Robert Armstrong discusses AI CapEx concentration, NVIDIA's circular finance, and political risks to infrastructure.
Summary
Robert Armstrong argues that the US equity market is structurally dominated by hyperscalers and semiconductors, driven by the largest capital expenditure cycle in US history. He contrasts the current AI boom with 2008, noting that today's investors are backed by highly profitable companies rather than subprime borrowers. Armstrong highlights NVIDIA's role as a bottleneck and its 'circular finance' strategy of reinvesting profits into AI model companies. He also warns of political 'vetoocracy' blocking data center expansion and criticizes AI leaders for failing to establish public accountability. The discussion extends to the social and political fallout of job displacement, drawing parallels to the French Revolution and the role of media incentives in fueling social anger.
Episode sections
- 0:00 – 9:36
Podcast Intro and Guest Background
Demetri Kofinas introduces Robert Armstrong, US financial commentator at the Financial Times and author of the Unhedged newsletter. Armstrong describes his role as writing opinion on finance from America and notes the FT's global readership, split roughly equally between the UK, US, and rest of the world.
- 9:36 – 12:22
AI Investment Boom and Market Concentration
Armstrong highlights that the US stock market is dominated by two groups: the hyperscalers (Magnificent Seven) and semiconductor companies. He notes this is the largest capital expenditure cycle in US history, driven by massive corporate debt issuance from hyperscalers.
- 12:22 – 15:21
CapEx Scale and Contract Rigidity
Armstrong cites an NYU professor's estimate that AI CapEx will reach 3.6% of US GDP over eight years, making it larger than previous cycles like railroads. He argues the cycle is currently locked in due to long-term contracts between data center builders, model developers (like OpenAI, Anthropic), and chipmakers.
- 15:21 – 17:49
CDS Spreads and Market Volatility
Armstrong explains that widening CDS spreads for data center firms reflect increased speculation and a rise in the perceived probability of default, rather than just insurance pricing. He views this as a signal of shifting market sentiment and animal spirits.
- 17:49 – 19:46
Hyperscaler Financial Strength vs. 2008
Armstrong contrasts the current AI boom with the 2008 financial crisis, noting that today's investors are backed by the four to six most profitable businesses in history (Microsoft, Apple, Alphabet, Amazon, Meta). He uses the analogy of a wealthy investment banker buying a house with leverage, rather than a subprime borrower.
- 19:46 – 23:25
NVIDIA as Bottleneck and Circular Finance
Armstrong identifies GPUs, specifically NVIDIA chips, as the current primary bottleneck. He describes a 'circular finance' dynamic where NVIDIA, the 'picks and shovels' seller, is recycling profits by investing directly in AI model companies like OpenAI and Anthropic.
- 23:25 – 26:40
Energy Bottlenecks and Political Friction
While GPUs are the current bottleneck, Armstrong acknowledges energy may become the next constraint. He discusses the 'vetoocracy' in the US, where local and state interests can block infrastructure projects, creating political headwinds for data center expansion.
- 26:40 – 33:16
AI Industry PR and Accountability
Armstrong criticizes AI leaders for their poor public relations, specifically their focus on existential risks rather than safety and accountability. He argues that leaders like Sam Altman and Dario Amodei have failed to take personal responsibility for potential harms, which could lead to severe reputational and legal consequences if accidents occur.
- 33:16 – 35:00
AI Impact on Commercial Driving Jobs
Speaker_1 argues that autonomous trucks are already operating safely and efficiently, suggesting that commercial driving jobs will eventually be taken over by machines. He notes that while local routes remain difficult for current technology, long-haul driving is already being automated.
- 35:00 – 40:45
Political and Social Fallout of Job Displacement
Demetri Kofinas and Speaker_1 discuss the potential political consequences of white-collar job displacement, comparing it to the 'China shock' in manufacturing. Speaker_1 cautions against assuming catastrophic economic outcomes, noting that manufacturing output grew despite job losses, but acknowledges the powerful narrative of displaced workers.
- 40:45 – 43:11
Personal Anxiety and Career Uncertainty
Speaker_1 admits to personal anxiety about his own job as a journalist, acknowledging that AI could eventually perform his role better. He contrasts his current financial security with the precariousness faced by younger workers, highlighting the emotional toll of potential forced career changes.
- 43:11 – 46:06
Complexity of Social Anger and the French Revolution Analogy
Speaker_1 references Simon Schama's book 'Citizens' to argue that the French Revolution was driven by economic dynamism and change rather than just poverty. He suggests that current social anger in the US is complex and not solely economically deterministic, involving factors like social atomization and media incentives.
- 46:06 – 49:31
Media Incentives and Engagement Economy
The speakers discuss how social media platforms and news outlets prioritize negative stories and engagement over factual balance. Demetri Kofinas argues that the 'engagement economy' drives content toward extremes, creating an environment of fear and anger that is difficult to change without regulation.
- 49:31 – 51:45
Transition to Second Hour Topics
Demetri Kofinas transitions the conversation to the second hour, focusing on the French bond market, rising yields, and the potential deflationary role of AI. He also mentions Meta's planned $100 billion stock offering as an example of the shift toward credit financing and equity dilution in the AI build-out.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| MicrosoftMSFT | equity | bullish | So you have Microsoft, Apple, Alphabet, Amazon, Meta. |
| AppleAAPL | equity | bullish | So you have Microsoft, Apple, Alphabet, Amazon, Meta. |
| AlphabetGOOGL | equity | bullish | So you have Microsoft, Apple, Alphabet, Amazon, Meta. |
| AmazonAMZN | equity | bullish | So you have Microsoft, Apple, Alphabet, Amazon, Meta. |
| Meta PlatformsMETA | equity | mixed | So you have Microsoft, Apple, Alphabet, Amazon, Meta. |
| NVIDIANVDA | equity | bullish | So, like such a large percentage of everyone's spending in this area is landing on NVIDIA's balance sheet. because they're the bottleneck. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
The FT readership is approximately one-third UK, one-third US, and one-third rest of the world.
Unidentified speaker (speaker_1) ·4:07
How to verify: Check Financial Times annual reports or audience measurement data from the verification period (2026-08-02 to 2026-11-02) or historical archives. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
AI CapEx is estimated at 3.6% of US GDP over an eight-year period, based on work by an NYU business school professor published on the Brookings website.
Unidentified speaker (speaker_1) ·12:42
How to verify: Search the Brookings Institution website for papers by NYU professors on AI CapEx and GDP impact during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
NVIDIA has a contract to provide chips for four years to a data center in Louisiana.
Unidentified speaker (speaker_1) ·14:46
How to verify: Check NVIDIA earnings call transcripts or SEC filings (Form 8-K or 10-K) for mentions of Louisiana data center contracts during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Primary sourceThe Magnificent Seven and semiconductor companies tower over large-cap indices in the US.
Unidentified speaker (speaker_1) ·10:07
How to verify: Calculate the market cap weight of the Magnificent Seven and semiconductor sector in the S&P 500 using data from the S&P Dow Jones Indices or SEC filings during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Primary sourceHyperscalers are the most profitable businesses in history.
Unidentified speaker (speaker_1) ·17:58
How to verify: Compare net income and operating margins of Microsoft, Apple, Alphabet, Amazon, and Meta against historical corporate profit data from SEC filings during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Primary sourceCDS spreads for data center firms have recently widened.
Demetri Kofinas ·15:27
How to verify: Check CDS spread data for major data center REITs or infrastructure firms from Bloomberg or ICE Benchmark Administration during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Consumer staple stocks like Kimberly Clark, Walmart, and Nestle rallied during a recent market fear event.
Unidentified speaker (speaker_1) ·11:34
How to verify: Check stock price performance for KIM, WMT, and NSR during the verification period (2026-08-02 to 2026-11-02) to identify a period of market fear and subsequent outperformance. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Francis Fukuyama described the US as a 'vetoocracy' where local, state, and national interests have veto power over building infrastructure.
Unidentified speaker (speaker_1) ·25:54
How to verify: Search for Francis Fukuyama's speeches or writings, particularly around the FT Weekend Festival, during the verification period for the term 'vetoocracy' and infrastructure analysis. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Employment for computer programmers has remained stable despite AI advancements.
Unidentified speaker (speaker_1) ·30:37
How to verify: Check US Bureau of Labor Statistics (BLS) data for 'Computer and Information Research Scientists' and 'Software Developers' employment levels during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
China has made the electrification of its economy a national objective.
Unidentified speaker (UNKNOWN) ·23:38
How to verify: Review Chinese government policy documents or state media reports on energy and AI infrastructure during the verification period. For recent events, verify 2026-08-02 through 2026-11-02. Use another period only when the speaker explicitly names it.
Do your own research
- Verify the 3.6% of US GDP AI CapEx estimate by checking the cited NYU professor's academic paper or presentation.
- Review NVIDIA's latest 10-K filing to confirm direct investments in AI model companies like OpenAI and Anthropic.
- Analyze CDS spreads for major data center REITs or developers to assess market-perceived default probabilities.
- Research state and local permitting timelines for data centers in key expansion states (e.g., Virginia, Texas) to assess 'vetoocracy' impacts.
- Compare current hyperscaler profitability metrics (ROIC, free cash flow) against 2008 subprime borrower profiles.
- Check Meta's S-1 filing for details on the $100 billion stock offering and its intended use of proceeds.
- Review Simon Schama's 'Citizens' for the historical argument regarding economic dynamism and the French Revolution.
- Assess the regulatory landscape for autonomous trucking, including federal safety standards and state-level labor laws.
Full transcriptexpand
- 0:00
- Demetri Kofinas: What's up, everybody? My name is Dimitri Grafinas, and you're listening to Hidden Forces, a podcast that inspires investors, entrepreneurs, and everyday citizens to challenge consensus narratives and learn how to think critically about the systems of power shaping our world.
- 0:17
- Unidentified speaker (UNKNOWN): My
- 0:17
- Demetri Kofinas: guest in this episode of Hidden Forces is Robert Armstrong, the U.S. financial commentator at the Financial Times, where he writes the unhedged newsletter and co-hosts. the Unhedged Podcast alongside Katie Martin. Robert and I spend the first hour of our conversation today examining the scale and structure of the AI investment boom, why Robert sees today's
- 0:37
- Demetri Kofinas: hyperscalers as operating from a stronger starting point than their internet era counterparts, and why GPUs and electricity have become the industry's defining bottlenecks. From there, we discuss the political and social background. backlash building around AI, from the local politics that make it harder to build data centers, to the anxiety spreading through the
- 0:57
- Demetri Kofinas: white-collar workforce, and why Robert believes the accountability industry leaders refuse to claim around the misuse of this technology could come back to bite them.
- 1:06
- Unidentified speaker (UNKNOWN): The
- 1:06
- Demetri Kofinas: second hour turns to France, where Robert argues that Is the unrest we are seeing in the streets and in the market for French government bonds is a preview of fiscal reckoning. reckonings awaiting other developed economies, including the United States. We discuss whether inflation is likely to stay structurally higher as governments choose the path of least resistance over genuine austerity, why
- 1:26
- Demetri Kofinas: Europe's chronic valuation discount to U.S. equities may be more structural than cyclical, and how investors should think about protecting themselves in a world where bonds can no longer be relied upon as a counterweight to once. one's stock portfolio.
- 1:40
- Unidentified speaker (UNKNOWN): If
- 1:40
- Demetri Kofinas: you want access to all of this conversation, go to hiddenforces.io/slash subscribe and join our premium feed, which you can listen to on your mobile device using your favorite podcast app, just like you're listening to this episode right now. If you want to join in on the conversation and become a member of the Hidden Forces Genius community, which
- 2:00
- Demetri Kofinas: includes QA calls with guests, discounts. discounted access to third-party research and analysis, and in-person events like our intimate dinners and weekend retreats. You can also do that on our subscriber page. And if you still have questions, feel free to send an email to info at hiddenforces.io and I or someone from our
- 2:20
- Demetri Kofinas: team will get right back to you. Lastly, because this conversation deals with investing, nothing we say on the this podcast can or should be viewed as financial advice. All opinions expressed by me and my guests are solely our own opinions and should not be relied upon as the basis for financial decisions. And with
- 2:39
- Demetri Kofinas: that, please enjoy this wide-ranging, informative, and enjoyable conversation with my guest, Robert Armstrong.
- 2:50
- Demetri Kofinas: Robert Armstrong, welcome to Hidden Forces.
- 2:54
- Unidentified speaker (speaker_1): It's great to be here.
- 2:56
- Unidentified speaker (UNKNOWN): It's
- 2:56
- Demetri Kofinas: great to have you on, Robert. This is actually your first time on the podcast. But we've met before. You were at our recent New York City dinner.
- 3:04
- Unidentified speaker (speaker_1): Excellent dinner. Excellent conversation. Excellent company.
- 3:07
- Demetri Kofinas: Yeah, you were a great addition to the group. And I've been a longtime listener to the Unhedged podcast that you co-host with Katie. You guys have a fantastic rapport and a great banter. And you also write the unhedged newsletter and you are the U.S. financial commentator at the Financial Times. What exactly does that mean, by the way?
- 3:27
- Unidentified speaker (speaker_1): Well, the joke at the FT is that the longer your title is, the less important your job is. And so it's unfortunate that I still have three words in my uh I still have three words in my title, but it just means I'm the guy who writes opinion about finance from America. Of course
- 3:47
- Unidentified speaker (speaker_1): I do write global commentary sometimes, but because you know it's three titles, you know, the ultimate title at the FT is just editor, and that's one. And so I still have three.
- 3:58
- Demetri Kofinas: So the FT is obviously it's a British newspaper, but you guys do global coverage. and you also cover the US. What does the readership look like for the newspaper?
- 4:07
- Unidentified speaker (speaker_1): The readership is about a third, a third, a third. Last time I checked in. So a third UK, a third US, a third rest of the world. And rest, you know, those shift around a little bit at different times, but that's the way I think of it. You know, as a finance writer, I feel like I have the clearance
- 4:26
- Unidentified speaker (speaker_1): closest connections to readers who are in the UK, like people in the city. Those are my most active correspondents. But I do talk to, you know, hear from a lot of people on Wall Street and so forth. So we're wired into both worlds. And, you know, one of the great things about the FT is it's a truly global audience. So,
- 4:46
- Unidentified speaker (speaker_1): you know, I write stuff and I hear stuff from New Zealand and Asia and India and China. and all over the world, Latin America. So this is our strength as a new, our competitive strength, I should say, as a news organization, is that we're really global and have a global reputation and a global readership.
- 5:03
- Demetri Kofinas: What do you feel like is the kind of content or the types of subject matter that resonate most strongly with everyone at the center of the Venn diagram? And then are there certain subjects? that resonate more with the international and UK audiences that are US topically focused?
- 5:22
- Unidentified speaker (speaker_1): Well, I'd kind of highlight two areas. The first, kind of serious grown-up area that people love is just we have a pretty nerdy finance audience. And I am a journalist who you know hits at the center of that audience. So I think of my work as
- 5:43
- Unidentified speaker (speaker_1): being a financial generalist who serves financial specialists. So one of the things that's been surprising to me reporting on finance and writing commentary about finance is most people who work in the industry are incredibly focused on their tiny.
- 6:03
- Unidentified speaker (speaker_1): square of the great spreadsheet of finance. And they actually find a lot of value in someone who's financially literate being able to widen their view a little bit. You know, so the guy who's trading swaps,
- 6:19
- Unidentified speaker (UNKNOWN): you
- 6:19
- Unidentified speaker (speaker_1): know, doesn't even know what the guy who's trading FX at the next desk is doing, right? Because he's just wired in or she's just wired. in on her thing. So I help people take a step back on serious financial stuff. The other thing that works really well for the FT audience, and to an extent for my newsletter and my podcast
- 6:39
- Unidentified speaker (speaker_1): is like any industry, the finance industry loves a little gossip and tittle tattle. So, you know, no matter the topic, people love to hear about personalities, who's rising, who's falling, where the power is, who are the players, who's made mistakes, who's been brilliant. And so, you know, the response to those
- 6:59
- Unidentified speaker (speaker_1): stories that kind of humanize things is always the strongest of any response we get. And so I have a friend, my friend Chris, who's a money manager, likes to joke that the FT is the hometown newspaper. of the rootless global elite. So there's all these like global people who
- 7:19
- Unidentified speaker (speaker_1): are like flying business class all over the world, and they either work in finance or consulting or they're lawyers. And for them, we're their hometown newspaper, where they read the police blotter and the marriages and the all all that stuff. So that's an important role for us. Yeah,
- 7:37
- Demetri Kofinas: that makes sense. That that actually kind of I feel like that resonates also. I find that relatable. Maybe just one more question about media since we both work in the media space. What do you think are most glaring differences between a podcast audience like yours and those who either watch CNBC NBC or
- 7:57
- Demetri Kofinas: Bloomberg on television or video on YouTube or their equivalents? Essentially like folks that are watching clips versus people that are listening to podcasts in finance.
- 8:08
- Unidentified speaker (speaker_1): I've been very pleasantly surprised by the mix of listeners we get. So we get the professional listener who writes in and they say, Look, oh, you were talking about how to hedge currency. And actually, you got the details wrong, and you can do better, you know? But we also get the random
- 8:28
- Unidentified speaker (speaker_1): person who works in a totally different industry who stumbled across the podcast and knows that the stuff Katie and I are talking about is important.
- 8:38
- Unidentified speaker (speaker_1): And they just enjoy listening to the show for 15 minutes. You know, one of the things that's really worked for our show is, you know, we're 15 to 20 minutes, we're twice a week. Katie and I do the work. for our writing, and then we rock into the studio and we just chat about the work that we do. And as you said, we
- 8:58
- Unidentified speaker (speaker_1): have a pleasant rapport because we've been friends for 15 years or whatever it is. And so we get nice responses from like Jane in California, who works at the Garden Center, but you know, knows that this stuff is important and wants to be in touch with it. and feels like we're an accessible way to hear those stories.
- 9:18
- Demetri Kofinas: Wonderful. All right. So let's put your financial generalist credentials to test here. What are the questions that you feel like investors, especially those who are narrowly focused on their domain of expertise, should be asking themselves that maybe they aren't right now?
- 9:36
- Unidentified speaker (speaker_1): Right now,
- 9:37
- Unidentified speaker (UNKNOWN): the
- 9:38
- Unidentified speaker (speaker_1): way people
- 9:39
- Unidentified speaker (speaker_1): think about
- 9:41
- Unidentified speaker (speaker_1): markets is very dominated by the AI trade. And fair enough. So we have two huge blocks of companies in the United States.
- 9:52
- Unidentified speaker (UNKNOWN): In
- 9:52
- Unidentified speaker (speaker_1): the Magnificent Seven, you can call them, or the hyperscalers, you can call them. And then we have the semiconductor companies, which have just become enormous.
- 10:01
- Unidentified speaker (speaker_1): And these two groups of companies kind of tower over
- 10:07
- Unidentified speaker (speaker_1): the large cap indices in the United States for stocks in the SP 500. And because the hyperscalers are issuing corporate debt too. This has become the biggest story in the corporate debt market. Nobody has ever had to raise so much money for such a big project in the history of the United States.
- 10:27
- Unidentified speaker (speaker_1): As everyone is saying, this is bigger than the railroads, it's bigger than electrification, it's just massive. And it's all hands on decks for finance. I think it's easy to be hypnotized by that trade, not only in the sense that you feel like whatever the last piece of news is the final word on this topic,
- 10:47
- Unidentified speaker (speaker_1): or whatever the story is this week, sometimes it's a negative story. sometimes it's a positive story. It's easy to forget
- 10:55
- Unidentified speaker (UNKNOWN): that
- 10:55
- Unidentified speaker (speaker_1): whatever we think about this stuff today, we're going to think something very different in six months. And it's easy to forget that the whole other world of finance is ticking along in the background and still requires thought and love and attention. Interestingly, we saw last week
- 11:13
- Unidentified speaker (UNKNOWN): a kind
- 11:14
- Unidentified speaker (speaker_1): of revolution. rotation trade that suggests people are starting to pay attention. So the biggest dogs in the market for a long time have been your boring consumer staple stocks, both in Europe and in the United States. So it's like Kimberly Clark, even Walmart's been having a hard time. Nestle, you know, the people
- 11:34
- Unidentified speaker (speaker_1): who make our food and our deodorant and, you know, our paper towels, right? And last week when we had that kind of moment of fear in the markets, those stocks
- 11:46
- Unidentified speaker (UNKNOWN): really
- 11:47
- Unidentified speaker (speaker_1): leapt and they'd been creeping up for a while before then. So there is this some sense that people are like, oh, I have to be interested outside of this huge AI euphoria that we're all living in. every day, that I have to look around and be ready for the time
- 12:07
- Unidentified speaker (speaker_1): when everything is not the AI trade.
- 12:10
- Demetri Kofinas: So you said this is the biggest capex cycle in history. How do we know that? What is the data that we're relying on to determine the effect that this trade is having on the broader economy?
- 12:22
- Unidentified speaker (speaker_1): So there's been a couple of papers, attempts by economists to write about. this. The one that is the standard, and I forget the guy's name, but we can look it up, is one that was written by an NYU business school professor, and it's on the Brookings website. And he's the guy who has come
- 12:42
- Unidentified speaker (speaker_1): up with this number 3.6% of GDP over over an eight-year period.
- 12:48
- Unidentified speaker (UNKNOWN): And,
- 12:49
- Unidentified speaker (speaker_1): you know, I've looked over his methodology. You know, I've read the paper, I've spoken to him. It seems like a pretty solid piece of work. Basically, what he's done is gotten a solid estimate of the cost to do a megawatt of AI computing capacity,
- 13:08
- Unidentified speaker (speaker_1): estimated the total number of megawatts that are going to have to be built, do the mathematics. and then compare it to the size of the US economy and the likely size of the US economy over the years to come. And so that's the standard estimate that people are using. And that makes it
- 13:27
- Unidentified speaker (speaker_1): bigger than previous CapEx cycles by like a percentage point. So it's big. And, you know, of course, we have to reckon with the possibility. that
- 13:40
- Unidentified speaker (speaker_1): at some point people change their minds. This is one of the most actually interesting and compelling questions, is
- 13:49
- Unidentified speaker (UNKNOWN): one
- 13:49
- Unidentified speaker (speaker_1): of the things that makes this capex cycle so exciting for investors is that a huge amount of capital flow has been, as it were, wired in, meaning there's contracts between people building the data centers and the people with the AI models like Enthropic and OpenAI
- 14:09
- Unidentified speaker (speaker_1): and the hyperscalers. And so it looks like the trade is all locked in for years to come. Not to mention the semiconductor makers. They have order books going out four years or whatever.
- 14:21
- Unidentified speaker (UNKNOWN): And
- 14:21
- Unidentified speaker (speaker_1): so it looks like nothing can stop this roaring train, right?
- 14:26
- Unidentified speaker (UNKNOWN): But of
- 14:26
- Unidentified speaker (speaker_1): course, contracts get broken. You know, like if this thing really does go sideways, and I don't think I'm a person who has very interesting opinions about whether it will go sideways or not, but things sometimes do. And it's not like all those contracts are going to save you if everybody realizes my God, we don't need this much computer
- 14:46
- Unidentified speaker (speaker_1): capacity after all. This isn't as interesting and productive of use of resources as we thought. In that case, the fact that NVIDIA has a contract with a huge data center in Louisiana to provide chips for four years isn't because that contract, there's going to be bankruptcies, there's going to be broken contracts, and it's all going to change.
- 15:06
- Unidentified speaker (speaker_1): So, you know, the future isn't written. And I think that is a point that, you know, you asked earlier, what is something people need to keep. in mind? And that's it, right? Is the future is not chiseled in stone.
- 15:21
- Demetri Kofinas: So haven't CDS spreads for data center firms recently widened? And and
- 15:26
- Unidentified speaker (UNKNOWN): they
- 15:26
- Unidentified speaker (speaker_1): certainly have.
- 15:27
- Demetri Kofinas: So what what does that tell us? Is that a meaningful indicator?
- 15:30
- Unidentified speaker (speaker_1): Okay. Well, one thing is that as we learned in technicolor during the great financial crisis. CDS spreads are both a kind of insurance. Although CDS people will remind you they're not properly.
- 15:46
- Unidentified speaker (UNKNOWN): This is when you're going to get some of those
- 15:47
- Demetri Kofinas: emails where you get yelled
- 15:48
- Unidentified speaker (UNKNOWN): at by listeners, which is you got the details
- 15:50
- Demetri Kofinas: wrong.
- 15:50
- Unidentified speaker (speaker_1): But think of a credit dis default swap as a financial instrument that pays you if a certain piece of debt, a company or whatever, goes badly wrong. So it functions in a way like. insurance. What we learned in Technicolor in the great financial crisis is that people also speculate in these
- 16:10
- Unidentified speaker (speaker_1): instruments so that they go up and down in value as the probability of default, however small that probability is, might go up and down. So people will jump in there to ride the roller coaster.
- 16:23
- Demetri Kofinas: Which is
- 16:23
- Unidentified speaker (UNKNOWN): one reason why it's not
- 16:25
- Demetri Kofinas: insurance, because you don't have to own the underlying.
- 16:27
- Unidentified speaker (speaker_1): Yeah, exactly. You know, it's fire insurance for someone else's house, right? So those have spiked. And I think what that's showing you is that the marginal chance
- 16:40
- Unidentified speaker (speaker_1): that not everything is going to go as expected has gone up. So I think, you know, there was a fair amount of huffing and puffing last week. You know, this implies a such and such probability of default. I don't think you can read probabilities of default off of these things, but you can sort of see the volatility
- 17:00
- Unidentified speaker (speaker_1): of opinion change, you know, and there is like, oh, we thought there was a 1% chance this whole situation goes sideways. Now it's a two and a half percent chance or something, right? And that's enough.
- 17:15
- Unidentified speaker (UNKNOWN): in
- 17:15
- Unidentified speaker (speaker_1): a highly complex and a highly opaque industry. Moving from 1% to 2.5% is a meaningful event that has to be reflected in prices and in kind of animal spirits broadly.
- 17:27
- Demetri Kofinas: What are the important differences between this boom, this investment boom, and the last time we saw anything like this, which was either a more direct comparison, would be the 1990s. but we could let's also throw into that the credit boom that led to the 2008 financial crisis. What are some of the important differences or similarities, especially between
- 17:47
- Demetri Kofinas: the 90s bubble and today?
- 17:49
- Unidentified speaker (speaker_1): The most important one is that standing behind this great adventure in how much money you can spend
- 17:58
- Unidentified speaker (speaker_1): is the four to six most profitable business. businesses in the history of the world. So you have Microsoft, Apple, Alphabet, Amazon, Meta. These are truly, depending on how you measure
- 18:16
- Unidentified speaker (UNKNOWN): it,
- 18:17
- Unidentified speaker (speaker_1): the most successful businesses the world has ever seen because they have highly scalable businesses with very low incremental costs to. grow and extremely high barriers to entry. These are money printing machines, right?
- 18:34
- Unidentified speaker (speaker_1): And all of these companies, collectively known as the hyperscalers, plus some other companies, have gone all in with their cash flow financing this. So let's say we're talking about a homeowner, right? Levering up to buy a house to use the 2008. We're starting with an extremely rich homeowner. And
- 18:54
- Unidentified speaker (speaker_1): that is different from previous
- 18:57
- Unidentified speaker (UNKNOWN): kind
- 18:57
- Unidentified speaker (speaker_1): of euphoric moments, new technology or new idea in finance, everyone is excited about. That's extremely important.
- 19:03
- Demetri Kofinas: In order to make the analogy work, in other words, the homeowner who was taking out a mortgage in 2006 or 2005 would also have to be very wealthy and have an independently large capital base.
- 19:13
- Unidentified speaker (speaker_1): Exactly. It's an investment banker who's buying the house. It's a rich plastic surgeon in Beverly Hills. So you're starting out with a good income flow and a lot of equity in the house. Now, on top of that, we are piling a lot of debt. And we have less strong companies that are piling in around the edges,
- 19:33
- Unidentified speaker (speaker_1): right? So we've started from financially a much better place than perhaps we started other.
- 19:39
- Unidentified speaker (speaker_1): bubbles or purported bubbles or euphoria. So that is good. I think it's unique
- 19:46
- Unidentified speaker (speaker_1): in this case that the bottleneck is at GPUs. That makes this case different. So, like such a large percentage of everyone's spending in this area is landing on NVIDIA's balance sheet.
- 20:03
- Unidentified speaker (speaker_1): because they're the bottleneck. The thing that everybody needs that there's not enough of is NVIDIA chips. So there's like this one passageway, 80% of the money that's being spent, or I don't know what it is, could be 50%, is just passing through NVIDIA. That's very important. So they are collecting a lot of
- 20:23
- Unidentified speaker (speaker_1): money, and they're not investing all of it,
- 20:26
- Unidentified speaker (UNKNOWN): but
- 20:26
- Unidentified speaker (speaker_1): they are.
- 20:27
- Unidentified speaker (speaker_1): investing some of it back in the industry. They're either providing guarantees to data center leases or investing directly in companies like OpenAI and Anthropic and et cetera. So this is kind of interesting. You know, I used the gold rush metaphor for this phenomenon, where
- 20:47
- Unidentified speaker (speaker_1): NVIDIA is the gold rush company. that is selling the picks and shovels and pans to the miners, right? They're selling the physical infrastructure that people need to do to do AI. And, you know, famously in the gold rush, the company that makes all the money is the one selling the picks and shovels, not the one doing the mining. Well, in this case, the picks and
- 21:07
- Unidentified speaker (speaker_1): shovels company, NVIDIA, India is actually buying mines too. They're taking their profits from the picks and shovels game. and they're investing some of them in mines.
- 21:17
- Unidentified speaker (UNKNOWN): So
- 21:18
- Unidentified speaker (speaker_1): this is what people call the circular finance aspect of this. And I'm not sure this is terribly dangerous, but it's worth noting that the people who are making profits at this stage in the game are recycling the money and making bets on the later stages of the game. I think that's important. So
- 21:34
- Demetri Kofinas: I'd like to dig into your observations about whether the bottleneck is GPUs or whether
- 21:38
- Unidentified speaker (UNKNOWN): it's not the same.
- 21:39
- Demetri Kofinas: actually electrification, the production and distribution of electricity to meet the projected demand from these data centers. But before we do that, does the recycling of profits imply that this cycle could go on for longer than many of the more bearish analysts expect? And that therefore early declarations about a bubble popping may maybe be premature?
- 21:59
- Unidentified speaker (speaker_1): You know, this stuff, as you know, is wickedly hard to predict because there is the fact of cash flows, and then there is the fact of human emotion. Right. And so you can draw all the diagrams you want. And this is a point that I was making earlier, just made in a different way. You can make all the
- 22:19
- Unidentified speaker (speaker_1): diagrams you want of how the cash flows are set up and where they're coming from and where they're committed to and what
- 22:26
- Unidentified speaker (UNKNOWN): they're doing.
- 22:27
- Unidentified speaker (speaker_1): the contracts are, and et cetera, et cetera. But you're always counting on a project that is based on future profits. You're always counting on people's emotional belief in their commitment to the future. So this is all just a long introduction to say your question is hard to
- 22:46
- Unidentified speaker (speaker_1): answer, whether this can keep going.
- 22:49
- Unidentified speaker (UNKNOWN): But
- 22:50
- Unidentified speaker (speaker_1): my hesitancy. hesitant answer is yes. And my tendency to think is that this euphoria period has legs. That is my sense, because there are committed cash flows, there are profitable companies involved, et cetera, et cetera. So I don't, I'm not one of these imminent collapse people. That's the short answer to your question.
- 23:09
- Unidentified speaker (speaker_1): But again, mood is everything, right? So we have to speculate. about mood even as we diagram cash flows.
- 23:19
- Demetri Kofinas: Okay. So let's go back to my question about GPUs versus electrification, because
- 23:25
- Unidentified speaker (UNKNOWN): the
- 23:25
- Demetri Kofinas: debate seems to have started to shift in the last couple of years from worrying about chip production as the primary bottleneck to the AI build-out, to increasingly worrying about energy. China, for example, has
- 23:38
- Unidentified speaker (UNKNOWN): made
- 23:38
- Demetri Kofinas: the broad electrification of its economy and the production, storage, and transmission of cheap electricity a national objective. Aspects of this policy were already in place before the U.S. began enforcing export controls on leading-edge semiconductors, but certainly chip restrictions have only further made the case for relying on cheap
- 23:58
- Demetri Kofinas: and abundant electricity as a competitive advantage. What is your view on this?
- 24:04
- Unidentified speaker (UNKNOWN): The
- 24:04
- Unidentified speaker (speaker_1): first thing I think about that argument, it's you probably know more about it than I do. You know, you've touched on an area I don't know loads about. They are building these data centers in the United States, the big ones, in places where they think they have access to energy.
- 24:18
- Unidentified speaker (speaker_1): Is that going to become when all this stuff is up, is it going to be an energy problem? I certainly read a lot of stuff that says so, but I hesitate because I haven't done the work. So in an extended way, I'm taking a pass, but you're right. This is a theme and it's another area of speculation. And I probably misspoke when I said
- 24:38
- Unidentified speaker (speaker_1): that the GPUs were the only bottleneck. They're the bottleneck now.
- 24:42
- Unidentified speaker (UNKNOWN): The
- 24:42
- Unidentified speaker (speaker_1): next bottleneck might be energy. Yeah.
- 24:44
- Unidentified speaker (UNKNOWN): Right.
- 24:45
- Demetri Kofinas: In that case, I'll direct listeners to something I read a while back. I'm not an expert on this topic either. Packy McCormick wrote a piece called The Electric Slide. I think that was what it was called, if I remember correctly. A while back. Yeah. And he's not the only one that's commented on this, but I think it's a
- 25:02
- Unidentified speaker (UNKNOWN): look,
- 25:02
- Demetri Kofinas: I mean, we have a very old grid and we have a very old electricity distribution, energy distribution infrastructure. Structure, And in some sense, China isn't saddled by that legacy. And also we have a democracy. and local municipalities and
- 25:16
- Unidentified speaker (UNKNOWN): constituents. The next
- 25:17
- Unidentified speaker (speaker_1): important point
- 25:19
- Demetri Kofinas: animated and can create points of friction between hyperscalers and data centers and their ability to actually fulfill the demand that their customers require because it causes higher electricity prices for consumers.
- 25:34
- Unidentified speaker (UNKNOWN): Trevor Burrus
- 25:35
- Unidentified speaker (speaker_1): I was listening at the FT Weekend Festival to Francis. Fukuyama, who is around all the time now because he's doing his book tour or whatever, and a much maligned, misunderstood, but also I think brilliant guy. And he was saying: you know, one of the major political problems our country has is that we are what he calls a
- 25:54
- Unidentified speaker (speaker_1): vetoocracy. Meaning, because of the structure of our democracy, when it comes to building things.
- 26:03
- Unidentified speaker (speaker_1): so many people have a veto over these processes. And his point was: in the world we live in now, being able to build things is an important part of the social contract. And for America, because local, state, national interests
- 26:22
- Unidentified speaker (speaker_1): all have veto power over building infrastructure.
- 26:28
- Unidentified speaker (speaker_1): This is part where our social contract is kind of being broken. He thinks this is actually very dangerous to the stability of our society. It's a very deep issue. And, you know, I think
- 26:40
- Unidentified speaker (speaker_1): I have written a little bit, and my team has written a little bit about the political risk around data centers, that it becomes harder and harder to build these things. And part of this comes down to the astonishingly bad PR job the whole industry has done since the beginning. You know, the fact that
- 27:00
- Unidentified speaker (speaker_1): the big AI labs decided it would be cool to talk a lot about how their product might destroy the world
- 27:10
- Unidentified speaker (UNKNOWN): at
- 27:11
- Unidentified speaker (speaker_1): the beginning. And, you know, they they thought, you know, they're a appealing to the kind of young men who tend to work in the industry. And they're like, you are going to be working on the Manhattan Project, something so powerful and cool that, you know, we're going to have intelligences that, you know, are greater than all human, you know, and it's
- 27:30
- Unidentified speaker (speaker_1): like, shut up, guys.
- 27:33
- Unidentified speaker (speaker_1): No. Like perhaps this is true, but if you want to be an industry,
- 27:38
- Unidentified speaker (UNKNOWN): what
- 27:38
- Unidentified speaker (speaker_1): you need to talk about, we have an incredible technology, it can be harnessed for good, and we're doing everything to make sure that it's safe and we're gonna make it safe. And what we should have heard from Sam Altman and Dario Amade from the word go
- 27:56
- Unidentified speaker (speaker_1): is you have my commitment
- 28:00
- Unidentified speaker (speaker_1): that if these technologies do harm, the buck stops with me.
- 28:04
- Unidentified speaker (UNKNOWN): And
- 28:05
- Unidentified speaker (speaker_1): to this day they can't say that sentence. And to me, if you can't say that sentence and you're trying to start and lead a new industry, you're doing it wrong.
- 28:14
- Unidentified speaker (UNKNOWN): They
- 28:14
- Demetri Kofinas: don't want to say that sentence also because it means that they're liable. And if they're liable, it means that their equity evaluation takes a hit.
- 28:21
- Unidentified speaker (speaker_1): Yeah, but they are liable.
- 28:23
- Unidentified speaker (UNKNOWN): You
- 28:23
- Unidentified speaker (speaker_1): know what I mean? They're on the hook. And if they're living their life. thinking, oh, you know, the accidents that we've had before, they've kind of said the response from the companies has been, Boy, society really has to do something about this. They may think they can get away with that. I don't think they can.
- 28:43
- Unidentified speaker (speaker_1): If there's a real accident,
- 28:45
- Unidentified speaker (speaker_1): they're done. Right? They're gonna going to be new CEOs of those companies. There will be live. liability. Nobody's going to treat this as a social problem when the rubber really hits the road, which it has not done yet. So I think they'd be better off saying,
- 29:01
- Unidentified speaker (speaker_1): we're going to keep you safe and staking their reputations and their wealth on that, because their wealth and their reputation already is at stake, in my view.
- 29:10
- Unidentified speaker (UNKNOWN): Yeah,
- 29:10
- Demetri Kofinas: no, I mean, it does feel like they're caught between a rock and a hard place because. it part of the way in which they are justifying their valuations is by leaning on this idea that they're going to create massive amounts of efficiency in part by displacing workers or reducing new hirings.
- 29:27
- Unidentified speaker (speaker_1): Yes.
- 29:27
- Demetri Kofinas: And that's not obviously a popular message. What was the I don't remember what the growth message was in the 1990s with the internet.
- 29:35
- Unidentified speaker (speaker_1): Yeah.
- 29:35
- Unidentified speaker (UNKNOWN): I
- 29:35
- Unidentified speaker (speaker_1): mean
- 29:36
- Unidentified speaker (UNKNOWN): Do
- 29:36
- Demetri Kofinas: you remember what it was?
- 29:38
- Unidentified speaker (UNKNOWN): As I
- 29:39
- Unidentified speaker (speaker_1): remember it, and I wasn't in finance in you know 99 through 01, but I think the message was everything is going to be so easy. It was much more of a consumer message. It was much less of a business technology. So it was like you're gonna get everything is gonna be delivered to your house. Your newspapers, you're
- 29:58
- Unidentified speaker (speaker_1): gonna going to read it on the computer, you know, everything. And so it was. like this convenience technology on the model of a dishwasher or a washing machine. Whereas this is a was pitched as a business technology. And I don't think, you know, the AI got thought through the fact
- 30:17
- Unidentified speaker (UNKNOWN): that
- 30:17
- Unidentified speaker (speaker_1): when they were saying, look, you're going to be able to run your business much more efficiently, Unfortunately, that also means, you know, a lot of people are going to get fired. By the way, we should emphasize while we're talking about this, nobody's been fired yet, if you believe the headline statistics. There's anecdotal stories, but if you look at employment for
- 30:37
- Unidentified speaker (speaker_1): computer programmers, hang it in just fine. Thank you very much.
- 30:41
- Unidentified speaker (speaker_1): And the history of economic change does tell us that new jobs appear and the transitions, you know, sometimes they're harder, sometimes they're easier. But we've had technological revolutions before, and those did not lead to higher unemployment, you know, even in the medium term. Now, we can have
- 31:01
- Unidentified speaker (speaker_1): a debate about whether AI, being the ultimate general purpose technology, may have a different relationship to employment than all the technologies went before. That may be the case, right? That because it can do everything in a certain way, it's not like electrification or railroads or the
- 31:21
- Unidentified speaker (speaker_1): highway system or the internet. Maybe it will permanently damage the level of employment in a way those technologies different, didn't. But the history of these kinds of transitions is actually reassuring in the medium to long terms.
- 31:35
- Unidentified speaker (UNKNOWN): You
- 31:35
- Unidentified speaker (speaker_1): know, even though, you know, a lot of people who used to drive buggies. maybe they had a hard time when there were no more buggies to drive. Right.
- 31:44
- Unidentified speaker (speaker_1): But, you know, the economies, you know, prove very elastic and resilient over the medium term. Is this time different? That is a question for a very different guest than the one that you have in front of you, I'm afraid. I'm not sophisticated enough on that topic. Are
- 31:59
- Demetri Kofinas: you familiar with Carlo de Perez's technological revolutions and financial capital?
- 32:03
- Unidentified speaker (speaker_1): I am.
- 32:04
- Demetri Kofinas: Okay, so putting aside for a moment whether or not we're in a late stage of a technoeconomic cycle that started 50 years ago, could this framework help explain or or give some weight to the argument that the speed with which jobs will be displaced this time could be greater than what we saw during the
- 32:24
- Demetri Kofinas: eras
- 32:24
- Unidentified speaker (UNKNOWN): of
- 32:24
- Demetri Kofinas: personal computing or the internet? Because so many jobs have been created as a result of those technologies than previously existed. And those jobs appear to be the ones most vulnerable to AI. So software engineering is a great example. So are call centers. I mean,
- 32:41
- Unidentified speaker (UNKNOWN): yes,
- 32:41
- Demetri Kofinas: we had call centers before the internet, but the internet helped scale them into offshore Shore high-volume high volume operations that are extremely vulnerable to automation. Also, aren't we seeing some early indication that this could be the case with postgraduate hirings? Like, is there any
- 32:55
- Unidentified speaker (UNKNOWN): data
- 32:56
- Demetri Kofinas: early?
- 32:56
- Unidentified speaker (speaker_1): Early indications. But we're at a very early stage in the game. So, you know, we can look up that data. I don't think it's there has been a little stress on the hiring of young people. If you dig down into the employment numbers, how long people are unemployed, young people are having a
- 33:16
- Unidentified speaker (speaker_1): harder time, relatively speaking, relative to the larger workforce. But I do not think it is clear that that is an AI effect. And it's not a massive effect. So that's a kind of wait and see.
- 33:30
- Unidentified speaker (UNKNOWN): To
- 33:30
- Unidentified speaker (speaker_1): pick up on the earlier part of your question, the job I think about is commercial drivers.
- 33:37
- Unidentified speaker (UNKNOWN): I
- 33:37
- Unidentified speaker (speaker_1): should have this number in front of me. And again, we can look it up, but the number of people whose job is basically driving. They drive a truck, they drive a van. This is a huge source of employment in America. And it is very clear that there's already 18 wheelers on the road that have no driver in them.
- 33:57
- Unidentified speaker (speaker_1): And they're doing, as far as I can tell, a great job. They're safer, they can drive 24 hours. They never speed. They never cut anyone off. It's better. And, you know, while local routes are still too hard for the technology to do for the long haul, they
- 34:17
- Unidentified speaker (speaker_1): already do a great job and they'll come for the local routes eventually.
- 34:21
- Unidentified speaker (UNKNOWN): So
- 34:21
- Unidentified speaker (speaker_1): that is a massive source of employment that, you know, is reasonably paid as a medium-skilled job.
- 34:29
- Unidentified speaker (UNKNOWN): And
- 34:29
- Unidentified speaker (speaker_1): I just think it's all going away, right? All of it.
- 34:33
- Unidentified speaker (UNKNOWN): You
- 34:33
- Unidentified speaker (speaker_1): know, there might be someone in the van who's like taking the packages out and carrying them to your doorstop or whatever.
- 34:39
- Unidentified speaker (UNKNOWN): But
- 34:40
- Unidentified speaker (speaker_1): all the commercial driving jobs, sooner or later, should be taken by machines because the machines will do them better and more safely.
- 34:48
- Unidentified speaker (UNKNOWN): Where
- 34:48
- Unidentified speaker (speaker_1): are those people going to go? I don't know. Some kind of service job.
- 34:53
- Unidentified speaker (speaker_1): So in my dark moods, that's the one I think about.
- 34:56
- Demetri Kofinas: So I have a couple questions there, but it's okay if you don't have the answers to them. Yeah.
- 35:00
- Unidentified speaker (UNKNOWN): First of all, one of
- 35:01
- Demetri Kofinas: them is an observation, just that
- 35:02
- Unidentified speaker (UNKNOWN): a
- 35:03
- Demetri Kofinas: lot of those folks, anecdotally at least, are actually driving either as a supplementary form of income to their primary job or because they got laid off and they're driving for Uber.
- 35:13
- Unidentified speaker (speaker_1): Yeah, that's true. That is true. So
- 35:16
- Demetri Kofinas: I don't know what that means. I don't know if that means if that means there's more room to cut there because a big chunk of those people are already underemployed or unemployed or lost their job, so they were in a transition phase already.
- 35:28
- Unidentified speaker (speaker_1): Yeah, that could be.
- 35:30
- Demetri Kofinas: I also wonder how those numbers. So do you know what how many million people drive for a living?
- 35:35
- Unidentified speaker (speaker_1): Yeah. I don't know. You're gonna have to ask AA. AI.
- 35:38
- Unidentified speaker (UNKNOWN): You can ask the AI
- 35:39
- Unidentified speaker (speaker_1): right sitting, right
- 35:39
- Unidentified speaker (UNKNOWN): sitting.
- 35:42
- Unidentified speaker (UNKNOWN): I'd love to have a
- 35:43
- Demetri Kofinas: co-pilot AI on
- 35:44
- Unidentified speaker (UNKNOWN): this.
- 35:45
- Demetri Kofinas: I I think that would be a that would actually be a value add. I'm sure that's gonna show up at some point. I mean, they already exist, but
- 35:50
- Unidentified speaker (UNKNOWN): in a way
- 35:50
- Demetri Kofinas: that's really easy for someone like me to use. But like essentially, I wonder because the useful analogy for a lot of these concerns is globalization and the offshoring of entire manufacturing industries to Mexico or China
- 36:02
- Unidentified speaker (UNKNOWN): and other
- 36:02
- Demetri Kofinas: countries.
- 36:03
- Demetri Kofinas: where everything could be done more cheaply and eventually better than here.
- 36:07
- Unidentified speaker (UNKNOWN): And
- 36:07
- Demetri Kofinas: so I wonder like to what degree the job displacement that we're going to see in the white-collar workforce will be comparable or even greater and happen faster. And that causes me to wonder, you know, what will the political fallout be? And which party will be in a position to channel the anxieties and
- 36:27
- Unidentified speaker (UNKNOWN): the
- 36:28
- Demetri Kofinas: rage of the newly displaced knowledge sector workers? And how does that show up in terms of policies like higher taxes
- 36:34
- Unidentified speaker (UNKNOWN): for
- 36:35
- Demetri Kofinas: certain companies or industries, new regulations that maybe attempt to curb the speed and scale with which the disaffected sectors are being impacted? Like how much is this something that you've given any thought to?
- 36:48
- Unidentified speaker (speaker_1): Yeah. So I'm looking.
- 36:51
- Unidentified speaker (speaker_1): at the all employees manufacturing series on Fred.
- 36:59
- Unidentified speaker (speaker_1): And between 2000 and 2010-ish,
- 37:04
- Unidentified speaker (speaker_1): that went from
- 37:06
- Unidentified speaker (speaker_1): 16 million-ish or 17 million to kind of 11 and a half million. So five million jobs went away. As a corollary to that, manufacturing output did not go down in America. Always important to keep in
- 37:26
- Unidentified speaker (speaker_1): mind, right? So that tells you what's China, what's technology, whatever. Maybe manufacturing could have grown more by output in the United States. But first of all, so just to scale it.
- 37:38
- Unidentified speaker (UNKNOWN): say
- 37:39
- Unidentified speaker (speaker_1): we lost four million jobs, and total employment is what 140 million jobs, 150 million jobs, something like that. So I say that to say
- 37:51
- Unidentified speaker (speaker_1): the
- 37:52
- Unidentified speaker (speaker_1): China shock, thought of as a manufacturing shock,
- 37:57
- Unidentified speaker (speaker_1): was significant,
- 38:00
- Unidentified speaker (speaker_1): but not catastrophic.
- 38:04
- Unidentified speaker (speaker_1): But
- 38:05
- Unidentified speaker (UNKNOWN): we
- 38:06
- Unidentified speaker (speaker_1): do have this big narrative about the left behind Rust Belt and
- 38:12
- Unidentified speaker (UNKNOWN): what
- 38:12
- Unidentified speaker (speaker_1): America used to be, and let's build things again. And so I think we have to be very cautious. And I'm not a politics guy, I'm a finance guy and a little bit of economics, right? We have to be very cautious about how powerful this network
- 38:26
- Unidentified speaker (UNKNOWN): is.
- 38:26
- Unidentified speaker (speaker_1): narrative is and separate that from the economic reality. You know, you got to acknowledge that the narrative is very important.
- 38:35
- Unidentified speaker (speaker_1): But the economic reality, so the economic reality, it doesn't look like the China shock was a catastrophic event for the manufacturing base of the United States. A lot of jobs disappeared. Most of it was probably technology, most of it was going to happen anyway.
- 38:50
- Unidentified speaker (UNKNOWN): Okay.
- 38:51
- Unidentified speaker (speaker_1): moving to this one. A lot of jobs are going to disappear. We can talk about driving jobs. We can talk about factory jobs. Driving jobs are probably more important. Service jobs are probably more important. You mentioned data centers. Even if those people find jobs pretty quickly, and by pretty quickly, let's call it, it takes them six months, which
- 39:10
- Unidentified speaker (speaker_1): is very
- 39:12
- Unidentified speaker (speaker_1): optimistic. Changing jobs is harder. hard. I should say, and this is a tangent. I should say that I was speaking once to the guy who ran one of the biggest consultancies in the world.
- 39:21
- Unidentified speaker (UNKNOWN): And
- 39:21
- Unidentified speaker (speaker_1): I kind of asked him, like, what is the hardest thing, you know, in the business world? And he was like, retraining people. We have no idea how to do it. It's really hard. Anyway, so let's say people are to leave the tangent. People are displaced, you know, in a period of transition for six months. I've been through those periods of transition in my
- 39:41
- Unidentified speaker (speaker_1): life. They're extremely uncomfortable. You're very uncertain about the future. You're wondering about your self-worth. Can you support the people you love? Can you support yourself? Do you have to move back into your parents' house? So, you know, even in the economically benign scenario,
- 40:00
- Unidentified speaker (speaker_1): there could be a shorter or longer. period where a lot of people in the electorate are upset. And that could really take over the narrative, it seems to me, even if things play out the way the technological and economic optimists hope that they will.
- 40:20
- Unidentified speaker (speaker_1): Right?
- 40:21
- Unidentified speaker (UNKNOWN): You
- 40:21
- Unidentified speaker (speaker_1): know, there's two things that are really, really upsetting
- 40:26
- Unidentified speaker (UNKNOWN): in
- 40:26
- Unidentified speaker (speaker_1): economics.
- 40:28
- Unidentified speaker (speaker_1): First one is inflation. It makes you feel like the world is crazy. And the one that's even worse than that is getting fired. So I'm not a politics guy. Is this a politically live situation? You bet it is.
- 40:45
- Unidentified speaker (UNKNOWN): Well,
- 40:45
- Demetri Kofinas: I mean, aren't
- 40:45
- Unidentified speaker (UNKNOWN): we
- 40:46
- Demetri Kofinas: already seeing evidence of this playing out in young people who are mobilized around their distrust of big tech and their fears and anxieties about AI making them unemployable?
- 40:55
- Unidentified speaker (speaker_1): Okay, Dimitri, let me just say plainly, I'm worried about this with my job. I'm a journalist, right? I collect a bunch of financial and economic information and then I try to think of something interesting to say about it. Then I write it down and send it to people. That's my job, or I talk about them on a podcast.
- 41:14
- Unidentified speaker (UNKNOWN): Can a
- 41:14
- Unidentified speaker (speaker_1): competition? computer do it better than me now? Maybe. Will it be able to do it as well as me later? Sure.
- 41:20
- Unidentified speaker (UNKNOWN): Do
- 41:21
- Unidentified speaker (speaker_1): I want to work another 10 years? You bet your ass I do. Might I have a forced job change in my life? I absolutely think that might happen to me. And yes, I worry about it.
- 41:31
- Demetri Kofinas: Did we talk
- 41:32
- Unidentified speaker (UNKNOWN): about this at
- 41:33
- Demetri Kofinas: our dinner
- 41:34
- Unidentified speaker (UNKNOWN): in September?
- 41:34
- Demetri Kofinas: Yeah, that makes
- 41:35
- Unidentified speaker (UNKNOWN): I mean I've
- 41:35
- Demetri Kofinas: been talking to everybody about this. I've I
- 41:37
- Unidentified speaker (UNKNOWN): talked
- 41:37
- Demetri Kofinas: about it
- 41:37
- Unidentified speaker (UNKNOWN): on the podcast.
- 41:38
- Demetri Kofinas: I think about this. every day.
- 41:39
- Unidentified speaker (UNKNOWN): I
- 41:40
- Demetri Kofinas: think every single day. And I think you need to be thinking this way. You need to be thinking very, very, very clearly about what is the value that you provide.
- 41:51
- Unidentified speaker (speaker_1): Yeah. And you may be clear about what that value is, and the machine might take it away from you anyway.
- 41:57
- Unidentified speaker (UNKNOWN): And
- 41:57
- Demetri Kofinas: more to the point, what is the monetizable value that we provide?
- 42:02
- Unidentified speaker (speaker_1): Look, I'm lucky. I'm 55 years old. My career has gone well. I've made enough money. I own my house or some portion of my house that the bank doesn't own. I have some pension savings. Right. So like it can't go that bad for me. It can go bad,
- 42:20
- Unidentified speaker (UNKNOWN): but
- 42:20
- Unidentified speaker (speaker_1): like I've made it this far and that's good. Right. But if I was 20 years younger, different story. Now, I should say that
- 42:31
- Unidentified speaker (speaker_1): reading directly from the economic situation to the emotional situation of the country is a very dangerous business. I agree with you that there is no question that there is anxiety afoot in this country, and that young people in particular seem more anxious than older people. Maybe they always are.
- 42:51
- Unidentified speaker (speaker_1): But it is going to take decades for us to figure out how much of that anxiety has to do with things like employment and wages, and how much has to do with other stuff. The unemployment late
- 43:11
- Unidentified speaker (speaker_1): rate is low, right? Real earnings are not. falling, at least, or not falling fast. And they had a period recently where they were rising. You know, I'm reading this book, Citizens, by Simon Shahma,
- 43:25
- Unidentified speaker (UNKNOWN): who
- 43:25
- Unidentified speaker (speaker_1): I'm proud to say is a colleague at the FT, a sometime colleague who writes stuff for us for the weekend section. And this is about the French Revolution.
- 43:33
- Unidentified speaker (speaker_1): And the French Revolution was a similar period in the sense where you had a big change in a society and you had a very simple narrative about it, which was you had the old regime dominated by the nobility. They had everything that was an imp a stale, impervious rich class.
- 43:54
- Unidentified speaker (speaker_1): They lorded it over everyone. And there was this bloodbath getting to the new regime on the other side. A lot of what I'm I'm, you know, in the first. third of Shama's book, but a lot of what the first third of his book is about is actually this picture of an economically sclerotic and class riven France is actually false.
- 44:13
- Unidentified speaker (speaker_1): The class barriers were very fluid. A lot of people were moving up and down the French class system. There was industrialization. There was economic dynamism, especially along the coasts and in Paris. And the real story might have been that dynamism, that change and growth
- 44:33
- Unidentified speaker (speaker_1): was actually
- 44:36
- Unidentified speaker (speaker_1): part of what fomented the revolution, right? That it wasn't about poverty being stuck at the bottom. It was about a change that was coming
- 44:49
- Unidentified speaker (UNKNOWN): and
- 44:50
- Unidentified speaker (speaker_1): was hard for. everyone to understand and made them all upset, right? The simple narrative turned out not to be true. And I'm telling you right now, Dimitri, I don't know what the reasonable, careful analysis of why we live in such an angry country is going to be, but it ain't gonna be simple and it's not gonna be economically deterministic.
- 45:10
- Unidentified speaker (speaker_1): It's not gonna be
- 45:12
- Unidentified speaker (speaker_1): man in the Midwest used to have a good job in factory, lost it, got addicted to oxycodone, and was angry. Right. It's not gonna be that simple. That's a story. There are people like that. It is true. Employment is part of it, but it always turns out to be a much trickier
- 45:32
- Unidentified speaker (speaker_1): picture than that. So we've got to be alert to those other narratives.
- 45:35
- Demetri Kofinas: Totally. I I couldn't agree more. I'm sure we've talked about this on the podcast in the in the past, that it's not just economic. There's a darkness in the country and in the discourse, and there's a roving anxiety and fear, and you can feel it.
- 45:48
- Unidentified speaker (speaker_1): Yes.
- 45:49
- Demetri Kofinas: And to what degree that is a response to a clear understanding of the risks posed, for example, by AI to one's job versus us attaching our fears to these things. because there's something else deeper going on.
- 46:06
- Unidentified speaker (speaker_1): And of course, there's the matter of how we communicate our fears. So there's the social media aspect. There's the social atomization story in America, the bowling alone story, which, separate from the economic story, is probably really important.
- 46:21
- Unidentified speaker (UNKNOWN): I
- 46:21
- Demetri Kofinas: think that's a bigger driver. You know, when you're more isolated, you're more vulnerable. And when you're more vulnerable, you're afraid. And when you're afraid, you can often become angry as a defensive response.
- 46:33
- Unidentified speaker (speaker_1): Of course. Yeah. No, and we all know this from the experience of our own lives. Everybody works, everybody has things they gotta do. And if you forget to hang out with your friends and your family, things can start to spiral down. And there is no, I mean, I think it's been demonstrated
- 46:53
- Unidentified speaker (speaker_1): amply by, you know, tons of historians and sociologists that we are a less sociable, community-oriented country now, whether that's church or work or the sports league or whatever it is that's or the social media, that's part of it. You know, we know in journalism.
- 47:14
- Unidentified speaker (speaker_1): unfortunately, that the negative story sells.
- 47:18
- Unidentified speaker (UNKNOWN): I
- 47:18
- Unidentified speaker (speaker_1): always joke that the story that says like thing goes as planned, doesn't get many clicks. You know, like economy continues to expand at around 2% most years or whatever.
- 47:31
- Unidentified speaker (UNKNOWN): And
- 47:31
- Unidentified speaker (speaker_1): I feel like we have politics now, and in particular a president who has grasped more completely than any president before the political analog of that, which is that negativity really sells, especially to an atomized country. And so, you
- 47:51
- Unidentified speaker (speaker_1): know, I don't know why it is, probably nobody knows why it is, but we are in an angry and anxious country right now. And economics is only part of that, but it is part of it.
- 48:01
- Unidentified speaker (UNKNOWN): Well,
- 48:01
- Demetri Kofinas: you know, you mentioned social media. and that's a really big one. You know, the economics of these platforms, the incentives that drive the type of content that does well, which of course informs what editors and hosts and producers and everyone decide they're going to focus on in terms of the narrative, the the structure of a program, the title, you know,
- 48:21
- Demetri Kofinas: the clickbait on the internet. Everything is driven towards extremes. And
- 48:25
- Unidentified speaker (UNKNOWN): that's the
- 48:26
- Demetri Kofinas: environment that we live in. And I don't really see how I don't see that changing unless when we're speaking specifically about the incentives of the platforms without some kind of government regulation. They're not going to change. If that's what works best for driving clicks and engagement, and we live in an engagement economy. That's the other thing.
- 48:45
- Demetri Kofinas: As well informed as possible. It is, well, rather, let me inform my audience as best as I can. Or let's say in the past, let me also gatekeep this information in accordance with what the government deems to be appropriate information for the public to have. Now a big driver, bigger than ever, is engagement.
- 49:05
- Unidentified speaker (speaker_1): It's tough. And you know, I always sort of shudder. I'm Um on the business regulation side, I'm a pretty, pretty much of a libertarian, not. nut job. I'm like a let markets get on with it guy.
- 49:18
- Unidentified speaker (UNKNOWN): But
- 49:18
- Unidentified speaker (speaker_1): I hear you, this is a hard problem to solve. And it's a hard problem to solve for my industry and your industry in particular. And maybe there's a regulatory solution. I, you know, I I kind of hate to think about regulation, but we may get to that point.
- 49:31
- Demetri Kofinas: So since you brought up the French Revolution, I think France is where I'd like to pick up in the second hour, as well as the French bond market, where we've seen oat yields rise sharply in the last few months. And this will also afford us a good opportunity to discuss the treasury market and whether owning bonds is starting to make sense again, especially if AI
- 49:51
- Demetri Kofinas: turns out to play a deflationary role in the economy over the long term.
- 49:55
- Unidentified speaker (UNKNOWN): And
- 49:55
- Demetri Kofinas: I also want to revisit our discussion about where the financing is coming from for the AI build build-out out and how it compares.
- 50:03
- Unidentified speaker (UNKNOWN): to
- 50:03
- Demetri Kofinas: the build-out of the internet, because we're beginning to see a move toward more credit financing as well as new stock issuance. Meta, for example, recently announced that it plans to raise close to $100 billion in new stock offering. And both of these things, more reliance on credit financing and equity dilution, may suggest
- 50:23
- Demetri Kofinas: that we are closer to the end than we are to the beginning of the cycle.
- 50:27
- Unidentified speaker (UNKNOWN): But
- 50:27
- Demetri Kofinas: we're going to address all of those questions in the second hour, Robert.
- 50:31
- Unidentified speaker (UNKNOWN): For
- 50:31
- Demetri Kofinas: anyone new to the program, Hidden Forces is listener supported. If you want to access the second hour of today's conversation with Robert, head over to hiddenforces.io slash subscribe and sign up to one of our three content tiers. All subscribers get access to our premium feed so you can listen to the rest of today's conversation on your mobile device using your favorite podcast.
- 50:51
- Demetri Kofinas: app, just like you're listening to this episode right now. Robert, stick around. We're going to move the second hour of our conversation onto the premium feed.
- 51:02
- Unidentified speaker (UNKNOWN): If
- 51:02
- Unidentified speaker (speaker_2): you want to listen in on the rest of today's conversation, head over to hidden forces.io slash subscribe and join our premium feed. If you want to join in on the conversation and become a member of the Hidden Forces D. Genius community, you can also do that through our subscriber page. Today's episode was produced by me and edited by Stilianos
- 51:22
- Unidentified speaker (speaker_2): Nicolau. For more episodes, you can check out our website at hiddenforces.io. You can follow me on Twitter at Kofenas, and you can email me at info at hiddenforces.io. As always, thanks for listening. We'll see you in
- 51:38
- Unidentified speaker (UNKNOWN): the next video. time.
More from Hidden Forces
Why Nothing Looks Cool Anymore | John Mauriello
Episode discussion: John Mauriello argues modern design prioritizes corporate efficiency over human desire, enabling scams and uninspiring products.
Speakers’ claims: Demetri Kofinas interviews John Mauriello, founder of Design Theory, about the decline of humanist design and the rise of corporate minimalism. Mauriello claims that contemporary industrial design, exemplified by Apple, has abandoned Bauhaus principles for profit-driven standardization and algorithmic taste-making. The discussion links this design shift to a broader cultural cynicism that facilitates scams like Theranos, where storytelling replaces physical reality. Mauriello contrasts Steve Jobs' respect for physical constraints with Elizabeth Holmes' dismissal of scientific facts. The episode concludes with Mauriello expressing a desire to transition from media commentary to physical product creation.
The Return of Economic Populism and the End of Two Percent | Vincent Deluard
Episode discussion: Vincent Deluard on Populism, Inflation Target Abandonment, and Foreign Capital Reversals
Speakers’ claims: Vincent Deluard argues that the US midterm elections signal a shift toward economic populism and the abandonment of the 2% inflation target due to structural deficits. He contends that inflation measurement is distorted by service-based subscription models and discriminatory pricing. Deluard details the reversal of the Mexican Peso thesis due to border enforcement and rate hikes, while highlighting Canada's political and economic turnaround. Speaker 1 warns that foreign capital repatriation, particularly from Canadian pensions, poses a long-term risk to US asset valuations. Speaker 2 describes France's debt crisis and political fragmentation, noting that Marine Le Pen's left-leaning populism threatens French bonds and Eurozone stability.