00:00:00
All right, everybody. Welcome back to
00:00:02
the number one podcast in the world, the
00:00:04
Fantastic Four, the original.
00:00:07
>> Oh, the cast is back.
00:00:08
>> The cast is back.
00:00:09
>> Our brothers in arms.
00:00:10
>> Brothers in arms. Here we go, boys. We
00:00:14
got a big Newsweek. David Saxs is back
00:00:17
and he's in the great state of Texas.
00:00:20
How's it been, Sachs? How's Texas been
00:00:22
for you so far?
00:00:23
>> It's been great. Although, I just got
00:00:25
back from DC. I got like three hours
00:00:27
sleep last night. So, but we had a lot
00:00:29
of news this past week.
00:00:31
>> Yes. And we'll be talking about PCAST
00:00:33
and your role in the and da and your
00:00:37
role going forward in the Trump
00:00:39
administration. Big news that we'll be
00:00:41
talking about today also relates to you.
00:00:43
Oh, Sultan of Science David Freeberg
00:00:45
with your background
00:00:48
from the iconic film for those not
00:00:50
watching. Looks like the iconic Thelma
00:00:52
and Louise. I wonder if that has
00:00:53
something to do with the budget of
00:00:55
California, which you've been outspoken
00:00:57
about recently.
00:00:58
>> Great rant, which I retweeted.
00:01:00
>> Mega rant with I retweeted it, too.
00:01:04
>> If only if only you could be allowed the
00:01:07
time and space to do those kinds of
00:01:09
rants on this pod.
00:01:10
>> Thank you very much. If you kept
00:01:14
talking, here he is. He's going again.
00:01:16
He's going again.
00:01:17
>> There it is. Dr. Doom. Dr. Doom, your
00:01:20
mayor, your new governor. Are you would
00:01:22
you consider it freeberg after Ohalo
00:01:24
running for governor? There is no after
00:01:26
oh halo.
00:01:27
>> Oh please do it after.
00:01:28
>> Oh please do it.
00:01:30
>> Wow that'd be so great.
00:01:31
>> I'm tempted to buy a hollow for like
00:01:33
five or six billion. So he just politics
00:01:36
California politics is dirty man.
00:01:37
>> I don't even know what it does. I'll
00:01:38
just have somebody else deal with the
00:01:40
oppo research. But the amount of
00:01:43
>> No no no no. We'd get him elected. He
00:01:44
would do an incredible job. He would
00:01:46
save the fourth largest economy in the
00:01:48
world. It would be incredible.
00:01:49
>> It'd be amazing. I would love the
00:01:51
>> Here's the oppo resource act. David
00:01:53
Freeberg went to a rave in 1999
00:01:57
and stayed up until 10:00 a.m. We have
00:02:00
witnesses
00:02:01
>> once he he got tilted at the poker game,
00:02:03
stole a bunch of pistachios and lactate
00:02:05
and ran home.
00:02:06
>> He did.
00:02:10
>> Let your winners ride.
00:02:17
said,
00:02:17
>> "We open sourced it to the fans and
00:02:19
they've just gone crazy with it."
00:02:25
>> Anthropic on a generational run and open
00:02:29
AI crashing out a bit, boys. Let's chop
00:02:32
it up here. Just looking at Anthropic
00:02:36
pretty major heater this year. January,
00:02:39
they launched co-work for business
00:02:41
users. You know what that does? Cron
00:02:42
jobs. You can connect to your Gmail,
00:02:45
your notion, whatever it is. And then
00:02:46
opus 4.6 which consensus wise everybody
00:02:50
thought this is a major step function
00:02:52
Jensen Michael Dell everybody's called
00:02:54
it out Jensen actually called it back in
00:02:57
November an inflection point and the
00:02:59
first agentic model and uh opus 4.6
00:03:05
has basically Dell said hit a threshold
00:03:08
that we haven't seen before in terms of
00:03:10
real productivity in teams. February,
00:03:12
they dropped a bunch of claw code
00:03:14
plugins that caused the SAS apocalypse.
00:03:17
Not the SAS apocalypse, the SAS software
00:03:20
as a service.
00:03:20
>> It was a SAX apocalypse, too.
00:03:22
>> Yeah, there was a little bit of that
00:03:23
back and forth as well.
00:03:25
>> No, I mean, as a SAS investor, it was a
00:03:27
SAS.
00:03:28
>> It was a bit of a you you were the tip
00:03:31
of the spear there. We'll get into it.
00:03:32
>> My exit comps were affected, that's all.
00:03:34
>> Yes. It seems like you may have divested
00:03:38
at exactly the right time. All right. $6
00:03:41
billion in annual run rate was added in
00:03:44
February alone. Brad referenced a couple
00:03:46
of weeks ago here on the pod. Earlier
00:03:49
this week, they announced computer use.
00:03:51
A new agentic system for enterprisegrade
00:03:54
kind of openclaw functionality. Now you
00:03:57
can use the clawed app from your phone
00:04:00
to control your desktop computer. Really
00:04:02
slick feature. Here's the calendar
00:04:05
release over the past two months for the
00:04:08
team at Enthropic. Dario, come on the
00:04:10
pond anytime. Uh Sachs, you've had a
00:04:14
couple of flare-ups uh and obviously the
00:04:16
administration uh and the Department of
00:04:19
War had their kurluffle,
00:04:21
but just, you know, looking at it
00:04:24
objectively, what's your take on the the
00:04:27
surging anthropic
00:04:29
generational run as I've described it
00:04:31
here?
00:04:32
>> Well, I've never been a critic of
00:04:34
Anthropic's products. I've always been
00:04:36
an admirer of their products. I think
00:04:38
last year I gave him credit for MCP. I
00:04:40
agree that they seem to be performing
00:04:41
very well now. The company made a big
00:04:44
bet on coding as the kind of big
00:04:47
breakout use case. Whether that was done
00:04:49
for business reasons or ideological
00:04:51
reasons, I'm not sure. Anthropic is sort
00:04:53
of the most AGI pill of all the frontier
00:04:56
labs. And I think they made this bet on
00:04:58
coding as their way to get to recursive
00:05:00
self-improvement. As it turns out, it
00:05:02
was a very good business move as well
00:05:04
because code is the gateway into
00:05:07
enterprise and enterprise IT budgets.
00:05:10
And so they've been able to grow revenue
00:05:12
pretty quickly as a result of getting
00:05:14
into enterprise. Also, coding seems to
00:05:17
be the basis for these other product
00:05:20
extensions. So like you said, they went
00:05:22
from cloud code to cloud co-work. The
00:05:25
idea being that well if you can generate
00:05:27
code you can also generate powerpoints
00:05:30
or spreadsheets and you do that by
00:05:32
generating the code to create that
00:05:34
output. So that was the first extension.
00:05:37
Now they are extending into agents. This
00:05:40
computer use product is kind of like an
00:05:42
open claw knockoff. So it looks like the
00:05:46
generational run for Mac Mini is just
00:05:47
about over. Look, I think they're firing
00:05:49
on all cylinders. My issues with them in
00:05:52
the past were related to what I have
00:05:55
called the regulatory capture strategy.
00:05:56
They do want a permissioning regime in
00:05:59
Washington for chips and models, meaning
00:06:02
you have to go to Washington to get
00:06:04
permission to release new models or to
00:06:07
sell GPUs anywhere in the world. I think
00:06:08
that's excessively heavy-handed. Their
00:06:11
motives for doing that may be pure. It
00:06:13
may not be regulatory capture. It may be
00:06:14
ideologically motivated. Regardless, I
00:06:16
do think it is a form of regulatory
00:06:18
capture because it plays into the hands
00:06:21
of the big companies and creates moes
00:06:23
that new entrance will not be able to
00:06:25
overcome. So I have, let's say,
00:06:29
philosophical objection to that part of
00:06:30
it. But again, I'm not a detractor of
00:06:32
their products by any means with respect
00:06:33
to what happened between them and the
00:06:34
Pentagon. I'm not involved in that. I've
00:06:36
stayed out of military procurement in
00:06:38
general. I don't get involved in what
00:06:39
are called party matters. I just focus
00:06:41
on policy matters which affect
00:06:44
>> the whole space. I saw Emil Michael
00:06:46
making this point a couple weeks ago on
00:06:48
our podcast that if you as a company
00:06:51
don't want your products to be used in
00:06:53
war, don't sell to the Department of
00:06:56
War. It's in the name, but if you do,
00:06:58
you know, if if you do decide to sell to
00:07:01
the Department of War, you should expect
00:07:02
it to be used for all lawful uses. So, I
00:07:07
think that was a very pragmatic
00:07:08
observation. Again, I just have to
00:07:10
underscore this again. I'm not involved
00:07:12
in that dispute. It's the base of a
00:07:14
lawsuit right now and I don't want
00:07:16
someone trying to draw lines between
00:07:18
dots that aren't there. So
00:07:19
>> yeah, and in fairness,
00:07:20
>> I'm staying out of that one.
00:07:21
>> Objectively, they've been treated the
00:07:23
same as any other large language model.
00:07:26
Even though they're not fans of the
00:07:27
administration, they're not donators to
00:07:28
the administration. They have
00:07:30
specifically been critical of the
00:07:32
administration as a company perhaps
00:07:34
cynically Freeberg as a strategy to get
00:07:38
uh you know it's one of the conspiracy
00:07:40
theories here in Silicon Valley is
00:07:42
Daario is taking the position of being
00:07:44
anti- this administration anti-president
00:07:47
Trump in order to get all the PhDs you
00:07:51
know there's like three or four thousand
00:07:52
of these highly sought after PhDs and
00:07:54
it's a way to have them you know vote
00:07:56
with their presence to come work in
00:07:58
anthropic your thoughts on that and then
00:08:00
just generally their general
00:08:02
>> I think he actually believes it and I
00:08:03
think they've actually created and
00:08:05
fostered a culture of that since the
00:08:06
beginning and I think that they're
00:08:08
representing it as a branding exercise
00:08:10
at this point but I don't think it's
00:08:12
made up I think it's directly
00:08:15
representation of the people that work
00:08:16
there and what they believe
00:08:18
>> yeah and it's a strategic advantage
00:08:20
because probably of those 3,000 PhDs 90%
00:08:24
of them are left-leaning uh and wouldn't
00:08:26
want to work necess necessarily. I mean
00:08:29
like like most things we see in the
00:08:33
world today, in economics today, in
00:08:34
markets today, in business today,
00:08:36
everything seems to be politicized. And
00:08:38
you have a left and a right version of
00:08:39
everything. You have a left and a right
00:08:41
version of media. You have a left and a
00:08:42
right version of what food to buy. You
00:08:44
have a left and a right version of what
00:08:46
AI tool to use. So, you know, this
00:08:48
effectively may just be the natural
00:08:50
manifestation in the AI market of what's
00:08:52
going on elsewhere in society as we all
00:08:54
kind of fracture and hustle over to our
00:08:56
side.
00:08:57
Uh, Chimath, before I go to Open AI and
00:08:59
their recent moves, any thoughts on
00:09:02
Anthropic and Daario's positioning of
00:09:04
the company?
00:09:05
>> Look, I think both are incredible
00:09:07
businesses. We're in the part of the
00:09:08
cycle where
00:09:10
we're trying to create drama where I
00:09:12
don't think drama exists because they're
00:09:15
still fundamentally in very different
00:09:17
gotomarket motions. Now they may
00:09:20
converge and compete over time but I
00:09:23
think it's important to separate where
00:09:24
each of them are good from an enterprise
00:09:26
lens which is where I see most of the
00:09:28
action particularly through 8090 it's
00:09:32
all anthropic all the time and I agree
00:09:34
with Sachs my philosophical issues with
00:09:36
the management aside around their
00:09:39
ideology and sometimes how they use some
00:09:42
of the capital for things other than
00:09:45
tech and R&D. I have issues with those
00:09:47
things. But in terms of the quality of
00:09:50
that technical team and what they
00:09:51
create, it's head and shoulders above
00:09:53
anything else. It allows us to build a
00:09:56
vibrant business. Now, do I have issues
00:09:59
with how much it costs? Yes. Do I have
00:10:00
issues with how fast we're consuming
00:10:02
tokens? Also, yes. But I think those
00:10:05
will get sorted out. And those are
00:10:07
really tactical issues. So the reason
00:10:09
why I think we're all breathlessly
00:10:12
trying to pit open AI versus anthropic
00:10:14
is because we want some drama. But the
00:10:16
reality is these are very different
00:10:18
businesses. And Nick found this tweet
00:10:20
which I thought was really interesting.
00:10:22
And because even at the absolute highest
00:10:24
level, these things are sort of
00:10:27
presented in an apples to oranges way.
00:10:30
And there's like these very basic issues
00:10:33
of revreck that are fundamentally
00:10:36
different. And you may say, well, who
00:10:37
cares about revenue recognition? Well,
00:10:40
the people that are trying to write the
00:10:41
headlines that say one is overtaking the
00:10:43
other and this or that sort of miss the
00:10:46
fact that they're in completely
00:10:48
different businesses, which has guided
00:10:49
how they even think about growth. And
00:10:51
so, if you normalize these two
00:10:53
businesses, what you would see is OpenAI
00:10:55
is still the overwhelming revenue
00:10:57
generator in this space. And that over
00:11:00
time, Anthropic is catching up. And so
00:11:02
this is this little diagram
00:11:05
that tries to explain this. OpenAI is
00:11:08
3/4 consumer subscriptions
00:11:11
and a quarter API. Enthropic is almost
00:11:14
the exact opposite.
00:11:16
OpenAI is used by consumers
00:11:17
overwhelmingly. Antropic is used either
00:11:20
directly or through things like GitHub
00:11:22
and cursor.
00:11:24
OpenAI as a result has a very
00:11:25
conservative way of recognizing revenue.
00:11:28
Anthropics, they sort of recognize gross
00:11:31
tonnage as their revenue. And so when
00:11:34
you start to hear these things about
00:11:35
like, oh, this thing is at 20 billion
00:11:37
and OpenAI is at N billion, they're two
00:11:40
totally different conversations. And I
00:11:41
think right now it's more about the
00:11:43
press cycle of trying to create clicks
00:11:45
than it actually is about the underlying
00:11:47
quality of each business. Both are
00:11:49
incredible businesses, as this
00:11:52
demonstrates. And by the time it goes
00:11:54
public, both of these two businesses
00:11:57
will have a very clean and I suspect
00:12:00
normalized way of telling a story so
00:12:02
that you can actually compare. But what
00:12:04
I would tell people right now is
00:12:06
everybody is running with numbers to try
00:12:08
to create a narrative that I don't think
00:12:10
makes sense or applies to either.
00:12:12
>> Yeah. And there's been a lot of strategy
00:12:14
change. Open AI, some people are saying,
00:12:16
is crashing out in panic mode.
00:12:19
Obviously, they own the consumer with
00:12:21
chat GPT. they are the verb like you
00:12:23
know taking an Uber or googling
00:12:25
something people consumers always just
00:12:27
say hey did you check chat GPT but
00:12:30
obviously other large language models
00:12:32
are catching up here's
00:12:35
>> can we say something to that Jason cuz
00:12:36
like
00:12:37
>> you mentor tons of startups
00:12:40
>> sax has done it free has done it I do it
00:12:44
>> what is the one thing we tell folks
00:12:45
focus focus focus focus
00:12:47
>> 100%
00:12:49
>> do one maybe one and a half things, but
00:12:52
do it incredibly incredibly well and
00:12:55
everything else you start to bleed and
00:12:57
smear. What was that Brad Garlinghouse
00:12:59
term? Peanut butter. Peanut butter.
00:13:00
Yeah, you smear the peanut butter
00:13:03
>> too far out. And so this is a good
00:13:06
moment, by the way, if either of these
00:13:07
two companies are in the smearing phase
00:13:09
to recalibrate and reset because
00:13:12
>> yeah, you just can't do everything.
00:13:14
Speaking of smears, I couldn't help but
00:13:16
notice that Emil Michael was smeared by
00:13:21
an article. Was it in Lever or something
00:13:22
like that accusing him of of having a
00:13:25
conflict in the anthropic dispute?
00:13:29
>> Did you guys see that?
00:13:30
>> I didn't. No.
00:13:31
>> Yes. I saw the art. There was an article
00:13:33
that said that he was an investor in
00:13:36
perplexity and therefore he's conflicted
00:13:39
in his negotiation with Anthropic.
00:13:42
That's what the article said pretty
00:13:43
much.
00:13:43
>> Perplexity is LLM agnostic. That's a
00:13:46
stupid claim,
00:13:47
>> right? Well, it's obviously written by
00:13:48
people who don't understand anything
00:13:50
about AI really.
00:13:51
>> Okay.
00:13:52
>> Perplexity is a rapper. You're like
00:13:54
you're saying, it uses multiple AI
00:13:55
models.
00:13:56
>> And I don't think they sell to the
00:13:59
Pentagon. They're not a competitor to
00:14:00
Anthropic. And moreover, as I understand
00:14:03
it, Emil's ownership of shares in that
00:14:06
company was blessed by the Office of
00:14:07
Government Ethics. Nonetheless, I think
00:14:09
the timing of this is very suspicious.
00:14:12
And it reminds me of what happened to me
00:14:14
when I started opposing Anthropic and
00:14:16
all of a sudden there was that hit piece
00:14:18
in the New York Times accusing me of
00:14:19
having conflicts. And I'll just say that
00:14:22
Anthropic may pose as this company
00:14:25
that's on the side of the angels, but
00:14:27
they've hired a number of very seasoned
00:14:29
brass knuckle political operatives in
00:14:32
Washington. And you know, members of the
00:14:34
Biden administration, Laura Loomer,
00:14:36
actually just had a piece today on one
00:14:37
of them. I'm not going to rehash that.
00:14:39
But the bottom line is this is I think
00:14:42
frankly a political operation that's
00:14:44
willing to get down and dirty and
00:14:46
they're not always on the side of the
00:14:47
angels. I think they can be quite
00:14:49
ruthless.
00:14:49
>> Saxs, remember I told you you get one
00:14:51
Biden mention a month in 2026. So you
00:14:54
just used it up. So I don't want any
00:14:55
more Biden Biden Biden. He's retired.
00:14:58
>> That was not a Biden mention.
00:15:00
No, but the truth is
00:15:03
these [ __ ] whoever wrote this
00:15:04
story, the actual best feature or
00:15:07
amongst the best features of Perplexity,
00:15:09
which is actually got a really great
00:15:11
co-work competitor called Computer I've
00:15:13
been playing with. I'm not a shareholder
00:15:14
to be clear. There's no book being
00:15:17
pumped here. The model council is like
00:15:19
the greatest feature.
00:15:21
>> Yeah.
00:15:21
>> That they have. And what is really
00:15:23
brilliant about it is you ask it a
00:15:25
question sachs. It will go to all three
00:15:28
different major models and you can pick
00:15:30
which ones including open source. Then
00:15:32
it tells you where they differ and it
00:15:34
tries to figure out why they differ.
00:15:36
This is like one of the great features
00:15:38
of the product. I think perplexity is
00:15:40
like could be a a really great company
00:15:42
as well even without a language model.
00:15:45
But let's talk a little bit more about
00:15:46
OpenAI here
00:15:48
and their market share because I think
00:15:50
you're correct Chimoth, but they are
00:15:52
getting off their game. Here's what's
00:15:54
going on. Quick look at the consumer
00:15:56
market
00:15:58
and obviously they started with 100%
00:16:01
market share, right? They created the
00:16:02
category in 2023,
00:16:05
dropped down to 85% market share 2024,
00:16:08
75% market share 2025.
00:16:10
>> But by how much has the market grown
00:16:13
>> precisely? So uh the market is still
00:16:15
growing. So on terms of number of
00:16:17
searches and queries, they're obviously
00:16:19
growing tremendously, but they have
00:16:21
major major competitors and the market
00:16:23
share is going down. I had my team over
00:16:26
at This Week in AI do a more thoughtful
00:16:29
analysis of where this is going. And if
00:16:31
you take a look at this, um there's
00:16:33
three players, and I'd like to get your
00:16:36
guys' take on this who really haven't
00:16:37
shown up yet. Apple,
00:16:40
uh Meta, and obviously Windows. All
00:16:43
three of those underrepresented. If you
00:16:45
give them credit for just getting, you
00:16:48
know, a half point of market share here
00:16:49
and starting to intercept, which I think
00:16:51
those three players were here, will be
00:16:53
here, they're going to be well under 50%
00:16:55
market share. And I think CHPT is going
00:16:58
to have some big challenges there on the
00:16:59
consumer side. While they're doing this
00:17:01
stuff in consumer, uh they uh are
00:17:04
cutting back on all their side projects.
00:17:06
So, uh you probably heard about the Sora
00:17:08
video app that's been shut down. This is
00:17:11
kind of major news because Disney was
00:17:14
going to put a billion dollars into
00:17:17
opening as part of it and they had done
00:17:19
a licensing deal and they were going to
00:17:20
integrate Sora, this you know short
00:17:23
video product into Disney Plus. All of
00:17:25
that's now been cancelled.
00:17:27
>> The billion is not going in.
00:17:28
>> The billion's not going in. The
00:17:30
licensing deal, all of that. And then in
00:17:33
addition, there's supposedly at OpenAI a
00:17:36
newfound focus on chasing Anthropic down
00:17:40
the enterprise path. So getting off
00:17:42
their game, getting a little
00:17:43
discombobulated perhaps or maybe getting
00:17:46
focused on what matters, which is
00:17:48
enterprise apparently in terms of
00:17:50
revenue. OpenAI also offered private
00:17:53
equity investors a guaranteed minimum
00:17:55
return of 17.5%
00:17:57
of as part of a joint venture that would
00:18:00
help PE firms deploy AI and ease the
00:18:04
high upfront cost of that. So, lots of
00:18:07
questions here. Chimov, I don't know if
00:18:09
you've tracked this PE uh model, but
00:18:12
obviously a lot of people are doing
00:18:14
rollups in services, uh, accounting,
00:18:17
legal, uh, Josh Kushner's got a big
00:18:19
effort here, a bunch of private equity
00:18:21
firms trying to essentially, I guess,
00:18:24
endrun the transition process and
00:18:28
arguably what you're doing with the
00:18:30
software factory at 8090. So, your
00:18:32
thoughts on Open AI and this pivot and
00:18:34
this private equity I think it makes a
00:18:36
lot of sense for Open AI to focus on a
00:18:38
few things and do them exceptionally
00:18:40
well.
00:18:41
>> I disagree slightly with your first
00:18:43
part, which is I think that people like
00:18:45
to make new decisions about new
00:18:47
experiences and I think that OpenAI has
00:18:51
incredible consumer mind share. I just
00:18:54
see like how my kids use it. They
00:18:55
started there and it's very hard to get
00:18:58
them to switch. Even when I say, "Hey,
00:18:59
have you tried Gemini?" They use Gemini.
00:19:01
And to your point, the reason is because
00:19:03
they stumble into it more.
00:19:05
>> Yeah. Yeah,
00:19:05
>> but if you give them a cold navigation
00:19:07
experience, they rely on chat GPT
00:19:10
and it's the same by the way on the
00:19:12
other side in the enterprise. If you
00:19:13
give us a cold,
00:19:17
my default reaction would be to use
00:19:19
anthropic. Now, I actually think that
00:19:21
that's quite healthy because you're
00:19:22
going to segregate the market. And I
00:19:24
think look if you go into the way back
00:19:25
machine when we first started talking
00:19:27
about this thing this is sort of how we
00:19:29
all postulated this would work where
00:19:31
even if open AI just won the consumer
00:19:33
business it is a multi- trillion dollar
00:19:36
company with enormous scale in value and
00:19:40
I think that that's okay. So I think
00:19:42
that what they probably need to do is
00:19:44
say where are we the strongest? Where is
00:19:47
there the most obvious traction?
00:19:50
Can traction in another market like an
00:19:52
enterprise bleed into consumer usage?
00:19:56
If it's true, then you have to win the
00:19:58
enterprise. I think winning the
00:20:00
enterprise though is a very different
00:20:01
game than winning consumer. Very
00:20:04
different set of features, very
00:20:05
different set of expectations. So, I
00:20:07
think people either have to decide
00:20:10
you're going to compete everywhere or
00:20:12
pick one thing and just nail it. And if
00:20:14
I was open AI and you had to pick one
00:20:16
thing, you would pick consumer because
00:20:19
they're the juggernaut and they're the
00:20:21
clear leader and they have an enormous
00:20:23
brand. Freeberg, let me pull you into
00:20:24
this because my base case here is that
00:20:28
all consumer queries are going to be
00:20:30
free. Apple's going to make them free.
00:20:32
They're already free for Google. I think
00:20:34
Meta is going to make them free and
00:20:36
actually have a decent product soon. And
00:20:38
Microsoft, same thing. Chat GBT has
00:20:40
decided to push off advertising. They
00:20:43
were going to put advertising in it. You
00:20:44
remember they got mocked by Anthropic
00:20:46
with their Super Bowl ads. So, what do
00:20:48
you think is going to happen on the
00:20:49
consumer side? Consumers generally don't
00:20:51
pay for services. That's usually five to
00:20:53
20% of the market is paid services and
00:20:56
everything else is free and ad
00:20:58
supported. But it looks like Apple and
00:21:00
and Google are going to just let it rip.
00:21:04
So that could take the revenue oxygen
00:21:06
away from Chat GPT. So what is your
00:21:09
thought here on who wins consumer?
00:21:11
>> I don't think that it's going to be
00:21:12
free. I think there's 290 million
00:21:16
subscribers
00:21:18
for Spotify. They're paying what are
00:21:20
they paying 20 bucks a month or
00:21:21
something?
00:21:22
>> Probably less than average because it's
00:21:23
global number, but yeah,
00:21:25
>> Netflix has 325 million paid
00:21:28
subscribers.
00:21:29
and AI that can book your travel, answer
00:21:32
questions for you, track your calendar,
00:21:34
do your email, etc., etc., etc., etc.,
00:21:37
is likely going to be the most valuable,
00:21:39
call it meta service that consumers have
00:21:41
ever seen.
00:21:42
>> And I think it's very likely that we're
00:21:44
going to end up seeing many more
00:21:45
consumers subscribe to a consumer AI
00:21:48
service than we've seen even with cable
00:21:50
television. I mean, think about your
00:21:51
cell phone. Everyone's paying 50, 60
00:21:53
bucks a month for a cell phone. Why not
00:21:55
pay 80 bucks a month? More bucks. you
00:21:57
pay 100 bucks and by the way in the
00:21:59
pandemic remember what we saw the two
00:22:01
things that people refused to cancel
00:22:04
>> was not your mortgage payment or any car
00:22:07
payment you were willing to go into aars
00:22:09
and into default the two things that
00:22:11
people would would always keep was the
00:22:13
cell phone number one and then
00:22:14
electricity number two chat GPT will be
00:22:16
there
00:22:17
>> so I think that's going to be the case
00:22:18
with these consumerized JL I think that
00:22:20
they're all going to be like ultra
00:22:21
valuable and they're going to layer in
00:22:23
services on top of them like for example
00:22:25
do you want to watch video embedded
00:22:27
in your consumer app. Do you want your
00:22:30
consumer AI app, you know, do your
00:22:32
finances for you? And it could be that
00:22:34
the consumer AI apps
00:22:37
much like the iPhone was for the app
00:22:39
economy. There could almost be whether
00:22:41
it's through kind of connectors or
00:22:44
embedded tools, an incredible ecosystem
00:22:46
that where traditionally advertisers
00:22:48
actually pay to be embedded and show up
00:22:50
inside of the AI app and the consumer
00:22:52
can either pay for it or the advertiser
00:22:54
can pay for it. So I think there's going
00:22:56
to be um a very different economic model
00:22:58
and still very early days. Sachs the
00:23:01
numbers
00:23:03
right now would uh be more in my
00:23:06
estimation about 50 million people
00:23:08
subscribe to chat GPT. They got a
00:23:09
billion users or they're they're trended
00:23:11
to a billion. I think they probably hit
00:23:12
it in the next month or two. Certainly
00:23:15
they had 900 million two or three months
00:23:16
ago. So it's about 5%. Where where do
00:23:19
you think this winds up? Do you think it
00:23:20
becomes, you know, 300, 400, 500 million
00:23:24
consumers are willing to pay 20 bucks a
00:23:25
month for this or do you think it's more
00:23:27
free and the data and the ad
00:23:29
supportedness of it? Go the meta and the
00:23:32
Google route.
00:23:33
>> I think it's possible that you could get
00:23:35
a few hundred million subscribers for
00:23:37
the premium tier. Look, I think most
00:23:39
consumers will take the free service in
00:23:41
exchange for advertising, right? Some
00:23:43
adup supported model, which by the way,
00:23:45
I think could be quite successful. when
00:23:48
chat GBT started displacing Google for
00:23:51
search, a lot of people were predicting
00:23:52
the death of of Google's model because
00:23:55
who'd want to look at 10 blue links. I
00:23:57
think that's true, but I think you can
00:23:58
do something much more compelling in AI
00:24:01
chat compared to just a list of links.
00:24:04
So, in any event, I think adup supported
00:24:06
models might make a comeback here in
00:24:08
addition to premium models. All that
00:24:11
being said though, you know, as an
00:24:13
investor, I always liked B2B businesses
00:24:15
better than B TOC because it is hard to
00:24:19
monetize consumers. Their willingness to
00:24:21
pay is not that high and they tend to
00:24:23
have high churn rates. Whereas
00:24:25
businesses tend to be very sticky and
00:24:27
you can upsell them and you can get more
00:24:30
than 100% net dollar retention
00:24:32
year-over-year. So, if you can make an
00:24:34
enterprise business work, it's always
00:24:36
been a model I've I've liked. But you
00:24:39
know that being said obviously some of
00:24:41
the most valuable companies in the world
00:24:42
are consumer companies. Meta, Google,
00:24:45
Apple, these are all consumer first
00:24:47
companies. And look I think ultimately
00:24:50
both models can work.
00:24:51
>> Obviously both can work. Question is
00:24:52
which one? I guess it really comes down
00:24:55
to how motivated do we think Google and
00:24:57
Facebook will be to build that bridge
00:25:00
from their ad networks to their AI
00:25:03
offerings. Obviously Facebook is kind of
00:25:06
MIA in all of this but Google is not and
00:25:09
I think that will be the determinant
00:25:11
here is
00:25:11
>> well Google is going to compete very
00:25:13
vigorously for the consumer because it
00:25:15
is existential to them I mean it's very
00:25:17
clear that search and AI chat are kind
00:25:21
of merging into one space that means
00:25:23
that ad links will kind of merge into
00:25:26
being in chat advertising so they have
00:25:28
to adapt with that and compete for the
00:25:30
consumer I also think that Google is in
00:25:33
an outstanding position to do the whole
00:25:35
claw thing because they already have
00:25:37
access to your calendar, your documents,
00:25:39
your email.
00:25:40
>> So, the agent doesn't really have to
00:25:42
earn your trust because you already
00:25:44
trust Google with all of your stuff,
00:25:46
>> right?
00:25:46
>> So, I'm kind of waiting for the Google
00:25:48
version of Open Claw because I don't
00:25:50
really want to share all my documents
00:25:52
with some new service.
00:25:54
>> They're the only one that has so much
00:25:56
free cash flow that they can almost view
00:25:58
it as two separate companies, which it
00:26:00
effectively is. you know, GCP over here
00:26:02
runs the enterprise play and then Google
00:26:05
consumer over here runs the consumer
00:26:08
chatbot play and they can keep them
00:26:10
segregated. That's so much harder for a
00:26:13
startup to do because on top of just
00:26:15
keeping everybody organized, you have
00:26:17
the financing problem of constantly
00:26:19
having to raise more money because you
00:26:21
don't yet have a profit engine that
00:26:23
spits out cash. They're probably the
00:26:25
only one. And you can see it in the
00:26:26
valuations actually which I'm going to
00:26:27
get to in a second but people believe
00:26:29
the durability of Google more than they
00:26:31
believe the durability of anything else.
00:26:33
>> And uh Sax I think you weren't on the
00:26:35
pod last week or maybe two or three
00:26:37
weeks ago but Google has announced
00:26:39
Google workspace
00:26:41
studio to do AI automation and it's uh
00:26:45
yeah it's online and people are playing
00:26:47
with it already. So they have joined the
00:26:50
open claw party. By the way, Jay Kcal,
00:26:53
you asked a different question earlier
00:26:55
as well, which is around the PE story.
00:26:57
And I think the PE story is a window
00:26:58
into the rest of the broader market. The
00:27:00
real open question is what are these
00:27:03
companies worth? There's like a very
00:27:05
threshold question. It's sort of like a
00:27:08
very important fork in the road right at
00:27:09
the outset, which is do you believe that
00:27:11
we're on a path to super intelligence
00:27:14
where everything is incredible, where
00:27:16
there's infinite abundance, where you
00:27:18
can magically describe things and
00:27:20
beautiful things appear? complex things
00:27:22
appear, groundbreaking things appear, or
00:27:25
do you believe that it's good next
00:27:27
generational software? And the answer to
00:27:29
that question is really important
00:27:31
because we're financing things like it's
00:27:32
the former
00:27:33
>> and uh GC has a big move in this. Hamont
00:27:36
was on the program in January when we
00:27:39
did the interview show and he's got GC
00:27:41
buying up
00:27:43
and rolling up accounting firms, etc.,
00:27:45
healthc care firms, hospitals, etc. This
00:27:48
seems to be part of the future of
00:27:50
venture capital is taking AI sacks and
00:27:54
actually buying out or I should say big
00:27:56
VC kind of starting to look like private
00:27:58
equity buying out hospitals accounting
00:28:01
firms business processing firms in India
00:28:04
putting them all together and then
00:28:06
running them with AI. So any any final
00:28:08
thoughts on that Sax as a business
00:28:10
strategy? Well, I think it's
00:28:12
interesting. They're kind of betting on
00:28:14
the idea that they can own the change
00:28:18
management around AI. And this is the
00:28:22
thing is everyone just kind of assumes
00:28:23
that you throw AI over a wall and a
00:28:25
business just automatically knows how to
00:28:27
use it and drive efficiency from it.
00:28:29
>> And what we're seeing is it's pretty
00:28:31
difficult. Chimath, you've seen that.
00:28:33
There's McKenzie studies showing that
00:28:34
like 95% of enterprise pilots aren't
00:28:36
successful. There's tremendous value,
00:28:38
latent value in AI, but it's hard to
00:28:41
know exactly how to deploy it at this
00:28:43
point in time. So, I guess what these
00:28:45
private equity firms are saying is that
00:28:47
we know how to drive value from this and
00:28:49
if we own the business and then own the
00:28:52
change management, then we'll be able
00:28:53
to, you know, create value that way.
00:28:56
>> Their business model makes solving this
00:28:57
problem existential.
00:29:00
And this is sort of along the lines of
00:29:02
this essay that I wrote. Let me just
00:29:03
give you the thought exercise and you
00:29:04
guys react. Today we live in a world
00:29:07
where the whole market is trying to
00:29:09
debate what is the PE ratio that you'd
00:29:12
be willing to pay. So Facebook is
00:29:16
incredibly durable. I'm willing to pay
00:29:17
30 times. Nvidia is really durable. I'm
00:29:20
willing to pay 40 times. Tesla is
00:29:22
incredibly asymmetric to the upside. I'm
00:29:24
willing to pay 200 times. And then
00:29:27
Caterpillar or John Deere, I'm willing
00:29:29
to pay 15 times. I'm just using these as
00:29:31
an example. And what it's effectively
00:29:35
signaling to you is how durable all of
00:29:37
these cash flows are. And all we do in
00:29:39
the public markets when we make an
00:29:41
investment is we're just guessing when
00:29:43
do the cash flows run out and we try to
00:29:46
say, well, here's how much it's worth
00:29:47
and here's how much I'd be willing to
00:29:49
pay for today.
00:29:51
But if you go back to this example and
00:29:52
you say, well, what if there's this
00:29:54
super intelligence on the horizon?
00:29:57
I think it's fair to ask the question,
00:29:58
what is anything worth?
00:30:01
And what is anything worth in year 10 or
00:30:03
year 15 or year 20? Because if you have
00:30:05
infinite abundance and you have all this
00:30:07
creativity,
00:30:09
won't all companies be disrupted? And
00:30:11
won't we be in this constant churn of
00:30:14
everything getting disrupted all the
00:30:16
time? And if you were faced with that
00:30:19
problem in the public markets, how would
00:30:21
you react?
00:30:23
And I think the canary in the coal mine
00:30:25
are the SAS talks. Yes, we jokingly call
00:30:28
it the SAS apocalypse, but I think it's
00:30:30
much more important. I think it's a big
00:30:31
societal question. How do you view
00:30:33
capital markets? How do you view the
00:30:36
health of a company in a world where
00:30:39
we've been told there's a super
00:30:40
intelligence on the horizon that makes
00:30:42
everything much more fragile than it was
00:30:44
before? And the market reaction is to
00:30:46
put all these companies on a spectrum
00:30:48
and they started here in software and
00:30:50
they're rerating everything down
00:30:54
and they're changing the way that things
00:30:57
are being framed from price to equity to
00:31:00
a multiple of the cash that you have on
00:31:02
hand. And I think that has huge
00:31:05
implications mostly to Silicon Valley
00:31:07
and largely to employees because we all
00:31:10
sell the dream. we start a company and
00:31:12
we're like okay small salary big equity
00:31:15
upside but that's implicitly saying in
00:31:17
15 or 20 years this thing is going to be
00:31:18
worth some gigantic number but if
00:31:21
instead every business gets disrupted
00:31:23
every 5 or 6 years all you're going to
00:31:25
end up with is just the cash and so what
00:31:27
should employees do the rational
00:31:28
reaction from employees will say you
00:31:30
know what I don't want your equity give
00:31:32
me more money
00:31:34
>> and if all of a sudden you do that the
00:31:36
valuation multiples and the complexity
00:31:38
changes again so I had my team put this
00:31:41
chart together. Okay, what is this? Took
00:31:42
a handful of SAS companies and took the
00:31:45
mag six
00:31:47
and I just said, okay, if you take the
00:31:50
market cap and you divide it by the
00:31:52
annual free cash flow, what that tells
00:31:54
you is how many years does it take to
00:31:55
get back? If you bought a share of
00:31:56
stock, how many years does it take from
00:31:58
the free cash flow to come back so that
00:32:00
you've earned back the cost of one
00:32:03
share?
00:32:04
Snowflake in 2023, it would have taken
00:32:07
you almost a hundred years.
00:32:10
And where is it now? It's been cut in
00:32:12
half. Service Now, it's last year.
00:32:14
Workday. You see it.
00:32:16
And I think what this speaks to is the
00:32:19
beginning of this rerationalization in
00:32:22
the public markets of saying if super
00:32:25
intelligence is coming, we have to be
00:32:28
very careful about what we're willing to
00:32:29
pay for these things. But if you look on
00:32:31
the right hand side and the Mag 7,
00:32:34
what's so interesting is Apple,
00:32:36
Microsoft, Meta, and Alphabet, the
00:32:38
market is completely ripped the other
00:32:40
way. And what they're saying is, we
00:32:42
believe that these cash flows are
00:32:45
essentially monopolistically durable
00:32:47
forever.
00:32:49
That's the only reason why you would
00:32:50
walk them up like this. except Nvidia,
00:32:54
which is the most unbelievably
00:32:55
accreative, well-run company, highest
00:32:58
margins, you know, making $200 billion,
00:33:02
and they're treating it like they're
00:33:03
treating Service Now and Snowflake. I
00:33:06
just think it's so interesting what's
00:33:07
happening. I can't explain this, but
00:33:09
here, this data sort of shows this reset
00:33:11
that we're going through, a very
00:33:13
complicated reset in the capital.
00:33:14
>> Free, what's your take on this reset as
00:33:17
Chimath describes it? You'd think this
00:33:20
is just a flight to the quality of the
00:33:23
free cash flow of the Mac 6 and just how
00:33:26
much cash they print and then maybe the
00:33:28
other ones are smaller footprints and
00:33:30
they're just more disruptable. We had
00:33:32
this discussion many years that Google
00:33:34
and Apple, Microsoft, they would all be
00:33:37
disrupted at some point, Facebook, and
00:33:39
that just simply hasn't happened.
00:33:40
They've gotten much more nimble at
00:33:41
copying products or incorporating, you
00:33:44
know, features and products into their
00:33:45
core offering. So, your thoughts? It's
00:33:48
probably generally correct that there
00:33:51
will be a decline, but there's also the
00:33:53
selective opportunity. Did you guys see
00:33:55
the LP slides that went on the internet
00:33:56
from Toma Brava's LP conference?
00:33:59
>> Yeah, those are great.
00:34:00
>> Nick, maybe you could find that. They
00:34:01
kind of highlight that within the broad
00:34:04
marketscape, there are companies that
00:34:06
are not just going to sit idly by and
00:34:09
let AI kind of delete their business
00:34:12
value, but they're integrating AI
00:34:14
themselves and they've got highquality
00:34:16
people to do so and they're reinventing
00:34:18
their product themselves. So, they've
00:34:19
already got a beach head, they've
00:34:21
already got customer access, they've
00:34:22
already got enterprise users. And in
00:34:24
fact, if they can integrate AI into
00:34:26
their products and into their tools,
00:34:27
there's almost this like selective
00:34:29
dispersion that happens in the market.
00:34:31
And so the winners are going to win I
00:34:33
think truly in every market not just in
00:34:36
software but across every market
00:34:38
including in industrial supply chains on
00:34:40
who is going to implement and utilize AI
00:34:43
tools and agents uh to do work and it's
00:34:46
going to expand the work uh productivity
00:34:48
of that organization not just create new
00:34:50
features which is what we focus on when
00:34:52
we talk about software companies but
00:34:53
really imagine a complex business being
00:34:56
able to do 10 times the output it can do
00:34:58
today with the same capital equipment
00:35:00
and the same labor force. What do you
00:35:02
pay for in a world of super intelligence
00:35:05
versus in a world of nons super
00:35:07
intelligence in terms of the durability
00:35:09
of a business?
00:35:10
>> Yeah, it's hard to say, man. I mean, we
00:35:12
don't know, right? And that's why all
00:35:13
the discount rates are going through the
00:35:15
roof and that's why the valuations are
00:35:17
collapsing because we just don't know
00:35:18
what multiple or what discount rate you
00:35:20
apply or what terminal growth rate or
00:35:22
which which effectively implies.
00:35:24
>> What do you do in your PA? Like do you
00:35:26
say I don't think Disneyland is going
00:35:28
anywhere? You know, I think there's some
00:35:29
stuff that you could say is the counter
00:35:31
AI portfolio. And the counter AI
00:35:32
portfolio, I think, is like physical
00:35:34
experiences.
00:35:35
>> Halo. They call it halo. High asset low
00:35:38
obsolescence.
00:35:39
>> Okay, that's a great Yeah, I'm not a big
00:35:41
investor trader like you guys, but that
00:35:43
makes a lot of sense. I mean, I think
00:35:44
that's like intuitively to me that seems
00:35:46
to be an area where people are going to
00:35:48
be spending a lot of time and they're
00:35:50
going to they're going to have
00:35:51
durability in those businesses. I I
00:35:53
think businesses like Nack Gas
00:35:55
Production, you know, I just bought LNG
00:35:57
that Shener company
00:36:01
>> I bought Shener given um all the
00:36:03
craziness in the Middle East and you
00:36:04
know I visited there with Doug Bergam so
00:36:06
it was the first time I'd ever been
00:36:07
exposed to this business and I checked
00:36:08
out a great business. It's a great
00:36:09
business really well business.
00:36:10
>> So I think that's got durability
00:36:12
obviously unless it gets blown up by
00:36:13
some enemy. Uh that that would be a
00:36:15
problem. So the Adams thesis that TK
00:36:18
>> shared Adams real life mining is a great
00:36:22
one. Obviously I think the space
00:36:24
industry is going to be a lot bigger
00:36:25
than we recognize. I actually think
00:36:27
there's probably a 15 to 30 trillion
00:36:30
dollar a year economic opportunity on
00:36:33
the moon. That kind of a business like
00:36:34
you're going to see with SpaceX's IPO is
00:36:37
going to have an insane multiple. So I
00:36:39
think to your point Jimat, there's a lot
00:36:40
of stuff getting steamrolled here with
00:36:42
crazy high discount rates where you just
00:36:43
don't know. And there's a bunch of stuff
00:36:45
where the pathway over the next 15 to 30
00:36:47
years is maybe independent or unlocked
00:36:49
because of AI. And then there's a bunch
00:36:51
of stuff that's just unaffected that and
00:36:53
that that starts to get a higher
00:36:54
multiple because that's where capital
00:36:56
starts to flow.
00:36:56
>> Sax there are probably three things that
00:36:59
we would agree are great modes for
00:37:01
businesses brands network effects and
00:37:03
the management team. Those come into
00:37:06
play here as well. Yeah.
00:37:07
>> I don't know if I would consider
00:37:08
management team to be a moat. I mean,
00:37:10
it's like Warren Buffett says that you
00:37:12
want businesses that are so strong that
00:37:13
they could be run by a bunch of monkeys
00:37:15
because one day they probably will be.
00:37:18
>> Well, I was thinking more like obviously
00:37:20
Elon and Tesla is going to just
00:37:22
relentlessly innovate.
00:37:23
>> It's a great point by the way and it's
00:37:25
true especially in the age of agentic
00:37:27
AI.
00:37:27
>> Yeah. I mean look I think though that
00:37:30
just upleveling a bit I think you are
00:37:32
right that the key question is moes
00:37:34
because I do think that there are still
00:37:36
strong moes in a lot of different kinds
00:37:37
of businesses and a lot of them are very
00:37:39
subtle like you said some are network
00:37:41
effects some of them are the the
00:37:44
difficulty of producing physical world
00:37:46
products things like that so there's a
00:37:48
lot of different types of modes out
00:37:50
there and that is the key question as we
00:37:52
enter a world of let's call it digital
00:37:54
abundance
00:37:56
>> yeah the the network effect of Apple's
00:37:59
ecosystem and they have hardware, right?
00:38:01
And they've been just on that path of
00:38:03
making their own silicon that's
00:38:05
incredibly defensive and they have
00:38:06
brand, right? So that that is pretty
00:38:08
strong for Apple. And then you take
00:38:09
Tesla, the same thing. You got Elon
00:38:11
relentlessly innovating and it's hard
00:38:14
hardware stuff.
00:38:15
>> And then if you look at Meta and Google,
00:38:17
these are incredible brands with great
00:38:19
management teams.
00:38:19
>> My assumption and constantly innovating.
00:38:21
>> If I had to bet, I'm gonna bet that
00:38:23
brands go to zero.
00:38:25
>> Really?
00:38:25
>> Yeah. Because I think that when you can
00:38:28
make things that are as good or better
00:38:32
and you can make them in a cheaper,
00:38:34
faster, better way, people want that
00:38:36
abundance more than they want an
00:38:38
affiliation to a brand.
00:38:40
>> Example.
00:38:40
>> So the perfect example is actually what
00:38:42
Tesla did to BMW.
00:38:45
You know what Tesla did to Mercedes,
00:38:46
what Tesla did to Porsche, what BYD and
00:38:49
Gily have done to the car manufacturing
00:38:53
cycle in China. This is a fundamentally
00:38:56
cheaper, faster, better product. Yes,
00:38:59
it's got a great brand, but nobody's
00:39:01
going to pay a premium for these
00:39:02
products. The reason why Y has outsold
00:39:05
everything else is because the Model Y
00:39:06
is priced better and it's superior on
00:39:09
every operational dimension of
00:39:11
comparison. That's also true for the
00:39:14
cars in China. So, I think it's the
00:39:16
opposite. I think that brands and the
00:39:19
pricing power of brands other than maybe
00:39:21
premium luxury goods but even that's
00:39:24
eroding like look at the stock like I
00:39:26
don't know Nick show the stock chart of
00:39:28
LVMH or Ferrari this is not
00:39:32
a commentary on the quality of the
00:39:34
actual product but what this shows an
00:39:36
erosion of pricing power all of these
00:39:38
things are being eroded away
00:39:41
>> yeah this would be the value prop most
00:39:43
people in brands would just say value
00:39:45
propositioning So JetBlue is a value
00:39:47
brand, a Tesla Model Y, perfect example
00:39:50
of a value brand. And if you look at
00:39:52
Apple's recent cohort, what did they
00:39:53
focus on? The Apple MacBook
00:39:56
Neo, which is a value laptop, 600 bucks,
00:39:59
700 bucks. So they're even going down
00:40:01
market to try to capture that value. And
00:40:03
to your point, maybe the right word is
00:40:05
abundance. Like the brands that bring
00:40:07
abundance, that bring more to the table
00:40:10
than their competitors, and they are
00:40:12
able to bring more at the same unit cost
00:40:15
or less.
00:40:16
capture share. That's probably true.
00:40:19
>> You know what's interesting about that,
00:40:20
Shama, as we open up the aperture of
00:40:23
this, if you the one of the thesis we
00:40:25
talked about here a couple years ago, I
00:40:26
say, "Hey, what happens with AI
00:40:29
disruption, job disruption, etc. costs
00:40:31
coming down on cars with Model Y getting
00:40:35
cheaper, Cyber Cab coming in, BYD,
00:40:37
obviously if you go to any foreign
00:40:39
country, BYDs are everywhere and they
00:40:40
cost 15 grand. Then you look at Apple
00:40:43
making the Neo, that's a $600 laptop.
00:40:46
Everything getting cheaper seems to be
00:40:48
happening.
00:40:48
>> I just think it's very hard to know
00:40:49
which of these companies going to be
00:40:51
disrupted. A year ago on this pod, we
00:40:52
were saying that Google was going to be
00:40:54
toast or some of us were saying it
00:40:56
because it looked like,
00:40:57
>> okay, fine. But, you know, it looked to
00:40:59
us like Chad GBT was taking a massive
00:41:02
share from Google search and the Adwords
00:41:05
model was becoming obsolete. Now,
00:41:08
because of the success of Gemini and the
00:41:11
potential for personal digital
00:41:13
assistants, personal agents, I think
00:41:15
we're probably pretty bullish on that
00:41:17
company. And look at their stock chart.
00:41:19
It's reflects that. I mean, I think it's
00:41:20
doubled in the last year. Now, you take
00:41:22
something like Apple on the other hand,
00:41:24
and I'm not saying this is going to
00:41:25
happen, but I'll just throw out a
00:41:27
counterfactual, which is what if your
00:41:29
personal digital assistant or a personal
00:41:31
agent gets so good that you don't need
00:41:34
to check your phone, you just tell it
00:41:35
what to do, and you don't need the wall
00:41:37
of apps anymore. I mean, the way that
00:41:39
you call an Uber won't be to punch a few
00:41:42
buttons. You'll just tell it what to do.
00:41:44
So, you could imagine the phone
00:41:45
operating system getting disrupted if
00:41:48
the agents are good enough. you're on to
00:41:50
something huge because what you're
00:41:51
saying is I can confirm this to you
00:41:54
89 we sell enterprise software I'll tell
00:41:56
you three conversations with huge
00:41:57
enterprises asked exactly what you just
00:41:59
said and we jokingly called it
00:42:02
strangulation as a service which is they
00:42:04
all say the same thing what you just
00:42:05
said okay look get all this complicated
00:42:08
UI out of the way get all of these
00:42:10
products out of the way find a way to
00:42:12
create a shim where I can just write
00:42:14
what I need tell it what I need it to do
00:42:16
it deals with all this complexity in the
00:42:18
background I never want to see these
00:42:19
things ever again. And that's what
00:42:22
people want to your point. They want
00:42:24
they want to be able to like say, "Okay,
00:42:25
pay this with my Venmo or use my MX in
00:42:28
this situation or get me that flight in
00:42:30
some way and have all these wonderfully
00:42:33
smart agents do all the work behind the
00:42:35
scenes."
00:42:35
>> And that's actually tracks with exactly
00:42:37
what I've seen with Open Claw,
00:42:39
Perplexity Computer, Clawed Co-work.
00:42:42
instead of going to your notion, instead
00:42:43
of going into your Gmail, instead of
00:42:45
pulling up your calendar, you ask it,
00:42:47
"Hey, what's on my schedule this week?"
00:42:48
It brings it to you. So, you could you
00:42:50
could have a dumb, you know, flat
00:42:53
terminal, you know, chat interface on
00:42:56
the $100 device and it would do just as
00:42:58
good. Just to make the counterargument
00:43:00
against myself, I mean, yeah, even
00:43:02
though I think that you'll just
00:43:04
increasingly tell your agent what to do
00:43:06
instead of be clicking around or, you
00:43:08
know, touching the screen, you still
00:43:10
need a a dashboard or a user interface
00:43:13
to like check on it and just see
00:43:15
readouts of information. You still need
00:43:17
to be able to visualize it. Where is the
00:43:18
Uber that I've asked to be sent to me?
00:43:20
How far away is it? Still want to see it
00:43:22
on a map. So, I could imagine that even
00:43:24
if let's say, you know, Siri++ becomes
00:43:27
the dominant way of interacting with
00:43:29
your iPhone, you'll still want that
00:43:31
Apple user interface. Hard to say. I
00:43:34
mean, I see a lot of people out there
00:43:35
tweeting that Apple is brilliant for
00:43:39
kind of missing the whole AI wave and
00:43:40
not spending a lot of money on data
00:43:42
centers or capex.
00:43:44
And then there's other people who say,
00:43:46
well, wait a second, they're they're
00:43:47
missing critical capabilities.
00:43:50
Then there's a question of will they be
00:43:51
able to make deals for you know an AI
00:43:54
powered Siri? I don't know. I think this
00:43:56
is very hard to know at this point in
00:43:57
time.
00:43:58
>> All right. And we will be discussing all
00:44:00
these hard topics at liquidity
00:44:03
May 31st through June 3rd. Chimath some
00:44:07
big announcements here from you. You
00:44:09
have taken control. This is what happens
00:44:12
in the game of thrones known as the
00:44:14
allin
00:44:15
>> corporation LLC. the partnership. If we
00:44:19
partnership,
00:44:20
>> the chaotic partnership,
00:44:22
>> no mids there, it literally is chaos.
00:44:26
>> Wait, wait, wait. Why are we doing this
00:44:27
in California? I don't even think I can
00:44:28
I can go back to the state at this point
00:44:30
in time.
00:44:30
>> You can come for 48 hours make an
00:44:32
appearance.
00:44:35
>> Zero days in state at this point in
00:44:37
time.
00:44:37
>> I love it. I love it. You can do 48
00:44:38
hours. It's okay. Well, Gavin Newsome
00:44:40
might meet you at the airport. You could
00:44:42
get up at the airport.
00:44:43
>> The By the way, the airport is about 10
00:44:45
minutes from the venue. You could fly in
00:44:47
and fly out same day and you don't have
00:44:48
an overnight.
00:44:49
>> No overnight.
00:44:50
>> Is that how it's counted?
00:44:51
>> That's how it's counted as an overnight
00:44:52
in California.
00:44:53
>> Head on pillow. Yeah. You can come in on
00:44:55
Monday,
00:44:56
>> then fly, fly out to incline, sleep at,
00:44:59
you know, at our friend's house and then
00:45:01
fly back in on Tuesday and then fly back
00:45:02
out. You'll be fine.
00:45:03
>> Or you could fly to Vegas, do a
00:45:05
blackjack run overnight blackjack run,
00:45:07
and then you come back in the morning.
00:45:09
We just freshen up. Just
00:45:10
>> get your fresh shirt,
00:45:11
>> nice and rested.
00:45:12
>> Let me announce the next two speakers.
00:45:14
>> Oh my god, I'm so excited. But just for
00:45:16
background, Chimath comes into this
00:45:18
thing. Liquidity was this little
00:45:19
conference I did. It's now part of Allin
00:45:21
and I start setting up all the speakers.
00:45:23
And then Chimath goes, "Not good enough.
00:45:24
I'm not showing up unless I pick all
00:45:26
speakers." And you know what Freeberg
00:45:27
and I did? Well, finally this melon
00:45:30
farmer is going to actually do some
00:45:31
work. Great.
00:45:32
>> What did you call me? What did you call
00:45:34
>> melon farmer? It's a when um they play a
00:45:36
Quinton Tarantino movie on like TV.
00:45:41
Instead of saying Mother Effer, Samuel
00:45:45
Jackson says Melon Farmer.
00:45:46
>> Oh, I see. So this Melon Farmer took
00:45:50
unilateral control of the 10 speakers.
00:45:52
10 speaker slots. Very coveted.
00:45:54
>> Okay. So we've announced Dan Loe.
00:45:56
>> We've announced Sarah Frier.
00:45:58
>> Incredible.
00:45:58
>> And I'm very very very honored and
00:46:01
excited to announce that Bill Aman
00:46:04
and Andre Carpathy will also be speaking
00:46:07
at liquidity.
00:46:08
>> Four heavy hitters, six to go.
00:46:11
>> Two more goats on the roster of goats.
00:46:13
>> So goatated.
00:46:14
>> And we have a handful of other
00:46:17
>> Yeah, there might be some surprise
00:46:18
guests and unannounced guests show up.
00:46:20
>> Can't can't say DVD.
00:46:21
>> TB.
00:46:22
>> More announced soon, but we're making
00:46:24
>> Let's get Daario to show up. Daio, come
00:46:27
to the show. Come hang out with the
00:46:28
All-In Boys.
00:46:29
>> I'm really excited. I You know what
00:46:30
Andre is gonna do? What Andre agreed to
00:46:32
do is he's going to do like uh five or
00:46:34
10 minutes of slides on like the future
00:46:36
of the world with AI and then we'll do a
00:46:37
fireside.
00:46:38
>> Love it. That's going to be tremendous.
00:46:40
And you know, he's uh he's been on a
00:46:43
heater himself with his recursive uh
00:46:45
GitHub.
00:46:46
>> Oh, dude. Auto research is incredible.
00:46:47
And Bill Aman has a ton of points of
00:46:50
view on all of this stuff. So, we'll get
00:46:51
a lot of his thoughts. I want to give a
00:46:53
shout out to Freebrook because when we
00:46:54
did the interview with Jensen, he's
00:46:57
like, "Guys," I was like, I like had a
00:46:59
glass of wine on a Sunday. I like coded
00:47:01
a replacement to my HR system. I asked
00:47:04
all my team to do and I was like, "Oh my
00:47:06
god, this is incredible." So I asked our
00:47:08
folks, I'm like, "Hey, I don't like the
00:47:10
website. This is 8090." And the next day
00:47:12
they're like, "Yeah, we vibe coded a new
00:47:14
one. We'll have it up." And then I'm
00:47:16
like, "Well, do you like the CTAs and
00:47:18
how it's doing?" They're like, "No, no,
00:47:19
no. We put it into auto research and we
00:47:21
doubled the click-through rate. And I
00:47:23
was like,
00:47:24
to Freeberg's point, this was this would
00:47:27
have been many man months, tens of
00:47:30
people,
00:47:32
and instead all these recipes, by the
00:47:34
way, in these playbooks, you know, like
00:47:36
the person that runs growth at OpenAI
00:47:38
publishes his recipes.
00:47:40
>> Yeah. You've been clawed. That's it.
00:47:42
You've been oneshotted.
00:47:43
>> It's really incredible. This is why like
00:47:45
I I mean, you commented on my post
00:47:47
yesterday. I wake up every day and my
00:47:48
head spins. I'm like, what is what is
00:47:50
going on in the world?
00:47:51
>> It's crazy. Like, every day feels like a
00:47:54
new era right now.
00:47:55
>> Yeah.
00:47:56
>> And
00:47:56
>> it's it's disoriented
00:47:58
>> because because I I think what's really
00:48:00
interesting is you pull what would have
00:48:02
normally been something that's call it a
00:48:04
period of time out, like a year out or
00:48:06
two years out, and you can pull it in
00:48:07
and say, I can get that done in 3 days,
00:48:09
and I don't need to hire people. All the
00:48:11
sequencing and staging that would
00:48:13
normally go into accomplishing something
00:48:15
has been reduced, and then the time
00:48:17
rushes in. So all your ideas rush into
00:48:19
you and they're all like immediately
00:48:21
accessible and that's why it's like so
00:48:23
every day it's like oh my god like
00:48:25
that is I I have I bought the domain
00:48:28
name annotated.com like 15 years ago for
00:48:31
four grand and I wanted to create a
00:48:33
service like a bookmark service where
00:48:35
you like highlight a paragraph from the
00:48:36
New York Times and then you write your
00:48:38
comments on it saves it and you
00:48:39
basically have this service and I was
00:48:41
talking to some developer about making
00:48:43
it and I literally vibecoded it and the
00:48:44
Chrome extension this past week and I
00:48:47
was like, "Okay, I've been sitting on
00:48:49
this domain and project for like 15
00:48:52
years and I did it in a weekend."
00:48:55
>> Isn't it crazy?
00:48:56
>> Weird. It's like very weird.
00:48:58
>> This makes it feel like a simulation to
00:49:00
me that everything can just manifest
00:49:02
itself. It's the Star Trek version of
00:49:04
the world. Like remember the replicator
00:49:06
sacks where you just say Earl Grey tea
00:49:08
at this temperature? That's happening in
00:49:10
business now. You're like CRM system,
00:49:12
buildan annotated.com, build this new
00:49:14
website for 8090. Boop. and it's a
00:49:17
replicator just gives it to you. It's
00:49:19
very strange.
00:49:20
>> I've only felt this
00:49:23
feeling twice. Once I was on the outside
00:49:26
looking in. I was a derivatives trader
00:49:29
in Toronto. I was looking at the first
00:49:32
dot
00:49:33
wave and I was like, I got to be a part
00:49:35
of this. How do I get to be a part of
00:49:38
this? And I got a job in at Win Amp and
00:49:42
kind of the rest was history. Win him
00:49:45
the original iTunes.
00:49:46
>> But I missed the wave financially. It
00:49:48
didn't do anything for me. But I was in
00:49:50
the right place. You got to the beach
00:49:53
with a surfboard. That's all that
00:49:54
matters.
00:49:55
>> Then the second wave I crushed and the
00:49:57
move to mobile and social. But this wave
00:49:59
feels like a hundred times bigger than
00:50:01
that.
00:50:02
>> It's a tsunami by comparison. I I I had
00:50:04
three of these.
00:50:05
>> It's probably what it's like to be at
00:50:07
Nazareth in Portugal. You guys ever see
00:50:09
those clips of like that crazy place
00:50:11
where like there's like 100 foot waves,
00:50:12
>> 100 foot waves or whatever and you're
00:50:14
just towed in and you're like, "Okay,
00:50:16
let's just
00:50:16
>> Here we go. This could end one of two
00:50:18
ways.
00:50:19
>> Let it go. Glory
00:50:20
>> or goodbye."
00:50:22
>> No, I It was like literally when I first
00:50:23
saw a PC,
00:50:25
when I first got on the internet, and
00:50:27
then saw Mosaic. Like those two moments
00:50:29
in the early 90s and then seeing the
00:50:31
iPhone. I think that like those three
00:50:33
moment
00:50:33
>> was very special.
00:50:35
>> All right, rough week for Zach. Two
00:50:37
verdicts went against Meta in two days.
00:50:41
They were first found liable for
00:50:43
allowing child predators to access
00:50:45
minors on Facebook and Instagram. New
00:50:47
Mexico jury ordered Meta to pay 375
00:50:50
million in damages. The AG's office
00:50:52
there ran an undercover investigation.
00:50:53
They created fake child profiles.
00:50:55
Facebook, Instagram, these accounts were
00:50:56
contacted by predators. People showed
00:50:59
up, yada yada yada. A whistleblower and
00:51:02
former Meta engineer testified that his
00:51:04
own 14-year-old daughter received sexual
00:51:06
solicitations on Instagram. Then on
00:51:08
Wednesday, an LA jury found Meta and
00:51:10
YouTube negligent for designing
00:51:12
addictive platforms that harmed a young
00:51:15
user's mental health. Basically, the
00:51:17
plaintiff in this case, Saxs, said they
00:51:20
started using YouTube at age six and
00:51:22
Instagram at age nine. and she testified
00:51:24
that features like notifications,
00:51:26
algorithms made the app so addictive
00:51:28
that it caused depression and anxiety
00:51:31
through that compulsive use.
00:51:34
You have some thoughts on this and we we
00:51:36
had um Jonathan hate, right? Didn't we
00:51:38
do an interview together? Wasn't that
00:51:39
one of the first joint interviews? Free
00:51:41
>> you and I did. You and I did. Yeah.
00:51:43
>> Incredible.
00:51:44
>> Uh these things are obvious. He he had a
00:51:47
very important point which is to try and
00:51:48
keep kids off
00:51:50
>> cell phones and social media until
00:51:52
they're 16. and he was kind of
00:51:53
cheerleading this this verdict. But, you
00:51:55
know, I'll take a little bit of a
00:51:58
contrarian to the popular kind of
00:51:59
sentiment on this and I'll just talk
00:52:02
broadly about this idea of tort
00:52:03
litigation. Tort litigation, you know,
00:52:05
costs our economy $900 billion a year in
00:52:08
the United States. 900 billion a year.
00:52:11
That's how much is spent on the
00:52:12
litigation costs, the settlements, the
00:52:14
judgments. It's 3% of GDP and it's
00:52:16
growing roughly 10% per year. And these
00:52:19
civil penalties decided by juries like
00:52:22
they're, you know, going against big
00:52:23
companies like Meta and YouTube, but
00:52:25
it's also food companies, restaurants,
00:52:27
everything. Anytime there's a window to
00:52:29
sue someone and and extract uh value
00:52:33
from them, tort firms are all over them.
00:52:35
You know, it's called the tort tax now
00:52:37
in America. It's not just losses paid by
00:52:39
the companies because fundamentally when
00:52:41
a big company pays out these tort taxes,
00:52:43
they're going to invest less. There's
00:52:45
less R&D, less product development.
00:52:46
Costs stay high. there's new fewer new
00:52:49
product launches and there's all these
00:52:51
crazy restrictions on stuff. So, you
00:52:53
know, look, I I agree. Social media
00:52:55
causes immense harm, particularly causes
00:52:57
harm for kids. Kids should not be on
00:53:00
social media until they're 16.
00:53:02
Absolutely agree. Maybe adults shouldn't
00:53:04
either be on, you know, as adults. But
00:53:06
fundamentally, I think there's an
00:53:08
important question that we often ignore,
00:53:09
which is who is fundamentally
00:53:11
responsible for that harm? You know,
00:53:13
should the sugar beat and sugarcane
00:53:15
farmers be responsible for diabetes in
00:53:17
America? Should the soda companies be
00:53:19
responsible? The retailers selling the
00:53:20
soda, the FDA for not stopping it all?
00:53:23
And fundamentally, I think we have to
00:53:25
ask the question, what role does
00:53:26
individual choice play and individual
00:53:29
responsibility play in this equation? If
00:53:31
everything is a liability, if anything
00:53:34
any I I think we have to take personal
00:53:36
responsibility. I think the parents that
00:53:38
are absent taking care of their children
00:53:40
are responsible for harm to their kids.
00:53:42
You shouldn't let your kid play with a
00:53:44
gun. You shouldn't let your kid go to
00:53:46
some sketchy neighborhood after hours by
00:53:48
themselves. You shouldn't let your kid
00:53:50
play video games 100 hours a week. You
00:53:53
shouldn't let your kids eat nothing but
00:53:54
soda and potato chips. You have
00:53:57
responsibility as a parent. And I think
00:53:58
parents should keep kids off screens and
00:54:00
keep kids off social media. Once the
00:54:02
harms are known of excess use or the
00:54:04
harms are known of exposure to this sort
00:54:06
of thing, I think there's responsibility
00:54:08
that sits with the parents.
00:54:09
>> What about things like tobacco or
00:54:11
processed food or
00:54:12
>> Yeah, this is the key.
00:54:14
Look, I mean the same the same is true
00:54:16
of alcohol. I mean dude, like alcohol is
00:54:17
terrible for you. There's nothing good
00:54:19
about alcohol. But I think I should have
00:54:21
a choice on whether or not I want to
00:54:22
consume alcohol. Whether or not I
00:54:24
consume tobacco, whether or not I
00:54:26
consume processed goods and the
00:54:27
recognition that it's bad for you should
00:54:29
be publicized by the government should
00:54:30
be publicized by the government.
00:54:31
>> Right. So if I had to summarize what you
00:54:33
would say is product liability law makes
00:54:36
no sense. There should be human
00:54:38
liability and human responsibility
00:54:40
expectations in a society.
00:54:41
>> We never talk about responsibility. We
00:54:43
always talk about where the government
00:54:45
failed us and where these companies us
00:54:47
and we never talk about what did we
00:54:49
individually do wrong. How did I
00:54:51
individually choose to eat a hundred
00:54:52
[ __ ] sodas a week? How did I
00:54:54
individually choose to get my kids
00:54:55
addicted to social media? Where the was
00:54:57
I as a parent? Like we don't talk about
00:54:59
our responsibility. And by the way, this
00:55:02
fundamentally addresses this point about
00:55:04
human agency which I think is more
00:55:05
critical in this era than ever because
00:55:07
AI is going to flood us with everything
00:55:10
all the time non-stop. What we choose to
00:55:13
do in a world where we're already
00:55:14
getting everything and how we choose to
00:55:16
not take everything that's being offered
00:55:18
to us, I think is a critical part of
00:55:19
what's going to distinguish human
00:55:21
success from human failure. And it's
00:55:23
going to become more apparent in the
00:55:24
future. And not everything is about
00:55:26
liability and not everything is about
00:55:27
the government failing us. It's about
00:55:28
people making choices. And we don't talk
00:55:30
about it.
00:55:32
>> On the counter, I I I agree. Personal
00:55:34
choice super important. What you're
00:55:37
probably leaving out here, uh, which
00:55:39
you're definitely leaving out here, is
00:55:41
when these companies know they're doing
00:55:43
something damaging and they do it
00:55:45
anyway. That was the key to the RJR
00:55:49
whiskey. What about what about a a
00:55:52
whiskey company selling whiskey to an
00:55:54
alcoholic?
00:55:55
>> So, what about a potato chip company
00:55:57
selling potato chips to an obese person?
00:56:00
>> I'll put those two aside because I don't
00:56:01
think we've seen major cases about that.
00:56:03
But I will say the auto industry knew
00:56:05
for a long time about seat belts. You
00:56:07
remember that? And they didn't deploy
00:56:10
them. RJR Nabiscoco, they knew that
00:56:14
these were addictive and they designed
00:56:17
the cigarettes to become more addictive
00:56:18
and they didn't tell people about the
00:56:21
health risks. Asbestos, same thing. Lead
00:56:24
paint, same thing. This has happened
00:56:26
over and over again where corporations
00:56:29
subvert the release of information to
00:56:31
make additional profits. So the question
00:56:32
here with Facebook is did they know how
00:56:36
addicted these were? Did they know kids
00:56:38
were being assaulted sexually and and
00:56:40
they could have done something about it
00:56:42
or didn't they? And so agree that
00:56:44
there's
00:56:45
>> that's absolutely true. The kids being
00:56:47
assaulted absolutely true. Not not
00:56:50
releasing information about the level of
00:56:52
addiction if they had it. That's
00:56:53
certainly bad. Should the product be
00:56:55
legal? Number one. And if the product is
00:56:57
legal, who's responsible for using it?
00:56:59
>> You know, and and where do we draw the
00:57:00
line? And if we don't want people to
00:57:02
make choice, then we shouldn't put it
00:57:03
out there.
00:57:04
>> But if you as a corporation know it's
00:57:06
dangerous and then you lean into that,
00:57:08
>> if you know, as is the case with
00:57:11
Facebook, that this is super addicting,
00:57:13
super damaging to young girls, and then
00:57:15
you lean into making it more addictive
00:57:17
and you don't put safeguards in place
00:57:19
and you can prove that like RJR, like as
00:57:23
like on whiskey. What's the What
00:57:25
safeguards does a whiskey company put?
00:57:27
What safe What safeg guards does a
00:57:29
casino put?
00:57:30
>> Hold on. you asked the question age is
00:57:32
one
00:57:33
second after age which is what we're
00:57:36
really talking about with kids and
00:57:38
Jonathan hate would agree that they
00:57:39
shouldn't be using this until the 16 so
00:57:41
I think that's a perfect analogy uh age
00:57:43
gating and then labeling and then if
00:57:46
they know of something that's really
00:57:48
damaging releasing that. So there was a
00:57:50
whole thing about alcohol in pregnancy
00:57:52
and they covered up in the alcohol
00:57:54
industry or didn't disclose exactly how
00:57:56
damaging it was to drink alcohol on a on
00:57:59
a fetus on a developing fetus and then
00:58:01
remember all those signs went up in bars
00:58:03
in the 70s and 80s that was directly
00:58:05
because of that. So labeling information
00:58:09
and agegating would be the logical
00:58:10
things to do for social media and that's
00:58:12
what's happened the best. So that that's
00:58:13
the answer to your question. I don't
00:58:14
know about age gating, but I think
00:58:15
informing parents about the risks is
00:58:17
fine. Like that should be a
00:58:18
responsibility, but you know that the
00:58:20
tort lawyers are one of the largest
00:58:22
donors in political elections in the
00:58:24
United States. They donate largely to
00:58:26
Democrats in local elections and
00:58:27
Republicans in national elections and
00:58:29
then they are the largest donor class to
00:58:31
elections of judges. And it's it's a
00:58:34
business and I think we don't talk
00:58:35
enough about the business of court law
00:58:38
um the business of litigation in this
00:58:40
country and we often ignore this
00:58:42
question about choice and I don't want
00:58:43
to live in a world JL where the
00:58:45
government and companies are telling me
00:58:46
what to do and what not to do to live my
00:58:48
life etc.
00:58:49
>> Exactly. Jump in here personal freedom
00:58:53
and personal um responsibility versus
00:58:57
hey corporations maybe knowingly
00:58:59
>> parental responsibility Jal
00:59:01
responsibility in there as well. Yeah.
00:59:02
What do you think's next?
00:59:04
>> Well, look, there's no question that the
00:59:05
trial lawyers want to turn Meta into RJR
00:59:08
and Nabiscoco and the cigarette
00:59:10
companies and try to fit their fact
00:59:11
pattern around that. I just think that
00:59:13
the activity is fundamentally different
00:59:15
than smoking. I mean, smoking is
00:59:16
manifestally harmful to you regardless
00:59:19
of what your age is. And the only reason
00:59:21
we allow it is because of assumption of
00:59:23
risk. It's a free country. I think in
00:59:25
the case of social networking, it's much
00:59:27
more unclear what the harms are and what
00:59:30
the benefits are. And I think it's much
00:59:32
more subjective and it's much more of a
00:59:34
personal choice for adults and also for
00:59:37
parents. Freeberg, I mean, one area
00:59:38
where I disagree with you a little bit
00:59:39
is you said the harms of this are
00:59:41
immense and well known and understood.
00:59:44
If that's the case, then let's just ban
00:59:46
it for under 13 or under 16 or whatever
00:59:49
it is. But I don't think that's the
00:59:51
case. And because it's unclear, I think
00:59:54
it's up to parents to decide what's
00:59:56
appropriate for their families. And at
00:59:58
the end of the day, I think the right
00:59:59
way to deal with this is parental
01:00:00
empowerment. You give parents the
01:00:02
controls to set screen time limits or to
01:00:05
decide what apps their kids install.
01:00:08
Now, the debate has now moved over to AI
01:00:10
apps. And there's a lot of parents
01:00:12
groups that want to ban kids or
01:00:15
teenagers from be able to use AI chat
01:00:18
apps because there was a couple of cases
01:00:20
of self harm. Now, my reaction to that
01:00:23
is, look, I've got a 10-year-old and if
01:00:27
he starts using Chat GBT to get answers
01:00:29
to questions, I would consider that to
01:00:31
be a good thing. I mean, I'll keep an
01:00:32
eye on the usage, but I want him to be
01:00:34
an AI native. I want him to be able to
01:00:37
do research, you know, I want him to be
01:00:39
able to know how to use these tools. I
01:00:40
want him to get the right skills to be
01:00:43
successful in the 21st century. In
01:00:45
China, they're incorporating AI into K
01:00:46
through2 education. Are we going to ban
01:00:48
it for our kids and teenagers? I think
01:00:51
that'd be a terrible mistake. So, I
01:00:53
think when it comes to AI and AI chat
01:00:55
apps, it has to be up to the parents
01:00:57
because there's too much manifest good
01:01:00
that can come out of kids learning how
01:01:02
to use these apps. Maybe social
01:01:04
networking is in a different category.
01:01:06
But I got to say, I do think that the
01:01:08
harms have been exaggerated because the
01:01:10
trial lawyers have an incentive. And
01:01:12
just to give you some facts about that
01:01:13
LA case, so just so you understand,
01:01:17
in this LA case, they were sued by a
01:01:20
20-year-old woman who claimed that she
01:01:23
became depressed
01:01:25
because of using social media, and she
01:01:28
apparently suffered body image issues
01:01:31
because of social media, and she was
01:01:33
able to win this judgment for millions
01:01:35
of dollars. Look, I think there's big
01:01:37
causation problems with that case. In in
01:01:39
the case of that pliff, the evidence
01:01:41
showed that she came from an abusive
01:01:43
home. Her father had abandoned her. Her
01:01:45
own mother had body shamed her. So, it
01:01:47
was very unclear where her body image
01:01:49
issues were coming from. I think it's
01:01:51
possible that social media contributes
01:01:53
to them. Or it could just be that social
01:01:55
media was a scapegoat. I also got to say
01:01:57
that I do think it's it's a dangerous
01:01:59
precedent. I mean, are we going to allow
01:02:02
plaintiffs, lawyers to sue Spotify
01:02:05
because you created a playlist of sad
01:02:09
music
01:02:09
>> and that music contributed to your
01:02:11
emotional distress? I mean, that's kind
01:02:12
of what we're saying. You got to
01:02:14
remember these are free services that
01:02:16
people have a choice whether to use them
01:02:18
or not. And if social media is making
01:02:20
you feel bad, if listening to the wrong
01:02:22
playlist is making you feel bad, then
01:02:24
stop doing it. But what's going on here,
01:02:26
I think, is the trial lawyers are trying
01:02:29
to create the next big tobacco and their
01:02:31
goal is to try and sue these companies
01:02:33
into oblivion. And I don't think that's
01:02:36
the right answer either.
01:02:37
>> The consensus, Chimoth, is kids who use
01:02:40
this two or three hours a day. this has
01:02:42
been across many um survey uh many
01:02:46
research studies is that it's massively
01:02:49
correlated with depression and anxiety
01:02:53
um and eating disorders specifically in
01:02:55
young girls like if you get to two or
01:02:58
three hours a day of this. So that
01:03:00
correlation's I think pretty well
01:03:02
established. you pretty famously said,
01:03:04
"Hey," and you worked at Facebook, so
01:03:06
you saw this coming, but I think at
01:03:08
Stanford 27 2018 time frame, you
01:03:11
actually had some comments on this. We
01:03:13
could either play the clip or you could
01:03:14
just describe it, I guess.
01:03:15
>> I said what was obvious to me then. I
01:03:17
lost a lot of friends at Facebook when I
01:03:19
said it, unfortunately. But essentially
01:03:21
what I said was, I don't let my kids use
01:03:24
it, and I don't think that this is a
01:03:25
constructive part of a developing
01:03:28
child's diet.
01:03:30
I do think that Sax is right that I view
01:03:32
AI chat differently. There are different
01:03:35
guard rails that are required so that if
01:03:37
you go down some sort of you know dark
01:03:39
corner
01:03:41
but that's possible to understand and
01:03:44
the product is architected in a way to
01:03:47
create culde-sacs. So if that if you're
01:03:49
thinking about self harm or you're
01:03:50
thinking about these other things,
01:03:52
social media is very different. It's an
01:03:53
incredibly fast twitch algorithm and
01:03:57
that's the optimization and until the
01:03:59
incentive for that optimization changes,
01:04:01
these outcomes will continue to
01:04:02
compound. I think the interesting thing
01:04:05
is that the LA lawsuit was an individual
01:04:08
lawsuit. The young woman I think was
01:04:12
awarded three or $6 million.
01:04:14
>> Yeah. The other one though that they
01:04:16
lost this week was in New Mexico and
01:04:18
that was for $375 million
01:04:21
and that was more around I think child
01:04:23
exploitation.
01:04:24
>> Yeah.
01:04:24
>> The thing that I'll note is that these
01:04:27
trial lawyers which I'm not a fan of
01:04:29
either have been trying as Sax said and
01:04:31
as Freebrook said to make these folks a
01:04:33
target because there's so much money on
01:04:34
the line and they've been batted back
01:04:37
pretty successfully. But this was the
01:04:40
first time where they were able to
01:04:42
navigate the section 230 protections
01:04:45
that Facebook and Google have typically
01:04:48
used to protect themselves because
01:04:49
Google was a part of the LA lawsuit
01:04:54
and they were able to go down the
01:04:55
pathway of product liability language.
01:04:57
Now to Freeberg's point, I think he's
01:05:00
generally right. I think it's my
01:05:02
responsibility as a parent to take care
01:05:03
of myself and my children.
01:05:06
But to the extent we don't change these
01:05:09
product liability laws that are on the
01:05:11
books, I think the door has been opened
01:05:13
and a map has been drawn, which is this
01:05:17
is how you navigate around section 230
01:05:20
and you can get a decisive lawsuit in
01:05:23
your favor against these large companies
01:05:27
with enormous cash flows. And so I
01:05:31
expect that this will be a death by a
01:05:34
thousand cuts kind of scenario where
01:05:37
folks are just going to rally around
01:05:39
this. I don't think it's right, but I do
01:05:41
think that that's the rational reaction
01:05:43
to what just happened. I mean, look, $6
01:05:45
million to an individual 20-year-old
01:05:47
girl or $375 million. These are huge
01:05:51
numbers. And the reality is that I think
01:05:53
we've opened the floodgates. I do think
01:05:55
there should be a better response and I
01:05:57
do think that parents should take a lot
01:05:59
more responsibility. But I do hope that
01:06:02
these products allow us a kill switch
01:06:05
for when our kids are under the ages of
01:06:07
16. Frankly, I would love a kill switch
01:06:09
under the age of 18. And there has to be
01:06:12
simpler ways to age validate. You know,
01:06:14
we used to work around this thing called
01:06:15
Kappa compliance. Kappa's [ __ ] It's
01:06:18
a nothing burger. Like a six-year-old
01:06:20
can vibe code their way around Kappa. So
01:06:23
it doesn't do anything then. So age
01:06:25
verification is completely broken. It's
01:06:28
harder to get age verified for a porn
01:06:29
site than it is to get unageverified for
01:06:31
Facebook.
01:06:32
>> Say more Chamoth. How did you get around
01:06:33
it?
01:06:34
>> Yeah. How do you know that? How do you
01:06:35
know that?
01:06:36
>> That's just for uploading. You were
01:06:38
uploading at the time. This is not your
01:06:40
only proof of proof of a
01:06:44
>> we actually deal with this in our new AI
01:06:47
framework that the administration
01:06:49
released last week because the number
01:06:51
one issue is around online child safety
01:06:55
which I think again is mostly about
01:06:57
social networking but it touches on AI
01:06:59
and so it's kind of gotten lumped
01:07:01
together and I think we're I mean I'm
01:07:04
referring now to the White House has
01:07:05
said it's willing to support some form
01:07:07
of age assurance technology and parental
01:07:11
controls. So, look, I think there needs
01:07:13
to be a conversation around how true it
01:07:15
is that social networking is just a
01:07:17
fountain of ills for young people. If
01:07:19
that's true, then why wouldn't you just
01:07:22
disallow it, right? But if if it's
01:07:24
unclear, and I think it's more on the
01:07:26
side of unclear, that's just me.
01:07:28
>> Uh then it's up to the parents and what
01:07:30
you want to do is again have the age
01:07:32
assurance technology and the parental
01:07:33
controls and let the parents decide. I
01:07:35
don't think that social media is net
01:07:39
badly
01:07:40
speaking, but I will tell you that if my
01:07:44
child uses it for two or three hours a
01:07:46
day for multiple days in a row, they
01:07:48
become [ __ ]
01:07:49
>> Yes, they act weird.
01:07:51
>> And so I don't know what to tell you
01:07:53
except I would love a kill switch until
01:07:55
these kids turn 18 on all these
01:07:57
products.
01:07:58
>> This has to do with dopamine, by the
01:08:00
way. these. What I do instead is I use
01:08:02
Apple's Family Sharing, but it's hard.
01:08:05
And I'll tell you how why it's hard. Our
01:08:07
schools give our kids Chromebooks and
01:08:11
say, "Use these Chromebooks because
01:08:13
that's the way you're going to access
01:08:15
your LMS, your information, your
01:08:17
content, your homework." And you know
01:08:19
what's an integral part of that?
01:08:20
YouTube. And then now it becomes a back
01:08:23
door. So even when we take the devices,
01:08:25
I take my devices, we lock them down.
01:08:27
You can't bring them to your room. the
01:08:28
whole thing.
01:08:30
But if they need to do homework and I
01:08:32
catch them in a little break, I'm like,
01:08:35
"What are you doing on YouTube?" They're
01:08:36
looking at shorts,
01:08:38
>> of course.
01:08:38
>> That's what they do.
01:08:39
>> Yes.
01:08:40
>> And so it is a whack-a-mole problem for
01:08:42
parents.
01:08:44
>> I think uh I'll I'll just give myself
01:08:46
the final word here on behalf of the
01:08:48
group. It's obviously addicting.
01:08:50
It's obviously the industry has not
01:08:52
policed itself, nor would they police
01:08:54
themselves because they want to get
01:08:56
people addicted now. Also, they're
01:08:57
addicted when they're adults. The rest
01:08:59
of the world has realized this.
01:09:01
Australia now has 16 enforced. Malaysia
01:09:04
16. Other countries are right behind
01:09:06
them. Spain, Germany, UK. And what
01:09:09
should happen here in the United States
01:09:10
is this should be done with the handset
01:09:13
manufacturers. If Android and Apple
01:09:17
showed leadership with Facebook and
01:09:20
said, "We're going to by default when
01:09:22
you buy a phone, you're going to have to
01:09:24
age verify, you know, kids under a
01:09:26
certain age," which you kind of do
01:09:28
already when you create a family plan
01:09:29
like it did. We could solve this
01:09:31
problem. And then parents would opt in
01:09:34
to giving their kids access to this. No
01:09:36
good comes out of kids under 16 using
01:09:39
social media. And Jonathan Hay in our
01:09:42
interview said you should be putting
01:09:44
phone lockers in and our school is doing
01:09:46
that. Other schools are doing it. It is
01:09:48
the greatest thing ever. Kids complain
01:09:50
and then they love it.
01:09:51
>> I I don't think it's that great. I'll
01:09:52
tell you why. Our school does the phone
01:09:54
condoms. I don't know what they're
01:09:55
called, but like you put them in the
01:09:57
>> the bags and you whatever.
01:10:00
The real problem is that there's
01:10:01
enormous social pressure when you're in
01:10:03
high school to use these products. I'll
01:10:04
give you a specific example. Our rule is
01:10:07
you cannot get Instagram or Tik Tok
01:10:09
until you're 16. I would love it to be
01:10:12
18, but we all agreed 16 and we're able
01:10:15
to maintain that rule. But then high
01:10:17
school comes around and I have two kids
01:10:19
in high school now. Oh, we need
01:10:20
Snapchat. No, you don't. No, yes we do.
01:10:23
No, you don't. Okay, great. We'll have
01:10:25
no friends. Thanks a lot. We'll just sit
01:10:26
here in our room dark and there and you
01:10:28
know, there's an enormous amount of
01:10:30
social pressure. So, when you talk to
01:10:31
the other parents, it's like, hey guys,
01:10:32
can we all agree that we don't need
01:10:34
Snapchat? You just can't get uniform
01:10:36
agreement across all parents because
01:10:38
everybody views this problem
01:10:39
differently. And so I had to change the
01:10:42
rule. Now when you get to high school,
01:10:44
you're allowed Snapchat but not
01:10:45
Instagram. And you know when you're 16,
01:10:47
you get Instagram. And that's the best
01:10:49
we could do and Tik Tok and all that
01:10:51
stuff. But
01:10:51
>> yeah,
01:10:52
>> it's an impossible task for a parent.
01:10:54
>> I find it indispensable to be able to do
01:10:56
iMessage with my kids. And also I like
01:10:59
the location awareness. You can track
01:11:02
location. You compare iMessage those
01:11:05
features to Snap. Is it really that
01:11:07
different? I don't think it's that
01:11:08
different.
01:11:09
>> Well, you really you gonna ban Snap.
01:11:13
>> No, I can't. My point is I had to give
01:11:14
them Snap. But you're right. I was like,
01:11:16
"Why can't you use iMessage?" They're
01:11:17
like, "Dad, don't be a loser."
01:11:19
>> Yeah.
01:11:21
>> No, I signed for a Snap account just so
01:11:22
I could message them where they are,
01:11:24
where they're messaging. I prefer if
01:11:25
they were just on iMessage, but anyway,
01:11:27
that's the reality. So, you're talking
01:11:30
about banning all these things, but
01:11:31
again, I think you're on a slippery
01:11:33
slope.
01:11:34
>> And I gota I got
01:11:35
>> No, I'm saying me as a parent banning.
01:11:37
No, I'm saying me as a parent.
01:11:38
>> Yeah, you're setting guidelines
01:11:40
>> in my house. This is the rules. When you
01:11:41
become a freshman, I'll give you Snap
01:11:44
and when you're 16, I'll give you
01:11:45
Instagram and Tik Tok. Otherwise, shut
01:11:47
the [ __ ] up. Put your head down.
01:11:49
>> By the way, I also have a tip for
01:11:50
parents. You know this issue with like
01:11:53
kids in headphones where they want to be
01:11:55
listening to stuff all the time and you
01:11:56
have kids walking around like zombies
01:11:58
with headphones. I replaced their
01:11:59
headphones with over the these brand
01:12:03
called shocks or something. It goes over
01:12:04
your ear but your ears open so you can
01:12:07
listen to an audio book or music and
01:12:09
still be able to talk. And so since we
01:12:12
did that it's like much less drama
01:12:14
around the overthe-ear headphone issue.
01:12:18
>> Well again it sounds like you're
01:12:20
exercising parental supervision. Does it
01:12:22
really make sense that later when your
01:12:26
kid turns 20,
01:12:29
they could sue these companies for
01:12:31
>> Well, to your point for emotional
01:12:33
distress. That doesn't make sense.
01:12:35
>> You're right. Because to your point,
01:12:36
that girl said she was using it from the
01:12:38
age of like six and 10.
01:12:40
>> Yeah, one product six and the That's
01:12:42
crazy.
01:12:43
>> Crazy.
01:12:44
>> I agree with you, Sax. That's crazy.
01:12:45
Where were the parents? All right, we
01:12:48
got time for I think one more topic. Be
01:12:51
amazing.
01:12:53
President Donald Trump, the 47th
01:12:56
president of the United States,
01:12:57
announced his council of adviserss on
01:12:59
science and technology. It's called
01:13:01
Pcast
01:13:04
Saxs. You have now, am I correct in
01:13:09
saying moved from the zar of crypto and
01:13:12
AI to now the leader of PCAS? Is that
01:13:15
the correct way to frame this? Well, the
01:13:16
president has appointed me to be a
01:13:18
member of his council of adviserss on
01:13:21
science technology and to co-chair it
01:13:23
along with Michael Katzios who's the
01:13:24
director of OSTP. I'm still a AI
01:13:26
adviser, but I do it on behalf of PCAST
01:13:30
now.
01:13:31
>> So, you remember last year I was an SGE.
01:13:33
We got up to 130 days. I use that time
01:13:36
up and the president appointed me to
01:13:37
this new role allows me to continue
01:13:39
being a technology adviser in fact on a
01:13:41
wider range of issues. So before it was
01:13:44
AI and crypto. Now it's whatever PCAST
01:13:47
wants to study or talk about or make
01:13:49
recommendations. I think in addition to
01:13:51
AI, other areas that are interesting are
01:13:54
nuclear power, quantum computing,
01:13:57
advanced semiconductors, all these
01:13:58
different areas. And I think we've got
01:14:00
some biotech. Thank you, Freeberg. And
01:14:03
we have some incredible people who are
01:14:04
now on PCAST to kind of run with this.
01:14:07
>> Yeah, I'm looking at it. Mark Andre,
01:14:09
Sergey, Michael Dell,
01:14:12
Larry Ellison, David Freedberg, Jensen,
01:14:16
uh, Lisa Sue, Mark Zuckerberg, and some
01:14:20
other folks there that, uh, maybe not
01:14:23
as, uh, recognizable by the audience.
01:14:26
How were they selected? Saxs, the only
01:14:28
criticism I've seen of this is lots of
01:14:30
business leaders, great, lots of
01:14:33
technology, great, but maybe a little
01:14:35
light on the scientists.
01:14:37
>> Well, I don't know. We have people
01:14:38
who've won a Nobel Prize for physics on
01:14:41
there. We're talking about people who
01:14:43
are experts in like I mentioned quantum
01:14:45
computing, fusion, nuclear,
01:14:49
>> biotech, pretty much everything across
01:14:50
the board. I would say that one
01:14:52
difference between this PCAST and
01:14:54
previous ones is you have more doers,
01:14:56
more builders, people who've actually
01:14:58
created products or companies. We think
01:15:01
that's a good thing. Why would it be a
01:15:03
bad thing?
01:15:04
I mean, is it a bad thing that Mark
01:15:06
Andre invented the first internet
01:15:08
browser or Jensen invented the GPU? If
01:15:11
you're going to make recommendations
01:15:13
about advanced semiconductors, don't you
01:15:14
want to have someone who actually
01:15:16
invented some of the key products in the
01:15:17
space?
01:15:17
>> The big question of course, Sax, is as
01:15:20
we've seen here with Science Corner is
01:15:22
how do you plan on staying awake for
01:15:23
these meetings if it's going to be like
01:15:25
all science is usually when you take
01:15:27
your your bio brain?
01:15:28
>> Well, it's it's science and technology.
01:15:31
Got it. So you'll be
01:15:32
>> So I'm going to focus on the tech stuff.
01:15:34
>> Got it.
01:15:34
>> And then we got Freeberg to focus on the
01:15:36
science stuff.
01:15:36
>> Got it. So Freeberg, this is incredible.
01:15:38
You've joined President Trump's
01:15:42
administration. Now
01:15:43
>> I will say I'm honored to be invited and
01:15:46
appointed by the president and I
01:15:47
appreciate Saxs and Michael Katzio. You
01:15:50
look back at PCAST, it's kind of rooted
01:15:52
in FDR when he formed this council of
01:15:55
advisers on science when nuclear physics
01:15:58
and quantum mechanics was starting to
01:16:00
kind of reinvent what was possible in
01:16:02
the world. We're sort of at a similar
01:16:04
era today because arguably AI is
01:16:06
reinventing what is possible in the
01:16:07
world. And I think there's this kind of
01:16:09
acute moment that we find ourselves in
01:16:11
in this extraordinary race against
01:16:12
China. I'll give you a statistic. 10
01:16:15
years ago, China published 50% of the
01:16:17
number of scientific research papers in
01:16:19
peer-reviewed journals as the United
01:16:21
States. Last year, they published 50%
01:16:23
more than the United States. This is
01:16:25
across all disciplines and domains,
01:16:27
including physics, material science,
01:16:29
chemistry, biochemistry, broadlife
01:16:31
sciences. There's this moment that we're
01:16:33
in right now where both the world is
01:16:34
being reinvented by AI, but there's this
01:16:37
extraordinary race with China, not just
01:16:39
in fundamental research and discovery,
01:16:41
but in the industrialization of new
01:16:43
discoveries and new technologies. You
01:16:45
could feel it in DC this week. I was at
01:16:47
the Hill and Valley Forum, but literally
01:16:50
everyone in Silicon Valley, everyone in
01:16:52
DC is is like absolutely honed and
01:16:55
focused in on what is going on in China.
01:16:58
It used to be in biotechnology for
01:17:00
example that China was kind of a copycat
01:17:03
and a me too or they were really good
01:17:05
for example in manufacturing but it is
01:17:07
now the case that in many subdomains
01:17:09
China is becoming the scientific leader
01:17:12
in biotechnology and in life sciences
01:17:15
and that is a scary thought because
01:17:17
ultimately China could end up engulfing
01:17:19
the entire pharmaceutical industry and
01:17:21
basically becoming the leader in things
01:17:22
like medicine but most importantly
01:17:25
foundational things like AI. So I just
01:17:28
want to defend and speak for a moment
01:17:29
about the choice about putting what I
01:17:31
would say are like industrial science
01:17:33
and technology leaders on this
01:17:35
commission because this is a moment
01:17:36
where there's an industrial race, not
01:17:38
just a discovery race underway. That's
01:17:40
why this moment is so critical. Anyway,
01:17:42
I'm very appreciative to to have an
01:17:44
opportunity to serve and thankful to
01:17:46
David for his leadership. In related
01:17:49
news, Chim uh will be joining the
01:17:51
president's advisory council on bomb
01:17:54
pots and wagering will be joining that.
01:17:57
And when I saw this sax, I saw PC and I
01:18:00
was like, "Oh, is it podcasting? I'm
01:18:01
in." Finally, I've been invited to the
01:18:03
President's Council on Podcasting. So,
01:18:05
big big news coming. I don't want to tip
01:18:07
anybody's cards.
01:18:08
>> We'll let you know when that happens.
01:18:09
>> Yeah, absolutely. when the when the
01:18:11
president's council on podcasting
01:18:13
emerges,
01:18:14
me and Lex Friedman will make our way to
01:18:17
Washington.
01:18:18
>> I want I want to thank the president as
01:18:20
well and it's a great honor to be named
01:18:22
to co-chair this. It is like Freeberg
01:18:25
was saying, this is a fairly august body
01:18:27
that's goes back a long time. I think
01:18:29
the modern incarnation was created in
01:18:31
1990 by George Herbert Walker Bush. So,
01:18:33
this has been around for over 30 years.
01:18:35
in its current formulation. I think we
01:18:37
have a slightly different take on it,
01:18:39
which is we're going to have these these
01:18:41
builders and doers on there. Like
01:18:44
Freeberg said, we got some catching up
01:18:45
to do on our industrial policy. We have
01:18:48
named 15 people. There's still nine more
01:18:50
slots. We can have up to 24. So, I think
01:18:53
it's possible that at some point
01:18:55
there'll be a second round. And if we're
01:18:56
missing some expertise, they can be
01:18:58
filled in. Obviously, that's up to the
01:18:59
president. He can decide later who else
01:19:01
might join this.
01:19:02
>> All right, everybody. That's another
01:19:03
amazing episode. The number one podcast
01:19:06
in the world, the Allin podcast.
01:19:08
Bye-bye. Love you, boys. Byebye.
01:19:12
>> Let your winners ride.
01:19:15
>> Rainman David.
01:19:20
>> We open sourced it to the fans and
01:19:22
they've just gone crazy with it. Love
01:19:23
you.
01:19:32
>> Besties are gone.
01:19:35
>> That is my dog taking your driveways.
01:19:40
>> Oh man, myasher will meet.
01:19:43
>> We should all just get a room and just
01:19:44
have one big huge orgy cuz they're all
01:19:46
just useless. It's like this like sexual
01:19:48
tension that they just need to release
01:19:49
somehow.
01:19:52
feet. Wet your feet.
01:19:55
>> That's going to be good. We need to get
01:19:57
merch. I'm going
01:20:06
all in.
