BIG INTV: Joe Weisenthal Has Predictions About How the AI Bubble Will Burst

0:00This show is supported by OutShift, Cisco's incubation engine. Today's AI agents operate in silos, limiting their true potential. We've been focused on building bigger, smarter models, but scaling up is just one approach. To reach superintelligence together, we need to do more. We need to scale out. And we actually have a blueprint from 70 ,000 years ago. Humans didn't just get smarter individually. The cognitive revolution transformed society because we began sharing knowledge, goals, and innovation. Agents are now at that same inflection point. They can connect, but they can't think together.

0:38That's why Outshift by Cisco is building the Internet of Cognition, transforming AI from isolated systems into orchestrated superintelligence. By creating an open, interoperable infrastructure, Outshift by Cisco is enabling agents and humans to share intent, context, and reasoning. The cognitive evolution for agents is here. Explore the Internet of Cognition at outshift.com. That's outshift.com. The world gets more interesting when we stop and ask questions. That's what Shortwave, NPR's science podcast, is all about. Shortwave spends each episode answering a big question like, Like, why do we have nightmares?

1:17How does AI affect my energy bill? And how can trash be good for the environment? In under 15 minutes, shortwave follows a question wherever it leads and brings you along for the ride. It's science, but not the kind you remember from school. It's surprising, playful, and full of wonder. Science isn't just for PhDs of the world. It's for everyone, including those of us who didn't realize we love science. At Wired, we love a good rabbit hole. We recommend listening to their recent episode, Where Did the Moon Come From? Earth didn't always have a moon. In the beginning of the solar system, when the planets were still forming, something happened that would change Earth's night sky forever.

1:59Co-host Regina G. Barber takes you on that journey. Follow NPR's shortwave podcasts and exercise your right to wonder.

2:12From WIRED, this is The Big Interview. I'm Katie Drummond. All anyone can talk about these days is how AI is revolutionizing everything from coding to copywriting. But AI is literally affecting the majority of the population in a more tangible way, the global supply chain. Companies are throwing machine learning at forecasting, logistics, operational efficiency, and tariff Tetris. Meanwhile, economic experts are increasingly sounding the alarm that this AI mania, and the billions of dollars being thrown into the tech, is creating a bubble that might very well burst, and take the economy down with it.

2:46To find out more and figure out whether or not to hoard gold bars and prepare for an AI economic apocalypse, I called a former colleague and very talkative, very smart economics expert. Joe Weisenthal is co-host of Bloomberg's Odd Lots podcast. Here's our conversation. Joe Weisenthal, welcome to The Big Interview. Thank you. Thrilled to be here. Nice to see you again. It's nice to see you. It's been too long. We were just talking that Joe and I worked together. What was that? Like nine years ago? It was 2014, 2015. So like maybe 10 years ago or something. 10 years ago. I worked at Bloomberg. I lasted about a year.

3:24It's not bad. Not bad. It was in my mid-20s, maybe not the most conducive environment to my professional success. Let's just say that. But Joe, you were there. You were loud. You were proud. You were excited. You were always very excited about the economy. I love talking about the economy. When you would just scream and shout and you would hoot and holler. And I was like, who is this guy sitting across from me and why is he so loud? But I really appreciated your enthusiasm. Thank you. So we always start these conversations with some very fast questions. Are you ready? Oh, yeah. Ready? Yeah. Okay.

4:01So you... I'm stressed out. You wake up extremely early. Yeah. What is the latest you've slept in this year? Oh, this year on a weekday? Well, probably no later than 7. But that's because of kids. Yeah. You know, I'm getting a little bit soft in my old age. If I didn't have anything, I could see sleeping until 8 or 9. I've really changed. Wow. Those days when I was waking up at 4 a.m. every single day. And you were tweeting. And tweeting. Yeah. I still get up early a lot, but I let myself slide a little bit now. What's the weirdest market indicator you actually take seriously? We did a really good episode of our podcast about a month or two ago about cardboard boxes, which are very important.

4:42I don't know how weird they are. I guess it's kind of intuitive because of e-commerce, but you can really dive deeply into cardboard box volume sales, the types of cardboard boxes that are hot right now. Like are they ones that go for individual packages? Are they more of the larger grocery-oriented ones? Yeah, cardboard boxes are top of my mind these days. I used to pay a lot of attention to Macau gambling receipts, et cetera, when that was more of a measure of Chinese capital market outflows. But I think they've clamped down on that, and that's not as useful as an indicator as much. So boxes more so than gambling receipts.

5:17These days, yeah. These days. If you could ban one financial cliche from headlines forever, what would it be? Oh, there's so many. It's hard to choose from. The stock market is not the economy. People love saying it. It's not true. The stock market is a very important part of the economy. Investors hate uncertainty. Things are always uncertain. That's never been a thing. So I could go on and on. But I would say the stock market is not the economy is one that I've hated for a long time. And I would say a very timely one, which we will get into. The stock market is very important right now. What is a chart you would get tattooed?

5:49Oh, I really like the chart of public sector employment over the last hundred years. No, you know why it's a fun chart? I'm imagining the tattoo. I don't think I would actually get it tattooed. It's just one of my favorite charts, though, because it's one of the – if you see this chart and you never knew what it was, you could probably figure out what it was because there is a little spike every 10 years, and that's additional hiring for the census collection. And so I've always just enjoyed this chart because if you really think about, okay, what is this indicator that always spikes right at the turn of every decade?

6:21There's one thing that happens in the economy every 10 years, 1990, 2000, 2010. And that's the census. You might be able to figure it out. Maybe it's like a fun – it's fun trivia. If the economy were a movie, what's the title? Oh, man. If the economy were a movie, what's the title? I don't know. Mind Breaker. Something like that. Because I do think that maybe more the market than the economy per se, but I do think the story of many people of the last several years is how many people's brains have been broken because things that are perceived to make sense or relationships that they had assumed to be stable, et cetera, have not held.

6:59And you just see like people who are in the space going crazier and crazier year after year because they have such a difficult time reconciling what they see on their screen. Wow. You are seeing these people go crazy? I think so. Yeah, I think there's a lot. I think so. I mean, I don't know if it's like sort of outward signs of, you know, literal insanity. But I think in markets, in many realms, you see this sort of intense frustration. The biggest thing that I think of, and again, I know we'll get into this, is for almost my entire career, but at least the last 10 years, people have said like, oh, like tech stocks can't go any higher.

7:35Or when are we going to rotate out of tech stocks? Things like this. Everyone is all in on this already. and yet these trends just persist month after month. And most people pride themselves in some way of not being part of the herd, right? Or taking some sort of contrarian viewpoint or finding something that everyone else doesn't know about, right? Everyone at this point has heard of NVIDIA, everyone in the world, and yet you could have just made a lot of money over the last couple of years just buying NVIDIA. And I think this phenomenon really breaks a lot of people's brains because the industry to some extent is supposed to encourage people to seek out information that others don't have.

8:09And I don't think that's been a particularly fruitful pursuit for a while now. And I think it's mind-breaking. All right. Last question. One economist, living or dead, you get to have dinner with them. Who is it? Living or dead. I'd love to talk to John Maynard Keynes, honestly. I mean, he was a very interesting character. Hung out with all sort of artists and dancers. And I love his writing. And yeah, if it's dinner, probably him. Dinner. Maybe one day. He's dead. But maybe one day the AI-generated, reanimated corpse of John Maynard Keynes. Don't put it past them. So let's talk big picture for a second.

8:46You work at Bloomberg. We've established that. I know firsthand, as I was saying, you really care about this stuff. You really care about what you cover. But you've said that the thing that really interests you is, quote, figuring out what's actually happening underneath the headlines. So my question for you, Wired publishes a lot of headlines. So does Bloomberg. What is going on underneath the headlines right now? Underneath the headlines. How would I put what's going on underneath the headlines? I think there's a few things. Look, there's obviously the headlines, I guess, are all about the AI build-out and how the AI build-out is represented in the market.

9:20So I guess that is what the headlines. I think probably what's less discussed is maybe some of the crowding out effect of this. Something that I think a lot about, we had this guy on our podcast. He's a real estate developer for drive-thru coffee shops. So like Starbucks when they want looking for a new drive-thru location or Dutch Brothers, et cetera. And we've talked to them. And in recent years, really basic electrical equipment that would be required to build a new – get a new facility opened up has been in short supply because there's all this capital, physical capital that's been going into data centers.

9:53We all know about that. Right. That you hear about all the time. But when you think – just think about that. Well, okay, where were these things going to go instead, right? if any supplier of any gear, whether it's cooling gear or heating gear or water filtration or doors or gas turbines, electrical equipment that allows any facility to hook into the power grid, all of this is an incredibly short supply right now. So the ISM manufacturing report, that's a monthly economic indicator I look at. One of the things that they always ask their survey respondents is what is in short supply right now. And with the exception of maybe two months in the last five years, electrical equipment has been in short supply for five straight years.

10:34And so if you think about what are the costs of the AI buildout, and I'm not trying to dismiss it per se as being wasteful or anything. I have no opinion on that. But I do think that it's interesting to think of some of these second order effects of who is not able to get natural gas turbine equipment, who is able to not get electrical gear equipment because some big data center operator is just willing to pay top dollar for it in order to get these things online. I think that's one of the most interesting stories there is right now. In a market economy, the way allocation of capital goods works, of course, is price signals and the builder of the data center is going to pay more for that same piece of gear than someone who is building a drive-thru coffee shop.

11:18But, you know, you could imagine in a more command economy, if you're thinking about how it would be done in China, they would just say, okay, right now we're prioritizing data centers for drive-thru coffee shops. And so they're going to, that is where we want to direct our resources. We do it through the price system, which works well. But if some of these investments don't pan out, et cetera, it's interesting to think what type of economic activity was constrained because it didn't have access to capital goods during this sort of incredible boom. I did economics in high school, just saying. So I know a little bit about supply and demand.

11:50But I'm curious about, Now, you just sort of mentioned electrical equipment has been in short supply for five years. Other types of goods relevant to these build outs also in short supply. Why has supply not increased to meet the obvious increase in demand that exists, right, between these, you know, people trying to make Starbucks drive-thrus, but also this data center boom? Like, why isn't the supply rising to meet that elevated demand? It's a really good question. I think there's a couple of things. I mean, I think, like, at the margins there has been more supply. So there's going to be more.

12:24But in times of uncertainty, building a big factory or building a new facility is very risky, right? You don't really know how long any boom is going to last for a period, right? So why not, if you have, if there are, say, three companies that are the makers of some piece, some widget, right? That's very crucial in the economy. And it's costly. why not keep supply where it is and just raise your prices? And jack your prices up, right. And then you have an order book that's consistent out to your 2030, in some cases, at top dollar. And, you know, it's very difficult to get out of that equilibrium because that's very nice if you're one of the sellers in this environment.

13:07And the alternative is to build another factory, make a big bet on upfront capital outlay. So that's money going out the door. and that's uncertain, and you're expanding supply, so the price is going to be softer than it otherwise would have. See, and it's really not just things related to electrical gear. I think one of the themes across the economy for a long time is the sort of paying the price for persistent underinvestment. I think another good example of that is housing. So, housing activity was pretty soft in the years after the great financial crisis. All these home builders went out of business because housing prices collapsed and so forth.

13:44And so there was just not that much, you know, new housing activity. When you look at the surge in home prices now, especially post-pandemic, part of what we're experiencing now is the price that we paid for underinvestment in the 2010s. And I think you can really go across a range of industries and find this phenomenon that when the economy lies fellow, when the economy is soft, investment goes down. And then when you get this boom in some either across the economy like we got in 2021 or 2022 or in specific sectors like we've seen in 2024 and 2025 with data centers, you run into a supply crunch pretty quickly.

14:21And that brings me to tariffs, which have been, I think, for me and probably for many people, one of the most insane, confusing, befuddling phenomena of this year of the Trump administration. Definitely. Can you explain to us, like, what is going on with the global supply chain with regards to tariffs right now? Yeah, the tariffs are not high enough to induce a massive amount of domestic investment in some area, right? So, you know, you like put tariffs on furniture. Like the tariffs aren't high enough that it makes a lot of sense to build big furniture factories in the United States. But the tariffs may be high enough such that someone who had been importing them from China or Vietnam, maybe they're looking for a supplier in India.

15:06And that's definitely happening. So the furniture still arrives and by and large, there aren't massive shortages of things in the U.S. economy right now the way people might have thought in early April. Part of it is that those early tariff levels, they all came down. Maybe the story would have been different. Maybe we would have had more persistent actual shortages had those initial tariff rates stayed in place, but they didn't. They were like – I think the administration thought they were unsustainable. So then there's this switching where, OK, we're going to maybe switch some vendor sources.

15:36we're going to move to a lower cost country. So anyone who's in the business of importing, maybe they can sort of keep prices roughly stable, but there is this costly process of, okay, now we have to spend money finding a new supplier, building a new relationship with them, changing our trade routes, figuring out all these things, figuring out how they operate. But then also this new supplier, they don't know the customer as well. They may be reluctant to make long-term commitments with the customer because they don't know what the tariff rate is going to be come April, and maybe I don't want to take delivery of the items.

16:08This is a worry for suppliers that they're going to make this, you know, whatever, whatever the good is, furniture, gardening goods, holiday goods, and then the tariff rate is going to go up, and then the customer walks away from them. So the way I think about tariffs overall, outside of some niche areas, is not that they've raised the cost of specific goods dramatically, though in some cases they definitely have. It's more that they've really raised the cost of doing business in the United States. And I think that is sort of the story. If I want to understand what is going on with the economy right now, I'd say there's basically two things, which is there's this slice of the economy that's doing phenomenally well, which has anything to do with big tech and AI.

16:47And then there's this economy that's sort of creaking. And I don't know, like the rest of the economy, is it in recession? It's fairly stagflationary, I would say. Hiring is certainly pretty mediocre right now. We're recording this on a day that we would have gotten a jobs report had the government been open, but we don't. But we know, we're pretty sure that hiring is certainly not booming. The cost of living is very high, and so it's still a lot of price pain. So I think the broad swath of the economy outside of AI is kind of creaking, and tariffs are probably part of the story. I wouldn't say the entire story, but they've raised the cost of doing business for the entire economy, even if not directly in goods, just in the effort that companies have to do to go into to reorient their supply chains.

17:39This week on the political scene from The New Yorker, Trump's rupture in the world order. Europe caught between two adversarial great powers. That's basically dialing back the clock to not only pre-World War II, but really it's a pre-20th century view of the world. And I would say it's a world of permanent insecurity that we're looking at. Join me, Evan Osnos, and my colleagues Jane Mayer and Susan Glasser every Friday on The Political Scene, available wherever you get your podcasts.

18:22so this brings me to i mean i have a lot of questions for you about ai especially in terms of the sort of increased murmuring or not even murmuring people saying out loud we're in a bubble yeah right we're in an ai bubble um it will burst at some point you know there are sort of varying degrees of hysteria around this notion. How is artificial intelligence affecting the global economy right now? Are there some examples that you can give us to, like, bring this down to earth a little bit? I would start by saying I'm very skeptical that AI as a sort of tool is having a significant impact on the global economy, including hiring.

19:01Now, I would caveat that a little bit. I do think it's possible that across a range of organizations right now that there is a lot of pressure on managers to show that they're using AI, whatever that means, that they're implementing, that they're finding ways to capture value from AI in their workflow. Everyone's doing experiments. Everyone's spending money. Everyone's trying to figure something out. One way that you could sort of demonstrate that you're using AI and getting value is to just cut headcount. And then you go back and say, oh, because we adopted some AI workflow, we were able to reduce hiring by 10%.

19:38That doesn't mean that actually the AI tool allowed you to do the same amount of work with only 90 percent of the labor. I mean, when I see those announcements, my immediate assumption is air cover. You did layoffs and you used AI as air cover to explain why you're eliminating 10 ,000 positions. I just don't see the technology being equipped to eliminate 10 ,000 white-collar jobs right now. I mean, I've asked so many different people and organizations. So have I. Yeah. I can't figure it out. And anytime I ask about actual implementation, I say excluding, don't give me the stuff about how your coders are using it because I know that's well-established.

20:14Yeah. The other possibility that connection is, okay, there is this mandate or there's impulse to spend more on AI. And so you're thinking about your budgetary allocations for the year 2026. We did an episode today and our guest was talking about this, that perhaps, okay, we know we're going to spend more on AI tools, services in 2026. So that means we're going to allocate less to hiring in 2026. I don't think there's a lot of evidence for it. And then, like I said in the beginning, you know, I think the stock market really is an important part of the economy. It's not just a scoreboard. It's an engine.

20:48It is a driver. People found companies in order to sell them to public companies or to sell into the IPO market. And when stocks are up, that means IPOs are going to be more valued. That means C rounds will be more valued. That means more people are going to move into this space. And that's fine. That's something, you know, Cain's talked about that, speaking of my ideal dinner guest. And so this has been understood for a very long time, that financial markets have this accelerant, that they're a force of production, so to speak, within the economy in a very real way. And I do think that the big rise that we've seen in asset prices, in stocks in particular, it's been related to technology and AI, is a big part of the economic story.

21:26But when you think about sort of that, you know, you look at like Nvidia stock, for example, which is just like through the roof, you look at sort of how tech and AI is maybe propping up is not the right word, but it is a significant factor in the in the stock market right now. Then you just talked about the fact that you are having a very hard time finding tangible examples of corporate productivity being enhanced by this technology. What does it mean when those two things are combined? I mean, intuitively, it's really hard to imagine that these big AI companies, whether we're talking about the NVIDIAs of the world or whether we're talking about, you know, Google's had a phenomenal year as well, that the AI premium that they're receiving, the AI activity that they're receiving can persist forever if there aren't more clear-cut examples of companies saying, you know what, this is a valuable thing that I want to keep paying more and more for.

22:19So where does it go? I mean, one possibility is that the productivity enhancing use cases will be found, which is certainly possible. Well, Joe, do you use AI to enhance your productivity? So it's a good question. Like, I certainly use it for interview prep. Sure. Research. For research. But I don't know. But I put out the same number of podcasts as I did before. And they're probably about as good as they were before. I don't know. It's a really good question. I think many people would say this, that they actually use ChatGPT or one of them almost every day for something. I'm in that camp. Yeah.

22:58But whether I can actually point to something and say now I'm more productive is really tough. Yeah, like this is my killer use case. But you know what's weird? So I've been thinking about this question, too, which is the Internet has obviously changed our lives. We didn't have some mega productivity boom in the 90s. Productivity was a bit higher in the 90s than it was previous to that as the Internet was rolling out. But it was not like insanely different. And same with mobile phones. Like intuitively, yeah, the world's changed because we all have an iPhone or whatever. But has there been some measurable productivity boom that we can find in the aggregate in economic statistics associated with the iPhone?

23:39Not really. And so I think part of the puzzle of figuring out where the productivity measure is going to show up with AI, maybe we should just bracket aside the AI conversation for a while and go back to the 20 years of technological gains, which have completely changed society. So I think there's some interesting questions here. But then the other possibility, of course, is, yeah, that companies pull back and they don't spend as much and then you get stock market crash. Well, that sounds fun. And I'm going to ask you about that in a minute. But I'm curious about the psychology of all of this.

24:11Sure. Right. I've been wondering in the last few months, as we've seen more discussion around this idea of an AI bubble, how does that affect psychology? And is there sort of like the potential for a spiral from that, right, that people start reading about this? It becomes part of the narrative. And then and then you you do see a crash. Right. That's maybe like less related to a lack of efficiency gains and more related to people flipping out a little bit. It could be. You know, I think there's a famous George Soros quote, which is something like when he sees a bubble, he runs towards it. He doesn't run away.