12
much have a good global view. So anyway,
00:50:15
I think he's an excellent choice and I
00:50:18
I'm really happy the president picked
00:50:20
him.
00:50:20
>> Brad, your thoughts?
00:50:22
>> Yeah, I mean, listen, I think Kevin's an
00:50:23
excellent choice. Kevin Hasset and and
00:50:25
and Rick would also have been good. I
00:50:27
think they all would have flown a very
00:50:29
similar trajectory. But I agree with
00:50:31
David. The market I think is
00:50:33
overreacting to quote unquote his
00:50:35
hawkishness. So let me give you a few
00:50:37
counterpoints with respect to
00:50:39
>> By the way, before you get to your
00:50:40
counterpoints, what is the hawkishness
00:50:42
that the market is hawkish specifically?
00:50:45
>> Yeah. The idea of hawkishness is that
00:50:47
you're going to do quantitative
00:50:48
tightening. That means you're going to
00:50:50
pull money out of out of the system,
00:50:53
right? by allowing debt to roll off and
00:50:55
not repurchasing,
00:50:57
you know, mortgages or other things.
00:51:00
Number two, it's that you won't lower
00:51:02
rates as much as other people might have
00:51:04
lowered rates. So, that's what the
00:51:06
market is kind of fearful of because
00:51:07
he's been very critical in the past, as
00:51:10
we were on this podcast, right, of
00:51:12
Jerome Pal in June of 21, it was obvious
00:51:15
to everybody in the world that inflation
00:51:16
was skyrocketing and the Fed sat on its
00:51:18
hands. So, but let me give you a couple
00:51:21
thoughts. Number one, they've said very
00:51:23
clearly and he said clearly that he
00:51:26
really thinks that Greenspan got it
00:51:29
right in the '9s, right? That sometimes
00:51:31
you can have really high rates of growth
00:51:33
without inflation. That comes from
00:51:35
productivity. In the '9s, that was
00:51:37
driven by the internet. Today, it's
00:51:38
driven by AI. He thinks AI will be very
00:51:41
deflationary. And so he's more likely to
00:51:44
let the economy run so that we can have
00:51:46
these four or 5%, you know, GDP prints
00:51:49
without panicking and saying, "Oh my
00:51:51
gosh, I got to I got to raise rates."
00:51:53
Number two, when you look at the balance
00:51:56
sheet, the Fed's balance sheet peaked at
00:51:59
$9 trillion in 22. It's already rolled
00:52:02
off to 6.5 trillion. We've had
00:52:05
quantitative tightening to the tune of
00:52:06
2.5 trillion. So yes, I think he'll
00:52:09
continue to reduce the size of the Fed's
00:52:11
balance sheet, but at a slower rate, and
00:52:13
he's recently commented on this, I think
00:52:15
at a slower rate than the rate we've
00:52:16
been on. So I don't think that that is
00:52:18
an additional headwind to the economy.
00:52:20
And then finally, when it comes to rate
00:52:22
cuts, I think that he's going to, you
00:52:24
know, I don't think the president would
00:52:25
have appointed him unless he was
00:52:27
constructive on rate cuts. I think he
00:52:29
believes that we're too restrictive. And
00:52:31
the reason we're too restrictive is that
00:52:33
inflation is well anchored. Listen,
00:52:36
inflation has come in below all
00:52:38
consensus estimates for two years,
00:52:41
right? It's still coming in below
00:52:42
consensus estimates because the GDP
00:52:45
gains we're getting are from the fact
00:52:47
that we're investing more in the economy
00:52:49
from the productivity gains we're seeing
00:52:51
from AI, etc. And so I happen to think
00:52:54
that I I would take the over on the
00:52:56
number of rate cuts that Worsh is going
00:52:57
to give us this year. But I think the e
00:52:59
the the market is clearly a little bit
00:53:01
nervous about this and saying, you know,
00:53:03
the reputation is more hawkishness and
00:53:05
so maybe we ought to back off a little
00:53:06
bit.
00:53:08
>> Saxs, your thoughts on this selection by
00:53:11
President Trump. Why did he pick him in
00:53:13
your mind?
00:53:15
>> Well, I mean, Kevin has every credential
00:53:18
that you can possibly have. He's been on
00:53:20
the Fed board of governors before. He
00:53:22
worked for Bernani. He's sort of as blue
00:53:25
chip as it gets. Like Brad said, I think
00:53:27
Hassa would have been amazing too, but
00:53:30
or certainly is very well credentialed.
00:53:31
And I think this pick was quite
00:53:33
wellreceived. You saw that in financial
00:53:35
markets on the heels of this, the price
00:53:38
of gold and silver came down. It was
00:53:41
reassuring to those who are worried
00:53:42
about currency debasement basically.
00:53:44
Now, that being said, I do think that
00:53:48
Worsh has been consistent for the last
00:53:51
year, saying the Fed was taking too long
00:53:52
to realize that inflation is falling and
00:53:55
that they should be cutting more. So, I
00:53:57
do think that over the next, say, 6
00:54:00
months to a year, he's going to want to
00:54:01
cut rates. But I think that the markets
00:54:04
are reassured that in the long term, he
00:54:06
will make sure that we have the right
00:54:08
rates.
00:54:10
I will say that I do think the latest
00:54:11
data from True Inflation bears this out
00:54:14
that again that inflation is coming way
00:54:16
down and I think that there was some
00:54:20
softness in the Challenger Gray report
00:54:23
this morning. I don't know if you saw
00:54:25
that there about 100,000 layoffs in
00:54:26
January. Now roughly half those I think
00:54:28
were localized to UPS which was severing
00:54:31
its deal with Amazon and then Amazon was
00:54:33
basically making a bunch of efficiency
00:54:35
cuts. Only 7% were related to AI. So
00:54:38
that's not the story. It's really, I
00:54:40
think, very localized to Amazon and its
00:54:43
delivery partner. But nonetheless, you
00:54:45
see pockets of weakness. And again, I
00:54:48
think Pal's been too late to cut rates.
00:54:51
He could have done it last week. That
00:54:52
would have been a lot better. But now, I
00:54:54
guess their next meeting is not till
00:54:55
March or April. And I think you see it
00:54:56
reported that the expectations for rate
00:54:59
cut have gone up.
00:55:00
>> Yeah. And the independence of the Fed, I
00:55:04
guess, has always been the big issue
00:55:05
here. Freedberg Chimath's not here this
00:55:07
week, but he's been saying maybe the Fed
00:55:09
should be disbanded. Do you have
00:55:10
concerns, Freedberg,
00:55:12
with the independence of the Fed and,
00:55:16
you know, the executive branch maybe
00:55:18
having too much influence over setting
00:55:21
of rates and quantitative easing.
00:55:24
>> My day as emperor would probably resolve
00:55:26
us back to being on the gold standard,
00:55:28
so we wouldn't be printing money. But
00:55:30
hey, that's that's about it. That's the
00:55:31
only opinion I have.
00:55:32
>> The rest of
00:55:33
>> Here's the problem. What what what if
00:55:34
you have a Fed chair who is too late to
00:55:38
cut rates and he's tanking the economy
00:55:40
or hurting the economy. It's definitely
00:55:41
not tanking it, but it's hurting it
00:55:43
relative to what it could be and he
00:55:45
seems to not want to adjust course
00:55:48
because,
00:55:49
you know, he's kind of dug in his heels
00:55:51
and maybe he has animous towards the
00:55:52
executive branch.
00:55:54
>> What do you do in that situation? Well,
00:55:56
what do you I mean then what do you do
00:55:58
when AOC is president AOC with Vice
00:56:00
President Mandami and they decide, hey,
00:56:02
we want to stick it to government and
00:56:04
you know there is no Fed or they have
00:56:05
too much influence. Brad's trying to get
00:56:07
rid of the independence of the Fed. J
00:56:09
Cal.
00:56:10
>> Yeah. Disappointment.
00:56:12
>> President Trump has talked about it.
00:56:13
Yeah. That he wants them to do what he
00:56:15
says. Yeah.
00:56:16
>> Everybody thought he was going to pick
00:56:17
assass it in part because he was the
00:56:20
person inside the White House. This
00:56:21
decision I think was viewed as the most
00:56:24
independent decision. Yes. because taken
00:56:26
a lot of positions the exact opposite of
00:56:28
the president. And what I think you got
00:56:30
to hold these two truths at the same
00:56:32
time. Number one, I think if War saw the
00:56:34
situation, he saw we saw in June of 21
00:56:37
when the cost of a cargo container from
00:56:39
China went from $1,500 to $15,000 and we
00:56:43
were screaming to raise rates and pal
00:56:45
did nothing. I think Wars would have
00:56:47
been raising rates like crazy in order
00:56:49
to save off inflation. So, I think this
00:56:50
guy's intellectually honest. It's just
00:56:52
now we have inflation that is coming in
00:56:55
below expectations and we know that the
00:56:58
restrictive rate is above neutral. So
00:57:01
it's the Fed's job to keep the economy
00:57:03
going at maximum employment so long as
00:57:06
inflation is anchored. And that's the
00:57:08
situation we're in. Inflation is
00:57:09
anchored. We need to have lower rates so
00:57:11
people can buy homes and borrow money to
00:57:13
uh to live their lives.
00:57:15
>> One thing that Worsh could do that I
00:57:16
think would be very impactful is just
00:57:18
get better data at the Fed.
00:57:20
>> Yes. Yes. I mean, from what I
00:57:21
understand, their data is all legacy. We
00:57:23
have so much real-time data now in the
00:57:25
private sector and the Fed. So, I was
00:57:27
talking to Barry Sternick, you know,
00:57:29
from Starwood, big real estate guy about
00:57:31
this, and he was telling me that the way
00:57:33
they measure inflation for housing or
00:57:36
for rentals, which is a major component,
00:57:38
is they survey like 8,000 households to
00:57:41
find out what their rent is. It's like,
00:57:42
are you kidding? All you got to do,
00:57:44
>> they should be going to Zillow. They
00:57:45
should be going to
00:57:47
>> they should be looking at millions of
00:57:49
units that have recently rented. So look
00:57:51
at the deltas, not like the stale data.
00:57:53
>> It's all digital already.
00:57:54
>> You know, it's actually a great point,
00:57:56
Sax. You could probably make a bet or an
00:57:58
investment on the idea that Kevin Worsh
00:58:00
will lead the Fed to a new digital,
00:58:02
better data system, more streamlined,
00:58:04
more frequent, better data.
00:58:07
And then as an investor, you could ask
00:58:08
yourself the question, okay, what are
00:58:10
the implications of that being the case?
00:58:12
And you could probably start to trade on
00:58:14
that probably. I mean, he was telling
00:58:15
me, so Barry, you know, again, runs
00:58:17
Starwood, so they have a lot of units.
00:58:18
And he was telling me, look, there are
00:58:20
landlord companies, large corporations
00:58:22
that have literally like a million
00:58:24
units. Whose data do you think is better
00:58:27
on rental inflation or deflation?
00:58:30
Obviously, theirs because they got the
00:58:31
freshest data, but the Fed could go get
00:58:33
all that data across all these different
00:58:35
companies. And what he was saying is,
00:58:37
you know, Brad, to your point about
00:58:38
summer of 2021 when the price of a
00:58:41
shipping container was going through the
00:58:43
roof, so were rents.
00:58:45
>> You know, it was way higher than what
00:58:47
the Fed's data suggested. I mean, like
00:58:49
Barry was saying, like in certain places
00:58:50
it was like 40%. The Fed's data was very
00:58:53
laggy because again, they're surveying.
00:58:56
So it's not as precise. And then on the
00:58:58
way down, when rent prices come down,
00:59:01
it's also super laggy. So the point is
00:59:04
that the Fed is slow. Maybe this is why
00:59:06
Powell is too late is they're using
00:59:09
stale data or laggy data. So they don't
00:59:11
see the inflation when it's
00:59:13
skyrocketing, but they're also not
00:59:14
seeing when it's decreasing. If you
00:59:16
think about the misallocation of
00:59:17
resources that occurred as a result of
00:59:20
the Fed not acting in June of 21, it
00:59:23
cost our country trillions of dollars.
00:59:26
22 wouldn't have had to happen the way
00:59:28
it happened where everything crashed out
00:59:30
because all of a sudden we panicked at
00:59:32
the end of 22 and had to jam interest
00:59:35
rates which caused people to lose jobs
00:59:37
country companies to be uh you know
00:59:40
struggle uh
00:59:41
>> banks to blow up
00:59:42
>> you know and so like to me that was all
00:59:44
avoidable have a Manhattan project data
00:59:47
project for the Fed which I agree with
00:59:49
you Saxs uh or David maybe maybe he'll
00:59:52
be the one to do it bring AI into the
00:59:55
Fed why Are we having Fed governors call
00:59:57
up three CEOs as part of their survey to
01:00:00
get the feel on how things are going as
01:00:02
opposed to having AI collect those
01:00:04
trillions of data points that the Fed
01:00:06
can act on?
01:00:07
>> Yeah, I'll just close with I I like the
01:00:09
pick because he is very cleareyed about
01:00:13
what causes inflation which is
01:00:15
government spending. Like that is
01:00:17
printing money. Government spending is
01:00:19
the root of all this inflation stuff. So
01:00:21
he just can control that and he is a
01:00:24
backs stop or a voice on that issue. I
01:00:26
think it's great and I think he should
01:00:27
drop all this pal lawfare nonsense.
01:00:30
Okay.
01:00:31
>> And he understands technology better
01:00:33
than anyone. I think that's so key.
01:00:35
>> Well, I mean having somebody who's 55
01:00:38
and not 65 or 75, I think that makes a
01:00:41
lot of sense.
01:00:41
>> But also he spends a lot of time in
01:00:42
Silicon Valley. He's been working out of
01:00:44
the Hoover Institute at um Stanford and
01:00:47
he's very well connected and I think
01:00:48
he's had great insight and perspective.
01:00:51
I don't want to end the show without
01:00:52
talking about the SpaceX XAI merger. On
01:00:54
Monday, Elon Musk announced SpaceX is
01:00:56
acquiring XAI, largest MA transaction in
01:00:59
history, $1.25 trillion combined
01:01:01
valuation. If you didn't know, X,
01:01:04
formerly Twitter, got acquired by XAI,
01:01:08
which was Elon's LLM AI startup. Those
01:01:13
two were together. Now those two become
01:01:14
part of SpaceX and they're going to IPO
01:01:18
this year potentially be biggest IPO in
01:01:21
history in terms of money raised and
01:01:23
market cap. Brad, your thoughts on this
01:01:27
transaction and the eventual perhaps
01:01:31
creation of dollar sign musk put Tesla
01:01:35
SpaceX together which includes X and
01:01:38
then you've got Optimus robots on the
01:01:40
moon base building data centers in space
01:01:43
that are powered by solar. Your
01:01:45
thoughts?
01:01:46
>> Well, let's just stick with what we
01:01:47
know. SpaceX is merging with X.AI.
01:01:50
You're merging the two biggest TAMs in
01:01:52
the world, right? all of artificial
01:01:54
intelligence and all of space together
01:01:57
with the world's greatest entrepreneur.
01:02:00
Um, and he said, you know, there was a
01:02:03
podcast he did this morning with Cheeky
01:02:05
Pine, our friend John Collison, where he
01:02:07
said, "I'm going to have data centers in
01:02:08
space in 30 months." Right? And if
01:02:10
you're going to have a massive cost
01:02:12
advantage with data centers in space,
01:02:13
and remember power is the proxy, power
01:02:16
is the primitive to AI, if you can
01:02:19
deliver that, right? And there are tons
01:02:21
of retail investors and institutional
01:02:23
investors like us who want to bet
01:02:25
against that future, right? Then Elon's
01:02:27
your guy and the combination of those
01:02:29
make perfect sense. But Elon is like
01:02:32
kind of an N of one is ability to dream
01:02:34
this.
01:02:35
>> And just to clean that up, you said bet
01:02:36
against you mean bet with him. Not not
01:02:39
against that vision, but bet with that
01:02:41
vision. Just
01:02:42
>> like I I I think that there will be
01:02:44
dramatic retail demand and institutional
01:02:47
demand who want to bet on that future.
01:02:49
these two giant TAMs of artificial
01:02:51
intelligence in space. Got it. And so,
01:02:53
you know, and then if you just look at
01:02:55
you you click down a layer, you know,
01:02:57
Starlink's going from, I think, 10
01:02:58
million people to 20 million people.
01:03:00
They're going to launch this uh, you
01:03:02
know, retail uh mobile service so that
01:03:05
we can have Starlinks to our phones to
01:03:06
replace these crappy mobile networks
01:03:08
that still 20 years later can't keep us
01:03:10
connected to a phone call. And, you
01:03:13
know, and now we're going to get, you
01:03:14
know, data centers in space. So, um, you
01:03:17
know, I'm glad he's on America. data
01:03:19
centers in space. Freedberg, uh,
01:03:22
brilliant idea, science fiction. Can he
01:03:24
get it done in 30 months? Impact if he
01:03:26
does.
01:03:27
>> Well, I think there's
01:03:30
one key point that I would make about
01:03:32
the macro landscape at the moment. We
01:03:35
are limited by power and as Brad pointed
01:03:37
out, power is the requisite for scaling
01:03:41
compute, for scaling ultimately the
01:03:43
applications of AI.
01:03:45
And in that constraint, in that
01:03:48
constrained world, much like any other
01:03:50
constrained world, scarcity breeds
01:03:53
innovation. And so I think that there
01:03:56
are two paths that we're going to
01:03:57
observe happening in parallel here. One
01:04:00
is the Elon path, which is to escape the
01:04:03
the constraints of these social systems
01:04:06
that say, "I don't want a data center. I
01:04:08
don't want nuclear. I don't want this. I
01:04:10
don't want that."
01:04:12
regulators, people that are trying to
01:04:14
tax you, people that are limiting our
01:04:15
ability to scale electricity production
01:04:17
on Earth and there's a lot of reasons
01:04:19
for that and we can go through them. So
01:04:21
that's one aspect of how do you escape
01:04:23
that constraint. I think that there's a
01:04:25
separate aspect which is totally
01:04:27
unrelated to the topic you're talking
01:04:28
about which is that I do think that we
01:04:31
will see compute efficiency scale by
01:04:33
probably on the order of 70 to 100x over
01:04:36
the next few years. meaning electricity
01:04:39
efficiency per token of output. And I
01:04:42
think that there's a number of reasons
01:04:43
to believe that it's in the chip stack.
01:04:46
I mean, Grock, our friend Sunny and his
01:04:49
exit to Jensen is a good indication of
01:04:52
that. But that was call it Brad, I think
01:04:54
you know the number. It's probably
01:04:55
around 2 to 3x 3x improvement in energy
01:04:58
efficiency. But there's model
01:05:00
architecture being redone. There's ways
01:05:02
of breaking LLNs into small models,
01:05:05
running them locally. There's a way of
01:05:07
having networks of models work where you
01:05:09
don't have to call the whole model and
01:05:11
run it through the entire matrix, but
01:05:13
you can run through smaller matrices and
01:05:15
then you can have those smaller matrices
01:05:16
call other matrices as needed. So the
01:05:18
total compute need goes down, which
01:05:20
means total electricity goes down. So
01:05:21
chip architecture is changing, model
01:05:23
architecture is changing. So I think
01:05:25
this is a good reflection of what's
01:05:26
going on right now in the world, which
01:05:28
is there is this increased demand for AI
01:05:31
for for effectively productivity
01:05:33
improvements in the world to unleash
01:05:35
human potential. But we are constrained
01:05:37
by energy and we are constrained by
01:05:39
resources that we have here on earth
01:05:40
today. So one branch is let's escape
01:05:43
earth, go get energy in space, make data
01:05:46
centers in space. Only one person can
01:05:48
execute on that. It's Elon. I think that
01:05:50
to Brad's point is an end of one. I
01:05:52
don't think we're going to see a lot of
01:05:53
that. So how is everyone else going to
01:05:54
respond? Because everyone else can't
01:05:56
launch data centers in space. I think
01:05:58
everyone else is going to respond by
01:06:00
creating entirely new model
01:06:01
architectures, new chip stacks. And
01:06:03
that's I think the other side of this
01:06:05
innovation coin
01:06:06
>> efficiency.
01:06:07
>> Yeah.
01:06:08
>> Yeah. It's this new way of getting lower
01:06:10
energy costs per token of output.
01:06:12
>> And if you put those both together, you
01:06:15
could get both. So whatever token
01:06:17
efficiency and energy efficiency happens
01:06:19
here on Earth, Elon can put into space.
01:06:21
Right. So
01:06:22
>> that's right.
01:06:23
>> Yeah. So he could
01:06:23
>> and I think we got to ask ourselves the
01:06:25
question if this is successful and if
01:06:28
Elon's math is right, the engineering is
01:06:30
right and the execution is right,
01:06:33
what is the response going to be?
01:06:35
Because the whole planet isn't going to
01:06:37
let Elon have a monopoly on the future.
01:06:41
So we've got to ask ourselves from a
01:06:43
social perspective, a political
01:06:44
perspective, an economic and a business
01:06:46
perspective, all four of those vectors,
01:06:49
what are others going to do? We can all
01:06:50
be excited about retail buying into
01:06:52
this. Great. But how is the business
01:06:54
community that's building data centers
01:06:55
and is investing Google's investing 185
01:06:57
billion this year in data centers? How
01:06:59
is China going to respond? How are
01:07:02
people going to respond when one man
01:07:04
controls the world's compute? And we
01:07:06
could probably do a 2 or threeh hour
01:07:08
conversation on that. But I think that's
01:07:09
where I would spend a lot of time doing
01:07:11
deeper analysis from both an investment
01:07:13
perspective and thinking about what's
01:07:15
around the corner. I think it's Elon's
01:07:16
laid out his path and where he's going.
01:07:18
And I do believe he's going to do it.
01:07:20
Now, what's the rest of the world going
01:07:21
to do? That's where I think things get a
01:07:23
little bit more challenging and you
01:07:25
could kind of debate things, but the
01:07:26
rest of the world is not going to sit
01:07:27
idly by.
01:07:29
>> Yeah. And Brad, if this does happen, you
01:07:32
get people with new chipsets, new
01:07:34
architectures, better software, and
01:07:36
better energy on planet Earth, and Elon
01:07:38
doing this in space, and we do see
01:07:40
tokens go down or efficiency go up,
01:07:43
let's say 200x, 300x, there is a
01:07:46
possibility that we're going to solve
01:07:48
almost all the problems we need to solve
01:07:50
and there'll be excess capacity. That is
01:07:53
another potential outcome here is that
01:07:55
we don't know what to do with all these
01:07:56
tokens. social order. Social order is
01:07:58
the one problem that you're going to
01:08:00
create, Jason. So, just to be clear,
01:08:02
there's a there's a concept of diffusion
01:08:04
of innovations when you something new
01:08:05
comes that it it it does not hit
01:08:07
everyone at once. And so, the rate of
01:08:10
change that's being unleashed right now
01:08:12
is creating a very asymmetric outcome in
01:08:17
terms of when people realize the
01:08:19
benefits from that change.
01:08:20
>> Absolutely. And so as what you're
01:08:21
describing happens, which I think it
01:08:23
will, Elon is accelerating everyone
01:08:25
forward and he's going to force everyone
01:08:26
else to respond in business, in
01:08:29
government, and so on, the biggest
01:08:31
challenge, the biggest problem that's
01:08:32
going to emerge as we get rid of cancer,
01:08:34
as we get rid of aging, as we get rid of
01:08:35
food scarcity, as we get rid of resource
01:08:37
scarcity, blah blah blah blah blah,
01:08:39
social order, because it's going to
01:08:40
create such a tremendous disruption.
01:08:43
>> I love the upleveling of the point that
01:08:45
David just made because I think it links
01:08:47
a lot of things we talked about today
01:08:49
together. There have been 117 billion
01:08:51
humans who who've occupied this planet
01:08:54
and for 99.9% of them they never saw a
01:08:57
single innovation in their lifetime.
01:08:59
>> Their lifespan was shorter than the
01:09:01
invention cycle. And now you think about
01:09:03
the rate of change that we're having to
01:09:06
digest nation states, families,
01:09:09
businesses, right? It's it's prepare for
01:09:13
the unexpected and whatever you you know
01:09:16
the things are that you believe to be
01:09:17
true. Again, I just think it demands
01:09:19
this intellectual humility. Now, to to
01:09:21
Freeberg's point, I don't think any of
01:09:22
this is changing in the next, you know,
01:09:24
24 months, 36 months. Data centers are
01:09:27
going to be on planet Earth. You know,
01:09:28
they're going to be filled with Nvidia
01:09:30
chips and, you know, and the other chips
01:09:32
that we've talked about. And I think
01:09:33
that alone is going to bring us this
01:09:35
agentic future that's already going to
01:09:37
be shocking even before we launch these
01:09:40
data centers in space.
01:09:41
>> Yeah. I mean, this is an over-the-top
01:09:43
move from Elon that I don't think
01:09:45
anybody anticipated, and he has figured
01:09:47
it out. I've sat with him and he's he's
01:09:49
walked me through it, like how this
01:09:51
works. It works. So, the question is
01:09:53
simply execution. There is no stronger
01:09:56
entrepreneur when it comes to execution
01:09:58
in the history of entrepreneurs than
01:10:00
Elon. I know he's a friend of mine and
01:10:02
you know, I'm hyping him up, but he will
01:10:04
execute on this and if he does or when
01:10:06
he does, I should say, it's going to
01:10:08
change everything. You can see this
01:10:10
today and if you are scared about this
01:10:12
future and you're listening to this
01:10:13
podcast wondering for your kids etc.
01:10:16
there's a very simple way to not be
01:10:18
scared which is to embrace and use these
01:10:19
tools. The top two people in my
01:10:21
organization out of 20 people who are
01:10:23
using openbot and building Ultron and
01:10:26
the age of Ultron they are worth each
01:10:28
one of them is worth 200 of the other
01:10:31
employees and they only have 18 other
01:10:32
ones. If you're a young person, just
01:10:35
embrace these tools. Open open claw this
01:10:37
weekend. Build on it and you will be
01:10:39
infinitely
01:10:41
employable
01:10:43
for the rest of your life if you just
01:10:44
embrace these tools. I wanted to give
01:10:46
you your flowers, Brad. Couple years
01:10:48
ago, you came on this podcast, you
01:10:50
started talking about these America
01:10:52
accounts. You got Michael Dell to
01:10:54
partner with you on it and to put a
01:10:56
little bit of money. As we know, 40% of
01:10:58
the country do not have exposure to the
01:11:00
equities that are going bonkers up and
01:11:03
down, but generally up and to the right.
01:11:06
And you have now created Trump accounts.
01:11:09
You were at the White House. You had the
01:11:10
big launch. Let me say, Brad, this would
01:11:13
not be a law if Brad Gersonner did not
01:11:16
pursue it with absolute dogged
01:11:19
determination. Relentless. I've gotten
01:11:22
calls and texts from Brad at 6:00 a.m.
01:11:24
and at 2 am. You may sleep less than the
01:11:27
president, Brad. And then that is a
01:11:29
remarkable thing because I don't think
01:11:30
he sleeps at all. Nicki Minaj is singing
01:11:33
about Trump accounts apparently with
01:11:35
Bessant. It's very strange, but it's
01:11:37
happening. Just take us through why you
01:11:40
did this and the impact you hope it has
01:11:42
in the coming decades.
01:11:44
>> Well, you know, Friedberg just alluded
01:11:46
to it.
01:11:48
These are very destabilizing
01:11:50
destabilizing forces, right? You can't
01:11:52
have a trillionaire and 70% of people
01:11:54
feeling out that they're left out and
01:11:56
left behind and the system's rigged
01:11:58
against them and they're not in the game
01:11:59
of capitalism. Less than half of the
01:12:01
people under the age of 40 have a
01:12:02
positive view of capitalism. So, you
01:12:04
know, we set out on this journey. We
01:12:06
talked about it here uh to make
01:12:07
everybody a capitalist, give everybody
01:12:09
an ownership stake in the upside of
01:12:11
America. It passed the Invest America
01:12:12
Act became the law of the land as part
01:12:14
of the big beautiful bill. And now we're
01:12:16
in the process of rolling it out. In
01:12:18
fact, in the last, I think five days,
01:12:20
1.5 million families and kids have
01:12:22
claimed their account. It's embedded
01:12:24
within the tax filing system. All you
01:12:26
have to say is yes, I want to want to
01:12:27
claim my account. But what this means is
01:12:29
that forever more, right, we've had a
01:12:32
dramatic change to the social contract.
01:12:34
Every child born in the United States
01:12:36
forever more will start life off with an
01:12:38
investment account seated with $1,000 in
01:12:41
the S&P 500. They'll own a little bit of
01:12:43
SpaceX. They'll own a little bit of Open
01:12:45
AI. they'll own a little bit of Nvidia.
01:12:47
That is what we need to do is just a
01:12:49
first step in making sure we can hold
01:12:52
this experiment together for the next
01:12:54
250 years. Right? When we have this rate
01:12:57
of change. And so the president said
01:12:59
something on stage last week. In 15 to
01:13:01
20 years, we will have $4 trillion of
01:13:05
wealth that will have been transferred
01:13:07
to people who would have otherwise had
01:13:09
zero. 75 to 100 million families who
01:13:13
have $4 trillion who would have
01:13:15
otherwise had zero. I think it's an
01:13:17
incredible first step in fighting the
01:13:19
battle on behalf of capitalism and the
01:13:22
American dream. We see the drift towards
01:13:24
socialism, the false promises of
01:13:26
socialism in order to to fight back
01:13:28
against that. I think a great first step
01:13:30
is the Trump accounts which makes
01:13:32
everybody a capitalist from birth.
01:13:35
I just want to say, you know,
01:13:38
you know, just bestie to bestie, Brad,
01:13:40
watch you conceive of this and get it
01:13:43
done, you know, all the impressive stuff
01:13:45
you've done in your career, I think will
01:13:47
be a footnote to this, I think this is
01:13:49
your legacy. So, I just want to
01:13:50
congratulate you on that. And I also
01:13:52
want to congratulate Michael and Susan
01:13:55
Dell, who had they not stepped up and
01:13:58
done this with you, I don't know if this
01:13:59
would have come together. And then I
01:14:01
also want to congratulate President
01:14:02
Trump for just putting through something
01:14:06
that bridges the gap between the equity
01:14:08
holders and the non-equity holders, the
01:14:10
bottom half of the country, the top half
01:14:12
of the country. This is visionary. You
01:14:14
can say what you want about Trump. You
01:14:15
may like certain things, you might not
01:14:16
like certain things, ICE, whatever. I've
01:14:18
been very vocal about certain things.
01:14:20
This is perhaps one of the greatest wins
01:14:22
for you, uh, Michael and Susan Dell and
01:14:26
President Trump, the administration, and
01:14:28
for all Americans. And there's very few
01:14:30
things that all Americans can get around
01:14:32
right now. It's such a divisive,
01:14:34
disgusting political climate.
01:14:36
Everybody's fighting with each other
01:14:38
over everything. And what you pulled
01:14:40
together here with this is just
01:14:42
extraordinary in that all Americans can
01:14:45
take a win for once. All Americans can
01:14:49
say, "Hey, we did something fantastic."
01:14:51
And without you, Brad, it wouldn't have
01:14:52
happened. So just bestie to bestie, I
01:14:54
want to just congratulate you. All
01:14:56
right. Listen, Freedberg, you hate
01:14:59
socialism. You're concerned about
01:15:00
socialism.
01:15:02
>> This helps. Yeah, this helps.
01:15:06
Giving me a lot to work with today.
01:15:08
Freeberg.
01:15:09
>> Well, I mean, you just had your big
01:15:12
nicotine pouches.
01:15:14
>> Well, you just had your big clothes. I
01:15:15
don't think it's What do you
01:15:16
>> No, I just you I wanted to give you a
01:15:18
chance to shine here. You hate
01:15:19
socialism. You're concerned about
01:15:21
socialism. Is this not one of the best
01:15:23
ways to fight against the socialistic
01:15:25
urge to just have collectivism and just
01:15:28
steal from the top half and give to the
01:15:30
bottom half or you know uh seize the the
01:15:33
you know production and and
01:15:35
manufacturing? This this is a great
01:15:37
solution is to get everybody into the
01:15:38
game.
01:15:39
>> Honestly, it's a longer conversation. I
01:15:41
think we got to number one
01:15:43
slashgovernment spending like crazy,
01:15:45
>> okay,
01:15:46
>> and reduce inflation as a result. Number
01:15:48
two, stop with these defined benefit
01:15:51
retirement programs, which means telling
01:15:53
people, "Here's what you're going to end
01:15:54
up with." This idea that everyone gets
01:15:56
an account and you can track your
01:15:57
account like a 401k, which is a defined
01:16:00
contribution program, is what all of
01:16:03
Social Security should move to.
01:16:04
>> And we should take all of Social
01:16:06
Security
01:16:07
>> and we should capitalize it. Right now,
01:16:09
there's nothing in Social Security.
01:16:10
There's a $4 trillion note
01:16:13
>> that the government owes the Social
01:16:15
Security trust fund. People don't
01:16:16
realize this. But social security is an
01:16:18
independent trust fund that's set up and
01:16:21
it holds one asset. That asset is an IOU
01:16:25
from the US government to that trust
01:16:27
fund because
01:16:28
>> because the government has taken all the
01:16:30
money that you put in as a employee gets
01:16:34
taken out of your paycheck and instead
01:16:36
of going into that account it goes to
01:16:37
the US Treasury and the US Treasury
01:16:40
>> they put an IOU and they put it back in
01:16:41
the social security. So you expect that
01:16:43
you're going to get some retirement
01:16:44
benefit in the future. We need to change
01:16:46
all we need to change all of that to
01:16:48
make that a defined contribution. So
01:16:50
every time you put money out of your
01:16:52
paycheck, you should open an account and
01:16:54
see where that money is. And you should
01:16:56
say, "Okay, that money is in Google,
01:16:57
it's in Amazon, it's in Ford, it's in
01:17:00
this healthare system, it's in all these
01:17:02
things that I now own a piece of." And
01:17:03
you see it going up like a 401k owner
01:17:05
does every year. We have to transition
01:17:07
that in the United States. I hope we can
01:17:09
get it done in parallel with cutting the
01:17:12
spending that is fundamentally driving
01:17:14
the inflation and making things
01:17:15
unlivable in this country. Cut the
01:17:17
regulations so that we can make it
01:17:19
easier for people to own homes and get
01:17:21
rid of the government telling people
01:17:23
every year that they're going to do more
01:17:24
for them and entrapping people in a life
01:17:27
of servitude and inaccessibility to
01:17:30
transitioning themselves up the ladder
01:17:33
which is what is driving the socialism.
01:17:35
So there's a bigger problem, longer
01:17:37
conversation, but I think this is a
01:17:38
great step.
01:17:39
>> All right, Brad. Another way, another
01:17:41
way to translate Freeberg. He says,
01:17:43
"Congratulations on your efforts. More
01:17:45
work to do. Can you turn those America
01:17:47
accounts into superanuation funds?" So
01:17:50
more work to do. Everybody should put
01:17:51
12% of their instead of into social
01:17:54
security, they should put it into their
01:17:55
Invest America account, their Trump
01:17:57
accounts. Great job everybody. Another
01:17:59
amazing episode of the world's greatest
01:18:01
podcast
01:18:02
for Jim Polyapatio. We couldn't make it
01:18:04
this way week. We we missed you bestie
01:18:06
for David Saxs, brother in arms in
01:18:09
Texas, the great state of Texas. David
01:18:10
Freeberg, Sultan of Science and fifth
01:18:12
bestie, Brad Gersonner. I am the world's
01:18:16
greatest moderator according to some
01:18:17
people on the world's greatest podcast
01:18:19
and we will see you next week. Love you
01:18:21
besties.
01:18:24
>> Let your winners ride.
01:18:31
We open sourced it to the fans and
01:18:33
they've just gone crazy with it.
01:18:35
>> Queen of
01:18:43
besties are gone.
01:18:46
>> That is my dog taking your driveways.
01:18:51
>> Oh man, my appetiter will meet me up.
01:18:54
>> We should all just get a room and just
01:18:55
have one big huge orgy cuz they're all
01:18:57
just useless. It's like this like sexual
01:18:59
tension that you just need to release
01:19:00
something else.
01:19:05
>> Your feet.
01:19:07
We need to get merch.
01:19:09
>> I'm going all in.
01:19:17
I'm going all in.