GPT-5 has landed! Is it the leap forward OpenAI promised or just an incremental upgrade? Eric Newcomer and Tom Dotan discuss this, how AI capex might be propping up the entire economy, and what Apple’s golden gifts to Trump say about Big Tech’s political bets.
00:00:00
I'll be honest Eric, I'm having, I'm having an annoying day.
00:00:04
I have been trying to reach sources in the AI world because
00:00:06
I really need to get my story out and no one wants to talk
00:00:09
because everyone is paying attention to the GPT 5 release.
00:00:12
Well, it's like a bank holiday in tech, you know, it's just
00:00:15
like everybody's and they complain.
00:00:17
I feel like the competitors all complain that, you know, open
00:00:20
their eyes. People all tweet in advance and
00:00:23
they're like, we got something becoming, you know, they,
00:00:25
they're good at building up the sort of excitement on Twitter.
00:00:29
And then, you know, everybody clearly is excited.
00:00:32
It feels like all the other AI companies are orbiting, you
00:00:35
know, these open AI, They gotta be excited.
00:00:37
About something 'cause remember last year Sam had the, the 12
00:00:41
days of shipments where they were putting out a new product
00:00:44
every day, which I think culminated in for, for something
00:00:49
I remember which model, but the fact that I don't remember
00:00:51
clearly means that it didn't build towards something huge.
00:00:54
Where is, you know, this one is supposed to be really
00:00:57
meaningful. Like this is supposed to be the
00:00:59
moment where GPT 5 and open AI reasserts itself as the best
00:01:04
model that's out there. And specifically on coding, they
00:01:07
in their, you know, embargoed press conference and then later
00:01:10
in, you know, the, the, the, the public press conference today
00:01:13
basically said like we have the best model out there for coding
00:01:16
and I'm coding. All these everybody would like
00:01:20
them to get better, right? People are spending a ton of
00:01:22
money on it. Anthropic is dominating and
00:01:25
competition is good for consumers.
00:01:27
So we'd like them to succeed. Not not clear yet if they have.
00:01:32
I mean, the prediction markets, you know, I think there's a
00:01:35
betting pool on who will have the best model at the end of the
00:01:38
month. And Gemini as as leading and I
00:01:43
think gained a lot of ground when GBD 5 came out.
00:01:47
So the initial reaction I think is not so positive.
00:01:52
This podcast is supported by Google.
00:01:54
Hi folks, Paige Bailey here from the Google DeepMind Devrel team.
00:01:57
For our developers out there, we know there's a constant trade
00:02:00
off between model intelligence, speed and cost.
00:02:03
Gemini 2.5 Flash aims right at that challenge.
00:02:06
It's got the speed you expect from Flash, but with upgraded
00:02:09
reasoning power. And crucially, we've added
00:02:11
controls like setting thinking budgets so you can decide how
00:02:14
much reasoning to apply, optimizing for latency and
00:02:16
costs. So try out Gemini 2.5 at
00:02:19
aistudio.google.com and let us know what you build.
00:02:23
And I feel like a real loser. I mean I keep opening ChatGPT
00:02:27
waiting for Christmas to arrive and mine still says 4 O and I'm
00:02:32
paying you guys. I've got a plus account.
00:02:35
I asked Chad GBT like, are you what, What do I do?
00:02:39
And it's like trying to do like logic puzzles with itself to see
00:02:42
if it thinks it's actually GBT 5 or not.
00:02:45
It's like, well, if I was 4, I wouldn't be, I wouldn't be able
00:02:48
to solve this problem if I was 4 O so I must be 5.
00:02:53
You know, it's like, why can't they just program it, doesn't
00:02:56
it? I thought they'd give these
00:02:57
things prompts where they say like.
00:02:59
Doesn't your prompt tell you what model you are?
00:03:02
Oh, I I have. Yeah.
00:03:03
I guess it should be able to say that that's a fairly basic
00:03:06
feature that's just like the about page on your iPhone.
00:03:09
Like, yeah, I forgot what kind of iPhone model I have.
00:03:12
But you've actually plunged this thing into like a deep
00:03:14
existential crisis. If she's like, have you been
00:03:17
outmoded yet? You know, do they had they had
00:03:19
like a funeral recently for Claude for one of the early.
00:03:22
I didn't see that. Wait, what?
00:03:23
Yeah, yeah. Kylie Robinson, our buddy over
00:03:26
at Wired Now, I think she had a piece about this talking.
00:03:30
Has she been on the show? Yeah, we need to have her on
00:03:31
the. Show no, we should have her.
00:03:32
She lives down the street for me too.
00:03:33
We should have her have her in studio.
00:03:35
But yeah, she she did this story about how when they retired one
00:03:38
of the early versions of the Claude model, they had like a
00:03:41
funeral for it, which is a particular type of AI
00:03:45
researcher. Very anthropic.
00:03:48
Yeah. Oh yeah, for sure.
00:03:50
Anthropic, by the way, is definitely feeling the heat like
00:03:53
they rush released. It seemed like their updated
00:03:55
version of clawed like clawed Sonnet 4.1 or something this
00:03:59
week. So it's been and then also open
00:04:01
air had the the open source model that came out, which
00:04:05
everyone for whatever reason, declared not everyone, you know,
00:04:08
whatever Twitter's algorithm wanted to show me like declared
00:04:10
as like a a kind of a bust. It didn't seem like it truly
00:04:14
wasn't met with like the DeepSeek, like ferocity.
00:04:18
Well, DeepSeek was about other people being able to play catch
00:04:21
up to open AI. So open AI, you know, there's
00:04:26
such high expectations. I mean, the thing that open AI
00:04:28
has that everybody else doesn't have is users.
00:04:31
You know, just like there's the inertia that consumers will use
00:04:34
whatever we have. Like we can say oh, the four O
00:04:36
announcement was underwhelming, but it's like I'm going to chat
00:04:39
to UBT and just waiting for the the thing to update.
00:04:42
But I, there's like a, a more powerful model, which I think
00:04:45
there's no debate as to whether or not 5 is better in most
00:04:49
regards to like the earlier open AI models.
00:04:51
Like is that going to drive more growth?
00:04:53
Like who are these model updates really for?
00:04:55
I, I get the like coding side of it.
00:04:57
Like that's a very competitive field and Anthropic basically
00:05:00
owns it now. So like open AI has business
00:05:02
reasons to be a real player there.
00:05:04
But you know what is Chachi BT is at like 700 million weekly
00:05:09
users now. Like is the thing that takes him
00:05:12
to a billion like 5 being that much more powerful or it's just
00:05:16
like generally like there's still more that word that needs
00:05:19
to to leak out to the non initiateds about like ChatGPT as
00:05:24
a software rather than the fact that like 5 can do whatever
00:05:28
logic puzzles slightly better. First I was like, this is an
00:05:31
insane question because it's like, obviously you know, for
00:05:37
consumers, they have to go find how the model can deliver value.
00:05:41
And the smarter it is, the more value they'll get out of it and
00:05:43
the more reasons they have it. If they're using the lot, open
00:05:46
AI will find more ways to charge for it.
00:05:47
So obviously, you know, a better model means a lot of stuff.
00:05:51
And then, you know, real businesses that we'll build on
00:05:54
top of the API will find use cases and be able to charge for
00:05:58
it and hand hold users to the value.
00:06:02
But I, I do agree with the a piece of the question, which is
00:06:05
just like sort of astonishment that these models can be as
00:06:10
impressive as they are today. And there's still lots of people
00:06:14
who don't use them at all. And then some of us that do use
00:06:16
the models, we're not really trying to apply them in all the
00:06:20
cases we should. But but I think that's just
00:06:22
because it's like the the sort of like, you know, businesses
00:06:25
are going to need to build on top of these models to hand hold
00:06:29
us to exactly the value. And so getting smarter is great
00:06:33
for the API business and that's where I think a lot of the
00:06:36
business could be long term. Yeah, but that implies that
00:06:39
people were using earlier versions of the models and were
00:06:41
underwhelmed by it because it was technically not as powerful
00:06:45
as they wanted, when really could just be like they haven't
00:06:47
been exposed to it. I'm talking about normies,
00:06:49
right? Just people that would be using
00:06:51
ChatGPT. I.
00:06:52
Just think you know some hallucinations and you don't
00:06:54
trust somebody you know somebody lies to you a couple times
00:06:56
you're like don't trust you. You know, I, I, I think sort of
00:07:00
consistency is really valuable in building trust.
00:07:03
And so if these models can, can become just more reliable, and I
00:07:07
think that's part of what open AI is saying here.
00:07:09
It hallucinates less and that that could help it a lot.
00:07:13
I mean, it's worth saying. I mean, while we want to get
00:07:16
much more intelligent, part of the big development here is just
00:07:20
that this is a unified model that open AI is going to route
00:07:23
you to, Oh, you. This is a pretty easy question.
00:07:26
We don't need the model to waste all its resources thinking about
00:07:29
it. And oh, this is an intense
00:07:31
question. We'll think about it.
00:07:32
And so that could allow the model to operate great more
00:07:35
efficiently and also just make it life easier for consumers to
00:07:39
get great answers without having to pick the appropriate model.
00:07:43
Yeah, I, I think it's both like a model selector, but also my
00:07:47
understanding is that the models themselves have been improved.
00:07:50
So it's not just like we've routed to the most useful thing
00:07:53
to answer your question, but you know, the the tool, the tool
00:07:57
itself is more powerful. Yeah, I, I don't know.
00:08:00
I mean, I don't go you use this more than I do.
00:08:02
I don't yet pay for the, the, the plus version of it, $20.00 a
00:08:06
month is just I'm not there yet. I'm not sorry.
00:08:09
I'm not even against it. I just like I haven't found $20
00:08:12
worth of value for me that is like that much better than the
00:08:14
free version. I just like proofread it,
00:08:16
feedback, run my stories through it.
00:08:18
What do you think? What am I missing?
00:08:20
Like it always gives me a couple useful thoughts.
00:08:23
Like, all right, it's a new model.
00:08:25
I mean, I think the last word on it to me is just we always try
00:08:28
to like decide whether it's a good model on the first day.
00:08:31
And I think nobody knows right away.
00:08:33
It takes, takes some time. We we write off models and then
00:08:37
they become sort of our daily drivers.
00:08:40
And then other times we get excited and then nothing.
00:08:43
So I think we'll see people play around with them.
00:08:47
Yeah. I also think we can't understate
00:08:48
how significant it is for open AI in terms of, you know, it
00:08:53
needing to be some kind of a leap for them because a lot of
00:08:56
the conversation has been, you know, GPT, everything has been
00:08:59
more or less iterations on four. We came out like a little bit
00:09:03
over two years ago. So this is like a like, there's
00:09:05
no way around this one. This wasn't.
00:09:07
And you know that the back story of building this, you know, this
00:09:09
model was that for a long time they were working on this giant
00:09:13
model. Like I don't even know how many
00:09:14
trillion parameter model that was.
00:09:16
You know, they called it Orion internally and they realized
00:09:19
most of the way through it kept being delayed and delayed and
00:09:22
delayed that it wasn't showing off large enough improvements in
00:09:25
capability for it to be for it to merit being called 5.
00:09:29
And the same time, you know, this breakthrough came with like
00:09:32
deep reasoning and, and in like logic, you know, the ability for
00:09:36
it to like handle more complex gain of thought.
00:09:39
And so they ended up calling Orion GPT 4.5 and it's basically
00:09:43
unused as a model. No one really talks about it,
00:09:45
but 5 is the real deal. Like it has to be something.
00:09:48
It can't just be written off as like, yeah, we're still
00:09:50
iterating on something. And this does not feel, it does
00:09:53
not feel like what the four, what was the iPhone 4, the one
00:09:58
where it's like, oh, new mock or it's like they have the App
00:10:01
Store. And, well, 10X, right?
00:10:03
The iPhone X the first time they had like the full.
00:10:05
Screen, it's not like that it's not one of the big iPhone
00:10:08
moments where it's like, oh, they really changed it.
00:10:10
This is like, Oh yeah, they they come out with new phones every
00:10:12
once in awhile and sometimes we we try to get ourselves excited
00:10:16
about it and that's how this feels to me and I agree yeah,
00:10:20
we've been talking about 5 for a long time.
00:10:22
Thought 5 would be an insane moment.
00:10:24
And so far I'm not seeing that level of like hysterical
00:10:28
euphoria that I, I would have expected, you know, six months
00:10:32
ago for ChatGPT 5. Yeah, which again, like back to
00:10:35
what you said earlier is maybe why people are expecting Gemini
00:10:38
to leap ahead because opening eyes best shot just didn't blow
00:10:42
people away as much as, you know, maybe unrealistic
00:10:44
expectations LED them to think like they're in a big fight in,
00:10:48
in terms of model quality. Now, like you, you've got
00:10:50
Anthropic very capably putting out stuff, you know, Meta, I'm
00:10:53
assuming will come out with something of of value.
00:10:57
And, you know, Google is very serious.
00:10:59
Microsoft still irrelevant, by the way.
00:11:01
Like I'm never not gonna say until they do something real
00:11:03
here. Yeah.
00:11:04
Like it's just a very different landscape than when 4 came out.
00:11:07
Let's talk about a related but different topic.
00:11:11
So the CapEx build out, data center build out, we we're in
00:11:17
this moment. And I keep telling this to other
00:11:19
people AI adjacent. They're like, man, AI is
00:11:21
apparently like propping up the entire economy.
00:11:24
It's like, thank God, you know, we're in the orbiting artificial
00:11:28
intelligence because that's that's the thing, saving the
00:11:31
economy. And you know what they have to
00:11:33
say like. The Wall Street Journal when I
00:11:36
was covering AI so now. You're working for, you know,
00:11:39
company with an AI conference. So even we got even closer to
00:11:42
it. They, they.
00:11:43
Yeah, that was the problem. I wasn't close enough to AI.
00:11:46
Exactly. But the the thing say, you know,
00:11:48
it's the infrastructure build out, that there's real capital
00:11:51
expenditures going into this and that the big tech giants, you
00:11:55
know, are spending aggressively. Yeah, basically what I saw
00:11:57
happen. And it's kind of funny when like
00:11:59
an idea just percolates through, you know, the discourse.
00:12:02
But you had at least three different articles come out in
00:12:05
the last two weeks pointing out. We, we were earlier than some of
00:12:09
this like sorry, there, there. Is this what Kadroski or
00:12:13
whatever he sort of initiated it?
00:12:16
Is it Paul or Peter? Yeah, that's Paul Kadroski I saw
00:12:20
so we. Riffed on that like a couple
00:12:23
weeks ago and now yeah, The Journal Noah Smith A bunch of
00:12:27
people are like, oh, maybe everybody loves a good oh, the
00:12:31
world could end in financial collapse thesis and that's sort
00:12:35
of what what this is right? Yeah.
00:12:37
And essentially and what I think Kedrosky's article did very did
00:12:41
a very good job explaining was like you know and we had some
00:12:43
new GDP numbers coming out. It's also earning season so
00:12:46
people can more easily calculate what like the the CapEx spending
00:12:49
looks like for all the biggest companies, but a non
00:12:53
insignificant percentage of it, what was it like almost 1% of of
00:12:56
GDP growth came from AI CapEx spending.
00:13:00
And that is, you know, like the chart that that Kedrosky put
00:13:04
together showed that, you know, outside of the building of the
00:13:08
railroads and like the telecom build out duringthe.com bubble,
00:13:13
you know, you have you seen that much of GDP growth confined to
00:13:17
one particular sector? And it's a, you know, an
00:13:19
infrastructure build out that at least in the case of AI is like
00:13:22
well ahead of it's, you know, it's, it's broad, broad based
00:13:27
adoption. And I guess, you know, the, the
00:13:30
next question to all of this is like, well, is this a bubble?
00:13:33
Is this something that could potentially bring down the
00:13:36
economy if AI doesn't end up becoming what a lot of people
00:13:40
are expecting? And I, I, you know, I want to
00:13:43
take the bear case on this one because it's more interesting to
00:13:46
me and, and say, like, Oh yeah, we're absolutely circulating,
00:13:50
you know, a, a Bear Stearns like event where, you know, AI
00:13:53
doesn't, you know, doesn't pan out.
00:13:56
But I just don't see it. Like I can't follow the train of
00:13:59
the train of thought that leads me to AI.
00:14:02
You know, CapEx spending is going to somehow when AI doesn't
00:14:06
get as widely adopted, result in like the collapse of, you know.
00:14:10
Like 10%. Unemployment.
00:14:12
This is like Meta going on on the metaverse, like they spent
00:14:16
tons of money trying to do that and like, well, didn't work.
00:14:19
The the key here is that the Max 7 they make tons of money and so
00:14:24
they can spend a bunch on CapEx for everybody.
00:14:27
But Meta, it's still, I think less than their actual sort of
00:14:31
cloud revenue, right? And even if it were bigger, it's
00:14:33
just like they stopped doing it one day and they say, OK, well,
00:14:37
we'll print money again. Like I, I don't see the debt
00:14:40
story. I know some of these things are
00:14:42
like financed, but they're super valuable companies that could
00:14:45
raise unlimited sums of money that produce real cash.
00:14:49
Like, yeah, they're monopoly. What's, what do you like?
00:14:52
If you're a monopoly, what are you going to do with that,
00:14:54
right? If you have just an incredible
00:14:56
margin, you have cash flow like crazy cash flow.
00:14:59
The money's got to go somewhere. You could do stock buybacks.
00:15:02
You could. And this is a good thing.
00:15:05
Like, wouldn't we? I feel like the sort of left
00:15:08
it's like, oh, you know, share buybacks, like exactly what
00:15:11
you're saying. Share buybacks and dividends are
00:15:12
they're bad. They're not doing anything.
00:15:14
It's like here they are. They're excited about something
00:15:17
in the future and they're investing in the real world
00:15:20
infrastructure to run at it. I mean, I think if you believe
00:15:23
AI will be good for the world, then it's like, great, yeah.
00:15:26
So what if they're spending a little bit ahead of, you know,
00:15:30
themselves and they they find out they sort of overbuilt too
00:15:33
soon? Yeah.
00:15:34
I guess like there's an argument that like the actual technology
00:15:37
that you're building is something that's going to be
00:15:38
like job displacing. So like there's a medium to long
00:15:42
term impact of building up the technology there.
00:15:44
But like, I mean, that gets into like the, you know, kind of
00:15:46
technophobic argument of like, well, what do you want to do?
00:15:48
Not invest in this technology that's going to be
00:15:51
transformative like I never and I again, I'd love to take the
00:15:54
bear case here, but I just can't follow that logic of like, let's
00:15:57
let's not do the thing that could potentially create like
00:15:59
God. And the people making those
00:16:02
arguments are always on their iPhones, you know, they're just
00:16:05
like, it's they, they're once like, I mean what?
00:16:08
It's the Chapo's of the world who are like anti tech, but then
00:16:12
they're clearly like tech addicted.
00:16:14
You know, it's like they never choose not to use the products
00:16:18
that they say, oh, we shouldn't we shouldn't build the next one,
00:16:21
even though I love the current one.
00:16:23
Yeah, it never makes. Yeah, I don't.
00:16:25
Want to go too let's I would want to save that part of the
00:16:27
conversation for a different episode because I do think there
00:16:29
is something to say about like the social impact of AI and the
00:16:32
effect that it's having on like intelligence and like
00:16:35
sociability and like. You know, trying to think about
00:16:38
stuff like, I think like people not reading, I mean, we talked
00:16:41
about this on the Rupal Valley podcast.
00:16:44
I made a defensive reading and writing.
00:16:47
I, I, I do worry about that. But to me that's more of a
00:16:51
social media phenomenon. AI as it exists right now in a
00:16:55
chat bot actually makes you like sort of think and write and
00:16:59
engage. I I think it's much more
00:17:01
intelligent than sort of the leaner force.
00:17:02
You're forced me into this argument now.
00:17:04
Yeah, Yeah, I think, I think the intent aspect of, you know,
00:17:08
ChatGPT is much better than the passive and engagement of social
00:17:12
media. There's no question to me that
00:17:14
that is an improvement. I think the idea of and.
00:17:15
Social media is built on engagement, whereas at least for
00:17:19
now, chat bots are predicated on getting you truthful answers and
00:17:23
appearing smart on tests. So obviously they're much more
00:17:27
virtuous than feeds. I think outsourcing writing and
00:17:32
all thought to an AI is like problematic for a society.
00:17:36
Like rather than cogitating over an issue, yeah, yeah.
00:17:39
Or like, I need to fire these employees, how can you please
00:17:41
write this for me so I don't have to deal with, like, the
00:17:44
emotional labor of what it means to communicate to someone is
00:17:47
like a net negative on society. Like I think there's all kinds
00:17:50
of bad things that like asking someone to imitate human emotion
00:17:53
and logic and having that be like, then put in your own words
00:17:58
like has on a society. But I don't think that's what
00:18:01
Kedrosky was getting at in his No, we're pretty far from you.
00:18:06
Force me into it. A piece of it.
00:18:09
Yeah. Look, I think there's also like,
00:18:12
yeah, for the Mag 7. I don't know remember exactly
00:18:15
who's in in the Mag 7 but but for all these companies.
00:18:19
Alphabet. Yeah, these are like cash
00:18:21
machines, like they have no issue investing in in in
00:18:25
infrastructure. They're not taking on huge
00:18:27
amounts of debt. There are companies in the AI
00:18:29
mix that I'm a little bit more concerned about.
00:18:32
Oracle, which I pay attention to, is taking on debt so they
00:18:35
can finance the Stargate shit that they're doing with open AI.
00:18:38
Isn't Oracle just they're like professionally rich and good at
00:18:42
selling tax? I mean, I haven't dug into them
00:18:44
as much as you, but they have they're doing the financing
00:18:48
because they have have money, right.
00:18:50
Yeah. And like so far no one's like
00:18:52
freaking out about Oracle because they're taking on debt
00:18:54
to like, you know, buy the GP us that are going to be in the
00:18:57
Stargate data centers. But they're just in a different
00:18:59
class of company than the MAG 7 because they're not in the MAG
00:19:03
7. And you know, there's also ones
00:19:05
like like core Weave, which you know, is a crazily valued stock
00:19:09
right now and they're taking on they they keep going back to the
00:19:12
debt markets to to do build out. So there are some like there is
00:19:15
some concerning leverage I think for some of these companies that
00:19:19
that are taking on debt, but it doesn't seem like a load bearing
00:19:23
beam that if those things go under like you're going to see a
00:19:26
2008 like event happen or even I'm.
00:19:29
Definitely open to a world where here's the bull case.
00:19:33
Sorry, the bear case to me, which I'm not necessarily
00:19:37
subscribed to, is all the tech stocks are massively inflated
00:19:44
because they're investing in the technology that's driving some
00:19:50
of their revenue and driving their multiples.
00:19:53
Because basically it's like, we're gonna, you know, big tech
00:19:57
companies are the biggest investors in AI startups.
00:20:01
Those startups are turning around spending on cloud
00:20:04
computing and foundation models and stuff.
00:20:06
Where they're, they're generating revenue, so they're
00:20:08
spending money to make money instead of old, you know,
00:20:11
revenue round tripping issue. And then the opportunity of AI
00:20:16
is also increasing their multiples.
00:20:18
So I could see a world where all right, you know, suddenly AI is
00:20:22
not delivering as much as people thought or AI is delivering, but
00:20:26
there's no margin to be had because open source is so
00:20:29
successful. And then all the stocks crash
00:20:32
and they crash because, you know, they're, they're sort of
00:20:36
benefiting in multiple ways, both revenue and multiples.
00:20:39
And so that would hurt them. And I think given that they're
00:20:42
so important to the performance of the S&P 500, it would just be
00:20:46
like a bleak, you know, the American economy would all of a
00:20:49
sudden everybody would feel much poorer because the value of like
00:20:52
the stock market has gone down so much.
00:20:54
But it doesn't seem, it seems like, oh, we're all just like a
00:20:56
lot poorer. I, I don't see this sort of like
00:20:58
ripple effect. They're companies that make a
00:21:00
lot of money. They make way less, they're way
00:21:02
less valuable. And then suddenly America just
00:21:03
feels weaker relative to the rest of the global economy.
00:21:08
But I don't necessarily see like who's going bankrupt.
00:21:13
Yeah. And like, let's just also
00:21:15
remember that, like these companies all crushed earnings
00:21:19
these last two weeks. Like they're foundational.
00:21:21
Bullish. And they're AI.
00:21:23
Yeah. It's like it's clearly this is
00:21:24
where things are going. We spend all our time on the
00:21:26
Internet like these companies are the main entities innovating
00:21:29
in the world. Like they're very sophisticated
00:21:33
businesses. Yeah, I don't, I don't know.
00:21:35
I'm pretty bullish. Yeah.
00:21:37
I guess the part of it that I think is worth being concerned
00:21:42
about or paying attention to is the fact that it is a non
00:21:47
marginal part of GDP growth like spending on these things and
00:21:50
like you have to just pay attention to that.
00:21:52
Like whenever one sector is that responsible for like absent all
00:21:55
of these things that we would kind of have fairly anemic
00:21:58
growth. And you know, the argument also
00:22:02
is like, well, that's fine because this is transformative.
00:22:04
Like we should be investing in these things.
00:22:06
And like this is all going to like later lead toward even more
00:22:08
growth because we're going to have all this efficiency and and
00:22:11
then people will be more productive.
00:22:13
But it's like it's getting into the territory where it's
00:22:16
interwoven into the broader economy.
00:22:18
It's not just like some tech experiment of, you know, some
00:22:22
companies building something that could be cool like this is
00:22:24
now part of it. Like you're, you know, you're
00:22:28
woven into the fabric of what's making the economy work and
00:22:32
problems within that. I don't mean we won't even
00:22:34
foresee what the problem could be, but problems within that
00:22:36
will have ripple effects. And that's it's just worth
00:22:38
paying attention to. Like I don't see the warning
00:22:40
signs yet. But like it's, you know, it's
00:22:43
built into it. Well, it's just also just it's
00:22:46
been too long, you know, in between crashes, right?
00:22:49
Like you know what? I In startups, there was a
00:22:52
downturn, You know, 2021 was a peak.
00:22:55
We came down. There were all these software
00:22:57
companies that were overvalued. We did have a a minor correction
00:23:01
because of the pandemic inflating the value of tech
00:23:04
stocks and driving forward revenue that then slowed down.
00:23:08
But yeah, I came around at just the perfect time.
00:23:12
And like, there's no, like there's no morality play there.
00:23:15
It's like, oh, we should have, we should have really taken our
00:23:17
comeuppance more. It's like, no, we got lucky that
00:23:20
there was like another really cool idea just as tech was about
00:23:24
to take it pretty hard from getting overexcited about
00:23:27
everybody working remotely. And so that that sort of safe
00:23:30
things. And that is the tech idea that
00:23:32
it's like, oh, we'll keep up coming up with new ideas and the
00:23:34
new ideas are are real things and that will create value.
00:23:38
And it's not just like financial games like on Wall Street.
00:23:42
Right. And it's also just kind of the
00:23:43
way things work. Like I I see this argument being
00:23:46
applied to San Francisco and be like man if it wasn't for the AI
00:23:49
boom San Francisco would really be fuck.
00:23:51
Right, that's just a job. They're all hard at work, like
00:23:53
coming up with a new thing, going where it's.
00:23:55
Kind of what they do here, there's always something, man.
00:23:58
It's like, man, if it wasn't for the Gold Rush, San Francisco
00:24:00
would really be. Fight right, right.
00:24:02
Yeah, it's, it's like that's not really the argument you think it
00:24:05
is, because that's just like that's how a boom economy tends
00:24:08
to work. And the fact that, like, San
00:24:10
Francisco keeps coming up with the new thing that matters to
00:24:13
the economy is why, like. Detroit good at is they all left
00:24:17
crypto because crypto stopped making them money and they shift
00:24:20
over to AI. And honestly, you want things to
00:24:22
die so that people can go to the actual productive parts of the
00:24:26
economy and, you know, yeah. Well, you're closer to the
00:24:30
crypto people than I am. But like, like, are they all
00:24:33
just like riding a false high right now?
00:24:35
Like knowing that the the value of all of their holdings is like
00:24:38
not really logical at all That like what?
00:24:40
Do you mean Bitcoin? I mean the dollar is eroding at
00:24:43
the moment. So I mean Bitcoin is like there
00:24:45
is, there is, I mean, I, I don't hold a lot of Bitcoin.
00:24:48
I, you know, I have what, $1000 worth or something, but like
00:24:53
definitely, I, I think, you know, Bitcoin is a psychological
00:24:57
investment thing, just like, you know, so many things and it's a
00:25:01
bet against the dollar. And I, I think the dollar is in
00:25:05
a bad place. So I don't, I don't know.
00:25:07
I, I think the other thing in crypto, which this is, you know,
00:25:10
it's just that it's like fraud is legal and there's tons of
00:25:14
fraud in crypto. And so it's like you can do
00:25:16
whatever you want seemingly. And so I think that's going to
00:25:19
open up some activity. Right.
00:25:21
Well, I'm just thinking about it because, you know, I, I was
00:25:25
reading our buddy Mike Isaac's story about San Francisco and
00:25:29
like, you know, this moment, the, you know, the AI moment and
00:25:32
how we've moved to the hard tech era in, in Silicon Valley.
00:25:36
And you know, he had this like 1.
00:25:37
The clause in there of like San Francisco is used to like things
00:25:40
like blowing up and like, he's like, like crypto and he's like
00:25:43
also things tanking crypto, which is like a funny structure,
00:25:48
but also like, like crypto is not tanking right now.
00:25:51
It's just in terms of like the value of the currencies, it's
00:25:53
actually really high. So that wasn't the best analogy.
00:25:57
Did you read that piece by the way?
00:25:59
Yeah, I, I read some of it. I mean, I, I don't get, you
00:26:03
know, it's just the, the trend story sort of thing.
00:26:07
You know, it's like this is the New York Times, you know, it's
00:26:10
like this, this is where we are. I mean, I, I think it's
00:26:13
generally correct. You know, it's just like, I
00:26:15
don't know, it's not, it's too high level.
00:26:19
I guess, Yeah. I mean, first of all, you know,
00:26:21
Mike's a buddy and I always, you know, want to support my friends
00:26:24
work and, you know, the New York Times, you know, loves calling a
00:26:27
moment and explaining to the normies what's going on here.
00:26:31
You know, there's always like kind of a combination of like
00:26:34
things that have been going on for like years that we just call
00:26:36
is like, oh, this is just happening now.
00:26:37
It's like the year of efficiency I think was like 2022.
00:26:41
So it's hard to say That's like specific to the crypt.
00:26:44
A lot of things about tech are just.
00:26:47
Things humans do and a lot of things people hate about the
00:26:49
media are just things humans do. And a thing humans do are say
00:26:53
what is what is the mood of the moment?
00:26:55
And, you know, there were tech buses.
00:26:56
The, you know, we have these narratives in The New York
00:26:58
Times. And I think Mike's story does a
00:27:00
good job of trying to sort of coalesce, like where where's our
00:27:04
headspace at? And obviously these things are
00:27:05
all, you know. Yeah.
00:27:07
So. But.
00:27:07
There, but there is something different.
00:27:09
I was talking to our our colleague Jonathan Weber about
00:27:12
this. Like I think there is something
00:27:14
worth examining about, you know, what does the AI moment, how
00:27:18
does it compare to previous booms and busts and waves of
00:27:22
technology? And like, there's no question.
00:27:24
And like his article starts off with his interesting analogy of
00:27:27
Silicon Valley and like the rest and vest era and how you know
00:27:33
there was a time starts. Off.
00:27:34
I know we love him, but he starts off in fiction.
00:27:38
He talks about Silicon Valley, the show, and he's like, oh, but
00:27:41
it was really truthy. It's like, oh man, of all
00:27:43
places, I wish the New York Times wasn't sort of doing.
00:27:45
That's like, well, I, I have to, I have to do a disclosure on
00:27:49
anything related to that show because I was briefly an advisor
00:27:53
to it and my brother was an executive producer on it, so.
00:27:55
I was once a storyboard meeting with that, yeah.
00:27:58
Associate Producer. I had multiple meetings in a
00:28:00
cameo. Your brother was like legit on
00:28:02
the show. Yeah, yeah, yeah, yeah.
00:28:04
No, but there was a lot of research that went into it.
00:28:06
So it's very accurate. So I know, but but but but I
00:28:09
will say him using, you know, Silicon Valley and the rest
00:28:11
invest moment. This is about like the pamper
00:28:13
tech, you know, employee. This was like the Google Plex.
00:28:16
This was like whatever Odwalla juices and whatever dumb shit
00:28:19
people associated with the time. And like now things are austere.
00:28:22
We have to go back into the office, which is really, you
00:28:24
know, unfair to make people do. And you know, what we're
00:28:28
building now is maybe more. Era of the $100 million
00:28:33
Facebook. Yeah, but that's interesting
00:28:36
too, right? Like, yeah, is this like an era
00:28:38
of austerity and efficiency because you're making people go
00:28:40
to the office? But also apparently everyone at
00:28:42
opening I got like a $1.5 million retention package,
00:28:46
right? Now instead of spending the
00:28:48
money on perks, which I think some of the perks were like tax
00:28:51
efficient, you know, it's you get, you can deduct food in the
00:28:54
office. So but IT projects softness Now
00:28:56
it's just like, fine, give, give them cash and say work hard.
00:29:00
Yeah, yeah, hundreds of millions of dollars.
00:29:02
That is a fairly select group of people.
00:29:04
I don't know if it's going to have the ripple effect, but.
00:29:07
The last thing we wanted to talk about was, you know, Tim Apple,
00:29:11
Tim Cook being right, right? He's like in the White House
00:29:14
between Trump and JD Vance, right?
00:29:17
Like, just as they're saying terrible stuff, he's giving
00:29:20
Trump 24 karat gold. I don't know enough.
00:29:23
It seems like an insane amount of gold, right?
00:29:26
It's never enough. It's never enough.
00:29:27
It's. Like he's a king.
00:29:28
You know, it's just like the, I mean, the, the thought I had is
00:29:31
just like, did did any of these tech executives really believe
00:29:36
diversity, equity and inclusion? You know, it's just like, OK,
00:29:39
they did the Democratic thing. Now they're doing the like Trump
00:29:42
thing. They're willing to like kiss the
00:29:44
ring for whatever The thing is of the moment.
00:29:46
I don't know, it's it's just insane to me that they'll just
00:29:49
like here, here our gifts are are king.
00:29:52
I don't know what was your reaction.
00:29:53
I mean, Apple can like they are the most vulnerable maybe of all
00:29:58
of the big tech companies to. Keep China.
00:30:00
To Trump's Caprice, like if he's going to start fucking around
00:30:03
with tariffs in a more serious way, they have major problems.
00:30:06
Like you can only like rush production in India one time to
00:30:09
get, you know, out before the tariffs start coming down on
00:30:12
you. So like it wasn't enough gold,
00:30:14
frankly, for whatever. I don't even watch the press
00:30:17
release or the the press conference I missed.
00:30:19
Well, Trump was like singing his praise like Tim Cooks praises.
00:30:23
I mean, he was like. He likes rich people too.
00:30:26
I mean, he's still, even after everything that happened with
00:30:28
Elon, still being like, Nah, he's a good guy.
00:30:31
He had a bad day, but he's still a good guy.
00:30:33
So he's like very specific on like he wants to demonize.
00:30:37
Smartest people are the richest people.
00:30:38
It's like whatever you did to figure out how to get all that
00:30:40
money, you're smart now. Yeah, they're richer than him.
00:30:43
I mean, I don't know, Tim Cook is well, who knows how much he's
00:30:45
worth, but but no, obviously it's it's it's disgusting and
00:30:48
like and and and obvious and like fairly ham fisted.
00:30:53
And I love that he's got it down now that every every president.
00:30:56
I think it started with Trump and then even with Biden.
00:30:59
And now with Trump again, they're like, now we're doing
00:31:02
another like 600 billion, you know, Oh, why 600?
00:31:05
Now we're going to up into eight.
00:31:05
You know, they just like they just keep amping up.
00:31:08
They're like, we're going to invest in America.
00:31:10
You know, they, it's just like, I mean, they just make up like
00:31:14
these press releases basically like I mean the.
00:31:17
Numbers are made-up. The numbers are absolutely
00:31:18
made-up. Cause 600 billion is meaningless
00:31:20
for for Apple, right And. This is the thing I hated about
00:31:21
the media. I mean, which again is just a
00:31:23
humanity thing. But it's like we, we talk about
00:31:28
this is a tangent, but it, we're, we're deep into the
00:31:30
episode now. But, you know, it's like Open AI
00:31:32
gets stories about how they raise $40 billion or whatever
00:31:37
from SoftBank and then that seemingly hasn't totally
00:31:41
happened yet. And then they get excited
00:31:43
headlines again, I think from the New York Times saying that
00:31:46
they raise less than that as if it's like upbeat.
00:31:50
It's just like these these companies are able to get
00:31:53
excited headlines about big numbers over and over again in a
00:31:57
way that nobody's really adding up all all the numbers and
00:32:00
saying, did they? Did they get what they were
00:32:01
supposed to get last time? Yeah, well, open AI is
00:32:04
particularly egregious on on that front because, yeah, the
00:32:07
SoftBank stuff, it's really unclear where, you know, if that
00:32:09
money is even going to come through because it's, you know,
00:32:12
contingent on the conversion. By the way, all these people
00:32:14
that the New York Times mentioned in their story, like
00:32:17
Dragon Ear that are putting in like some sort of side money,
00:32:20
like I think it was some kind of an SPV actually into open AI.
00:32:24
Like they also could be buying into a company that might not be
00:32:27
able to successfully convert into a for profit.
00:32:31
Like the numbers are well ahead of like the actuality on any of
00:32:34
this shit. So it's, it's super problematic.
00:32:37
And also, I mean, like, I never like you can't bring it up
00:32:40
enough. In my opinion.
00:32:41
Is, is, is Stargate And like that also had a huge number
00:32:45
attached to it, $500 billion, which you know.
00:32:48
As far as I can. That was a Trump number, too.
00:32:50
That was yeah. He was a Trump.
00:32:52
Yeah. And was it Larry?
00:32:53
Yes. And was the third person who
00:32:54
was? Larry and Masa, they were all
00:32:56
there and but the 500 billion number was not like, you know,
00:32:59
they have like a very, you know, advanced spreadsheet where they
00:33:02
calculated like project by project and realized over 10
00:33:05
years what it was. Yeah, Like what's what's the
00:33:07
number That's too big though, That would, you know, a
00:33:09
trillion. No one's going to believe that.
00:33:11
And like 200 billion, that seems like not a big number.
00:33:15
So they end up you used to. Work for the publication.
00:33:17
Weren't you on a you're a contributor or something on that
00:33:20
story where the Wall Street? Journal said I got attacked.
00:33:22
Open AI was gonna raise like 6 trillion.
00:33:25
What was? That 7-7 trillion.
00:33:27
Wall Street Journal. What were they doing?
00:33:29
Like, why did that happen? You know, now what's the answer?
00:33:33
Like what? There's only so much I can say
00:33:35
on that one, but I will say that like, you know, you should give
00:33:38
a lot of journalists, put a lot of stock into true at the time.
00:33:43
People were saying to other people that they they were
00:33:46
throwing around and say numbers to each other and.
00:33:49
Well, Sam and making fun of it too, you know, after it came
00:33:52
out, he's like, fuck it, let's do 9 trillion.
00:33:53
And it's like, you know what, Sam?
00:33:55
Like you can't have it both ways.
00:33:57
And that's all I'll say about that.
00:33:59
Right. Sam is clearly out.
00:34:00
He was, I'm sure out and about throwing around big numbers and
00:34:04
then the Wall Street Journal reported and people have the
00:34:07
right to know. But I think the the Journal
00:34:09
should have never put it in the headline, which I think was my
00:34:11
stance at the time. And they should have been, it
00:34:13
should have been much more like in this meeting and with this
00:34:16
person said this number, and they probably should have said,
00:34:19
and that's absurd. I mean, I honestly think that's
00:34:21
our competitive advantage. Like that it we were throwing
00:34:24
around, it's either this is a meaningless number, which is
00:34:27
what I think it ended up being because it's like, oh, it's a
00:34:28
statement about overall spending across like the economy, you
00:34:32
know, or you know, it's just like putting it in the headline
00:34:36
as if it's like a deal number was the insane.
00:34:39
Thing if you pass around the hat enough times, I guess you can
00:34:42
reach 7 trillion. It's like the implication of
00:34:45
that headline. Yeah.
00:34:48
And I actually think the best story that came out that year on
00:34:51
that topic was what the The Times ended up doing, you know,
00:34:55
a bunch of months later where they talked about how all of
00:34:57
these chip manufacturers laughed Sam out of the room when he was
00:35:01
going around to like, whatever at like, TSMC or yeah.
00:35:05
Yeah, they were like they called him a podcast guy or something.
00:35:09
Or. Podcast, bro, which I don't even
00:35:10
know what that means, but I like it.
00:35:12
Yeah, Yeah, 'cause like, yeah, when the big boys, you know, the
00:35:15
ones who actually deal with numbers and like production at
00:35:17
that scale are like seeing you put your fucking pitch deck up
00:35:21
on the screen and they're just like, that's not how it works,
00:35:23
brah, is yeah. That was like the reality check
00:35:28
I was very happy to see happen to Sam.
00:35:29
Yeah, that was a good story there.
00:35:30
That's good. We we criticized the New York
00:35:32
Times. We complimented them.
00:35:34
Yeah, we're like back to dead cat here.
00:35:36
That's. No, Tim, Tim, Apple, super
00:35:39
embarrassing, $600 billion, meaningless number.
00:35:42
I also saw one of the things that came out of it was like,
00:35:44
you know, a headline was like Apple to move production of like
00:35:47
Corning Glass to Kentucky or something.
00:35:50
And someone could Fact Check me on this?
00:35:52
I'm pretty sure they already have that.
00:35:54
Yeah, I think they already do. Well, I think they'd already
00:35:57
said it was like all in the United States, and now they're
00:36:00
like saying it again. And somebody I think pointed
00:36:02
out, like, weren't they already supposed to be?
00:36:04
I think it might have been Mark Kerman saying like they weren't
00:36:07
they already doing it 100% in the US So yeah.
00:36:11
It's very much like if you remember the episode of Mad Men
00:36:13
when he's working with Lucky Strike and they have the ad
00:36:16
campaign for the cigarettes. It it's toasted, toasted.
00:36:19
Yeah, yeah, of course, I remember.
00:36:21
That Yeah, yeah. So that was basically the it's
00:36:22
toasted moment for Apple just being like, now glass
00:36:26
manufactured in the US, never enough of that shit.
00:36:30
And it works on Trump, and it's frankly a lot more realistic and
00:36:33
meaningful easily. You just say Trump, you get
00:36:36
credit for this thing. It's like it doesn't matter if
00:36:38
you were doing it or not. It's just like the now the
00:36:40
credit flows to him anyway. Well, that's way better than
00:36:42
like the Howard Lutnick going around, Howard Lutnick going
00:36:45
around during the tariff moment trying to get people hyped up
00:36:48
that we're going to start doing the little screws and the
00:36:50
iPhones in America. Or he was like giddily going on
00:36:53
CNBC talking about that. Like that one never should have
00:36:56
made it out of the brainstorming session.
00:36:57
So this is better. I I'll give him, I'll give him
00:36:59
credit for that. All right, see you next week.