

Written by AI. May contain mistakes.
Discussion of the concept of the triangle of confusion in economics.
“And I'm also joined by head of the Financial Times Lex Column and a man who sometimes wears shorts and an earring in the office at the same time, John Foley.”
Analysis of current consumer sentiment and its implications on the economy.
“So it's social media what done it, essentially.”
Examination of job market stability and its impact on economic outlook.
“moment when it was like unemployment rate is rising.”
Discussion of the current state of the housing market and affordability issues.
“if this is a safe space to ask stupid questions.”
Insights into market fluctuations and fears regarding AI's impact on various sectors.
“You have software companies, and those are being hit very hard at the moment because people have suddenly decided that AI is going to destroy the traditional software business.”
The hosts discuss their experiences with Duolingo and language learning.
“and it's going to get me something that works pretty well, you know, it's going to code me up something that works pretty well and is specifically for me, why do I need to pay my CRM provider anymore?”
A deep dive into the implications of generative AI on software companies.
“You were going to make an argument about Duolingo.”
Analysis of recent market trends and the impact of retail investors.
“But we're not going to suddenly all stop buying software and doing it ourselves.”
Discussion about the current state of the economy and confusion in the market.
“down a little bit over the last couple of days or week or so.”
The hosts explore the concept of data centers in space and its feasibility.
“John, what do you like or not like today?”

Get the gist, explore an idea or ask about something said.
On Unhedged, what are the key points of this episode?
Automatic transcript. May contain errors.
0:05Pushkin. We are lost in the woods and our compass needle is spinning. Economic indicators say one thing, markets say another, sentiment surveys, still a third thing. Today on the show, welcome to the Triangle of Confusion. This is Unhedged, the markets and finance podcast from the Financial Times and Pushkin. I'm Rob Armstrong. I'm coming to you from Unhedged World Headquarters in a still icy New York City. I'm very happy to be joined today by my co-author on the Unhedged newsletter and my indispensable aide-de-camp, Hak-Yung Kim. Very happy to be here. And I'm also joined by head of the Financial Times Lex Column and a man who sometimes wears shorts and an earring in the office at the same time, John Foley.
1:05Hi, Rob. Thanks for exposing my midlife crisis to the audiosphere. All right. The triangle of confusion, which is a bit of geometry I came up with literally 30 seconds ago, is a beast with three points. One point is sentiment, which says things are terrible. One point is hard economic data, which says that things are good. And the third point is markets, which say a lot of different things all at once. Today, we're going to try to see if we can change this shape a little bit, maybe into the circle of knowledge. Let's talk about sentiment. Hak-yung, you've written about what consumers say when economists ask them how they are feeling.
1:54What do they say? So if you go by just the sentiment numbers right now, the economy is worse than it was during the pandemic. We're seeing this big disconnect basically between the headline numbers, consumption, broader growth, even the unemployment rate, which isn't too bad right now. But consumers are just becoming more and more pessimistic as time goes on. And there are a lot of theories that kind of go to that. one point that's been brought up is we're just in a really negative news cycle and you have these algorithm-driven news feeds and people are just being bombarded with negative information about the economy.
2:33So it's social media what done it, essentially. Yeah, you can blame a lot of things on social media, this being one of them too, right? The fact that sentiment right now is worse than it was during the 1970s and 80s when inflation was far, far worse, that does suggest to a degree that it is this information overload that's overwhelming people too. There's also this view that it's hopelessly partisan, that all the Democrats say things are terrible and all the Republicans say things are fine. Is that fair? Well, definitely when Trump took office, there was a big divide, and there still is a divide between Democrats and Republicans on how the economy is going, but there has been a worsening of sentiment amongst Republicans as well that does suggest that, especially with the tariffs that really set a worry across consumers, that this fear is becoming a little more widespread.
3:26And I think by far the biggest, most common response we got to the sentiment piece is that people are really, really scared that AI is going to take their jobs away. I thought this was absolutely fascinating. You know, we sort of, we always in the newsletter, we kind of put it to readers. like we don't really understand what's going on is kind of a theme of the newsletter and please help us readers and it was like by far the biggest answer was ai scares the crap out of everyone yeah and it was it was crazy it was just people from all different ages people like new college grads saying that ai was worsening their chances of getting a job starting their careers but also So people fairly well into their careers were saying they've been at it for decades and now they don't have a sense of confidence anymore because you have AI that can say code within minutes.
4:21How's LexGPT going, John? I know that your columns are all written by an algorithm. Do you know the other day I did, I had an article to write and I was just like, I wonder what would happen if I asked Gemini to just do it for me. It was terrible. I was so glad. Oh, God, it was awful. But it won't be for long. The point is, what everyone says is like six months ago, people who wrote code were like, this code is so-so. And now you talk to experienced coders and they're like, oh, yeah, the machine's way better than me now. Yeah. And that fear is real. And, you know, I've been looking at this from a slightly different angle and looking at agentic AI.
4:58So the adoption of AI agents, which we may talk about later as well. but the adoption of these sort of fancy autonomous bots is really slow. Companies just are not putting them to work as fast as everyone kind of hoped they were, which is partly feeding into the worries about tech generally. One reason when I talk to executives about this is that employees are so scared of using it that they just kind of refuse. It's very hard to get people to use these tools. So now a lot of companies are forcing them to. They're saying, we're measuring your use of our AI tools, and you'd better use them, otherwise you're out.
5:32I mean, it's interesting you say that. I have a lot, you know, I'm in my mid-50s and I have a lot of friends who have kind of fairly senior roles in, you know, finance companies or whatever. And they kind of use these tools for everything but their core job. Like they don't want to write a letter to someone or draft a memo. They hate that stuff. And AI helps them with their calendar. It helps them write memos. But it's not their core job, right? It's not managing the portfolio. it's all the kind of stuff they hate on the periphery that they're using it for. But again, that's just anecdotage. But so we have this terrible background in terms of sentiment, like basically as bad as it's ever been.
6:11And when I look at the numbers, it looks like things got a little better sentiment-wise early last year. And now that has just kind of flipped over. Let's contrast that now with the economy. and the most important piece of news we've gotten out of the real economy, it seems to me, and you guys may disagree, is that the job market is stabilized. Two months ago, we were in this moment when it was like unemployment rate is rising. The jobs market is not adding any new jobs. And the few jobs it is adding are in like healthcare and other cyclically unimportant, non-cyclical sectors. And suddenly now it looks like jobs maybe are okay.
6:54I want to caveat that because apparently this morning we got some private payrolls data that is not looking very optimistic. There is the notorious challenger report, which came out today, which gathers official company announcements of layoffs. And that kind of rose a bit too. But I'm always slightly suspicious of that data because it's just the companies that announce. Anyway, yeah, it's worth caveating. But that's strong. Consumption numbers keep going along. earnings reports seem good. Is the economy just good? John? I think about share prices a lot, right? But like we think about companies and the value of companies and the value of a company is all of the future cash flows it will ever make discounted back to today.
7:42And in the economy, we don't really have that, do we? So it's all right looking at where jobs data is now and what people are consuming now. But it's the surveys and the sentiment that tell you what people think their future cash flows are going to look like, what their future salary income is going to look like. So clearly we're seeing that that is falling, even if today's activity still looks good. A person I was speaking to a few months ago used the phrase that I'd never heard before, which is the risk adjusted wage. And he used the example of a truck driver, which is a pretty well-paying blue collar job, long haul truck driver.
8:16But you know that the long haul truck driver is going to be a robot at some point in the future. So what's your risk adjusted wage in that job? It's falling. I guess another thing that's an area of worry, and this touches on your work on sentiment, Hak Young, is that the housing market's terrible. And Americans are so focused on housing. Two things are amazingly happening at once. A, houses are not affordable. And B, they're not going up enough in price to make the people who already own them happy. So like home seekers are unhappy and homeowners are unhappy at the same time. I'm fascinated by this dynamic because this is something Trump has talked about, isn't it?
8:56That he wants to bring, he wants to please both camps, both the people who don't have a home yet and the people who do. And I just wondered, is that even possible, Hak-Yong? Can you do both? Can you please the homeowners and the home seekers? I mean, I guess if I knew, I wouldn't be here, right? But yeah, no, it's true. No one's really happy with the housing market at all right now. Even if you are sitting on a home, you can't really get much out of it, right? Yeah, it's hard to sell, period, and it's not going up in value either. Yeah, but at the same time, people who haven't yet owned a home, that age is getting continuously pushed back.
9:32I guess I am evidence. I know very few people in their 20s who are actively planning on buying a home. It's just that's not within the measure of reason, it seems like, economically for most young people, which adds on to that element of despair, I guess, that we're already talking about. There is this affordability kind of, it's almost like a meme in describing the economy that only the rich people are spending any money, that the lower half of the spectrum is really under pressure. I'm not sure I totally buy it. Real wages are still rising. But to your point about risk-adjusted wage, like, would I want to commit to a 30-year mortgage or would a bank want to commit to lending me a 30-year mortgage if my job may not exist in five years' time.
10:22Yes. Yeah, yeah. No, I think that's right. That's another thing, too, I think, with housing, borrowing costs just being so high. We don't see that necessarily measured in a lot of price change surveys or data collection. Very good point. But that adds on to this level of expenses that people feel really overwhelmed by. And this is why the president is so focused on interest rates. He knows that they're going to affect whether his party does well in the midterms or not. If he can't get the 10-year yield down, he and Scott Bessana, I should say, can't get the 10-year yield down, and therefore the mortgage rate down, they're going to get smoked.
11:02But if you bring down the more, like I have a question about this, if this is a safe space to ask stupid questions. If you bring down the interest rate, which he's talked about doing to make housing more affordable, don't you just push up the price of houses? Doesn't that just create asset price inflation? It probably does. Famously, house prices are sticky down. Because when you get a sense that your house is less valuable than it once was, you just don't sell it. Right? So when house prices are under pressure, they just stop for a long time. They don't fall. But if you make mortgages cheap, that does increase the price of houses quite quickly.
11:41So you're absolutely right. You know, it's optics. He's got to do something. The president only has so many levers and he's going to pull them by God. I mean, he was talking about Trump houses. He was going to do some kind of public private partnership to build a lot more houses. I don't know if that's going to work. I mean, obviously, that's the solution to this problem is more housing units, less stupid zoning rules in America. But we've been talking about that for 20 years. Let us turn then to the third corner of the triangle of confusion, which is markets. We've had a weird couple of days.
12:18And just looking at what markets are doing right now, things are getting weirder still. Markets open down. The VIX is way up, the measure of volatility, but gold is down at the same time. We've had a massive tech company sell off in the next couple of days, but other stocks are doing well. what is going on, friends? These guys are just looking at each other like, you want to take that? Well, okay. I think that what is happening is that we're starting to move towards a more nuanced view of what we mean when we talk about tech. So the tech covers a lot. I mean, tech is 40 % roughly of the S &P 500 by market value, although that number is probably shrinking as I speak.
13:00But you have within tech, you have what we now call the hyperscalers, Google meta platforms, people who are building lots of cloud computing and AI stuff. You have software companies, and those are being hit very hard at the moment because people have suddenly decided that AI is going to destroy the traditional software business. Within software, you have a bewildering array of different kinds of software companies, CRM, ERP, observability, blah, blah, blah, all of which have different kinds of response to the new AI threat. But I think the worry, I don't know how you guys might feel different, and I'm now speaking from the gut here rather than from my great knowledge of software.
13:40The worry basically makes sense. you've been paying a subscription in the office for customer relationship software. Who do we sell to? What do they buy? How do we organize the information? Who's the contact person and all of that? Well, if I can whip up a custom version of that for myself by just talking to an AI assistant and it's going to get me something that works pretty well, you know, it's going to code me up something that works pretty well and is specifically for me, why do I need to pay my CRM provider anymore? My question is, how easy is that for just, say, you or me to do? Because I'm seeing stocks like, I think Duolingo is one of the software stocks that's getting really heavily hit, right?
14:28And hypothetically, I would like to not pay for - I don't even know what Duolingo does, by the way. What? Is that a joke? No, it's 100 % true. You don't know the owl? I don't know the The owl. Tell me about the owl. This is such a, like, okay boomer moment. You could be, we could be having this conversation in any of, like, ten different languages if you were a Duolingo subscriber. Oh, really? It's like a foolproof language learning app. You can learn Welsh. You can learn Sanskrit. Awesome. I don't even know if you can speak Sanskrit. It's a babel fish, like in Hitchhiker's Guide to the Galaxy.
14:57It teaches you languages. You put the little fish in your ear and you can understand all languages. And it has this very scary owl that will constantly bombard you with alerts threatening you to keep up your street. What are you using? What language are you learning with? Well, I'm trying Japanese, and I'm not going to try that on this podcast, but I have a 400, five-day streak. How's your Welsh, John? My Welsh is not good. I'm actually not using Duolingo. I like the old-fashioned way of learning languages. Okay. In any case, I'm hopeless with languages. This has all been a tangent. Back to you, Hakung.
15:25You were going to make an argument about Duolingo. My point being, say, I don't want to pay for Duolingo anymore. For someone like me or you, Rob, if you suddenly wanted to learn Russian, say, can we just go to our generative AI services and just ask, can you just make me an app that will teach me Russian or Japanese? Yeah, that seems far-fetched. And that's why, like, what you say, Rob, about this, this is the vibe code. There are two threats to software. And one is vibe coding, which is the idea that you just say to your LLM, you're like, make me something that can do my tax return and it just does it.
15:59And so you no longer need to subscribe to things like Intuit, to, like, software companies. that is massively overstated because if you ask companies what they're paying for with software they're paying not just to get the software they're paying to maintain it and the cost of maintaining the software is actually higher often than the cost of installing it so they don't want to just make something that might break that might be out of date that they don't they have to then pay to maintain they'd rather just get microsoft and salesforce to do that so although there are other margins you're going to be vibe coding is going to be a thing and lots of individuals that you say, how can you go playing with it?
16:31It's not going to eradicate the case for like paying a company to do it for you, knowing that your peers are using something similar. An analyst put it to me, I thought well yesterday, he said for a good software company, an established software company, the code is not most of the company. There's like distribution, maintenance, customer relationships, all this kind of commercial and intellectual infrastructure that's around the code. So he was like, the question is, which software companies are just a good idea written into code? Yeah, the code is cheaper and cheaper by the day. And which have that kind of larger economic and kind of social infrastructure that makes the business lasting?
17:14And maybe the market is trying to figure that out right now. So the first and the second threat, which I think is a more potent threat is not not just about you vibe coding everything for yourself. And so you don't use, I don't know, word anymore but it's about ai agents taking the place of software so instead of having your traditional software package will you have that like you you know there's a window and you type something into a form and it stores it in a database and gives you a dashboard will you have these agents that just do all that for you and that is a real threat because that's where we are going but the software companies are doing this like salesforce has agents service now has agents they're not they have they have sort of seen this coming so what i think we're going to get is a reshuffling and we're going to get some new companies that are what they call AI native that just sprang out of nowhere, basically, who are competing for that market share.
18:01But we're not going to suddenly all stop buying software and doing it ourselves. What's been wild in the last couple of days is that subjectively, it feels like the market is crashing, but it's not. Like yesterday, the equal weighted S &P was actually up and the S &P itself was down like half a percent, even though stuff is breaking all over the place, right? So there's like these offsetting gains and losses and the stuff that's gaining is boring stuff. Oil. Yeah, oil. Excellent. Campbell's soup, you know, like dull, opaque liquids are doing really well. Value stocks. And there's also this thing going on where that everything that did well last year is just automatically doing badly this year.
18:46I don't know how to sort of pick that out because it could be we're just shifting towards value. People want safe stuff, but it could also be people are just taking risk down, right? They're just like where I have profits, I'm going to take them. And if a stock is volatile, sell it, shift into Campbell's soup and ask questions later. Do you guys have any kind of vibes or feelings about what's been going on the last couple of days? I have kind of a far-fetched theory. My favorite kind of theory. Unsubstantiated, or I haven't really looked into it yet, but I do wonder if, because we've seen so much easy money come into the markets and you've seen areas like gold and tech and Bitcoin being flooded by all these retail investors in the past few years, right?
19:35And I wonder if maybe with what we're seeing just a broader worsening perception of the economy, making people more scared. And that's kind of leading into market spillovers into these areas that they've kind of flooded over the last few years, kind of triggering the insane volatility we've seen in gold, silver. I mean, I'm glad you mentioned Bitcoin because I wonder how many of these retail traders, you talk to people who trade at home, they all dabble, at least dabble in Bitcoin. And how much is that thing down? Bitcoin's fallen below like$70 ,000 now. So I wonder if Bitcoin is playing a role in market sentiment, even for people who aren't actually owners of Bitcoin.
20:17They see that as a measure of kind of froth or animal spirits. And so Bitcoin is kind of killing us. And weirdly, gold, which had been strong in the face of Bitcoin's weakness, is also now down a little bit over the last couple of days or week or so. Yeah, that's another corner of the market where there is a lot of retail investors flooding into gold-backed ETFs over the past few months. So a lot of the metals analysts, they're saying the price action that's, I mean, you have the central banks and then they kind of provide a nice price floor. And it seems like a lot of people just flooded into it with that assumption of safety into gold.
21:01And they've made loads of money and they might be thinking, well, take a little bit off the table. All right, Hakkyung, John, have we exited the triangle of confusion? I don't think so. I think we're caught deeper. Yeah, we went deeper. We've just turned into a pentagon. Yeah, it has more points than it did when we started. The polyhedron of confusion and we are immediately at the center. But as we journalists like to say at moments like this, hopefully we are asking the right questions and we will ask more questions after the break.
21:37Listeners, welcome back. This is Long and Short, which is the portion of the show where we go long things we like and we go short things we don't like. John, what do you like or not like today? Data centers in space. I love this story. Elon Musk wants to put data centers in space. He's merging SpaceX and XAI because he thinks it would make sense for them in space because there is no rent to pay. You've got the cut. It's noon all day, so you can get amazing solar power. And it's also very cold up there. Maybe we should just move to space. Sounds great up there. Can't be any worse. I just want to say that it is really difficult to put data centers in space.
22:17Even though it's cold, it's very hard to cool them. I wrote a story about this this week, and many of my readers told me I was an idiot. But I'm doubling down and saying that I think the future is data centers in space. How can it be hard to cool a data center in space? In order to cool something, the heat transfers through mass, through matter. And there's no mass out there. And there's no matter. So you have to radiate it. You can't conduct or conduct it. So you need enormous radiators. But this is the kind of thing that Elon Musk thinks he's good at. Hak-yung, top that. Oh, I can't top that.
22:49But I'm going to go long on tea, specifically matcha, after having given Rob his first matcha in the office this past week. And it was good. Katie once went short matcha because it tastes like dirt. But when you gave it to me, I was like, this is the most delicious dirt I've ever drank. I think when I visit London, I can make Katie a matcha and convince her to go along on it. Listeners, we'll be back in your feed next week. Until then, for goodness sake, be careful out there. Unhedged is produced by Jake Harper and edited by Bryant Erstat. Our executive producer is Jacob Goldstein. We had additional help from Topher Forhez.
23:31Cheryl Brumley is the FT's Global Head of Audio. Special thanks to Laura Clark, Alistair Mackey, Greta Cohn, and Natalie Sadler. FT Premium subscribers can get the Unhedged newsletter for free. A 30-day free trial is available to everyone else. Just go to ft.com slash unhedged offer. I'm Rob Armstrong. Thanks for listening.
24:01We'll be right back.
Markets, data and sentiment are often looked to as indicators of the present and guides to the future. But these data points aren’t helping as much as they used to. Today on the show, Robert Armstrong, Hakyung Kim and John Foley try to understand why major numbers are in such constant conflict. Also, they go long Matcha and long data centres in space.
For a free 30-day trial to the Unhedged newsletter go to: https://www.ft.com/unhedgedoffer.
You can email Robert Armstrong and Katie Martin at [email protected].
Read a transcript of this episode on FT.com
Hosted on Acast. See acast.com/privacy for more information.