원문 보기
Listening to Katie Wood respond to Bill Ackman, I really think no one is prepared for what is about to happen. After the Fed raised interest rates in September, Bill, who runs Pershing Square, warned that the hike could worsen inflation because the race to create artificial intelligence won't slow down just because loans cost more.
But Katie Wood, who runs Ark Invest, says otherwise. And she believes that AI is about to make so many things cheaper that prices may actually start to fall. So we'll hear from Katie, Bill himself, and Steve Eisman, who is the real investor of the character in the movie The Great Threshold. Halfway through, Steve says that the entire AI boom is based on just two companies that are losing money.
And by the end, you'll understand why I think Bill and Katie can be right, just not at the same time. Katie started the latest episode of her show, "In the Know," with a warning to Bill. And she starts it like this. So pay attention to what she says about inflation. In this episode of "In the Know," you'll hear our reaction to a very interesting tweet from Bill Ackman.
You'll also learn why we don't care about inflation and what credit markets think about this artificial intelligence revolution. Bill's question is simple. What if higher rates didn't slow anything down? Companies that create AI will continue to spend at any cost, so additional interest is built into the price of everything, inflation rises, and the Fed has to raise rates again.
But listen to how Katie responds, because she's heard this argument before. So, let's start with Bill Ackman's tweet. I find it very interesting, and one of the reasons is that when I first started in this business, I was in college, it was the late 70s, but this argument, the one that Bill makes, was common back then.
And it was based more on inflation, which was stuck in double digits. And this idea that interest costs will continue to rise, fueling inflation rates. So, it's a vicious circle. We actually think this time will be very similar to the 80s. In the 80s, we saw real growth due to productivity, lower inflation.
One reason for the decline in inflation is that we were experiencing a small technological revolution, started first by PCs and then by software. This contributed to the disinflationary effects we saw then. Today, we believe there are five major innovation platforms. So, artificial intelligence, robotics, energy storage, blockchain technology, and multi-ohm sequencing.
These five major technologically prepared innovation platforms will lead to a reduction in inflation to much lower levels than most people expect. And in fact, we wouldn't be surprised to see deflation, especially if oil prices rise as we expect. Katie had seen this before in the late 1970s, when inflation rose to double digits.
And that's exactly what happened. But she believes we are heading towards something more like the 1980s, when inflation declined. And she says the PC and software revolution was one of the reasons for this . And she goes further because she won't be surprised to see deflation, which means an actual fall in prices.
This is a pretty serious decision. But most of us have watched our grocery and energy bills soar over the past few years. So the idea that prices are moving in the opposite direction sounds a little crazy. So why should we take Bill's point of view seriously? Because when Bill made big decisions about the economy before, he put real money into it, and it paid off.
This is Bill on The Knowledge Project podcast hosted by Shane Parrish, and he's talking about COVID-era deals, not today. So listen to his inflation rate. And also, we are not just a long-term equity investor . Uh, we opportunistically made a lot of money on a few occasions when we had a different view of macroeconomic events than others.
And when you have that, uh, and you can find an asymmetrical way to make that bet, you can make a lot of money. So we made a lot of money starting with COVID because we were, I would say, weeks ahead of the world, and maybe even some of them were talking about the need for a global economic shutdown.
Uh, and we made a lot of money because they said there was going to be huge inflation. So we bet that the rates would go up. So this is someone who argued that inflation and higher rates would come when many people didn't, and he was right. That was a few years ago, and his warning today is related to that post on X, not this clip.
But it shows why people listen when Bill Ackman talks about inflation, but why he doesn't think AI is breaking the old rules this time. It all comes down to one phrase. Here Bill is talking about Brookfield, a company that builds infrastructure such as data centers and power plants. So listen to him describe the demand for computation, which essentially means the computing power for AI.
So, I think Brookfield is an incredibly well- run, amazing company. They are very good at infrastructure, they are very good at financing, they are very good at construction. So think about data centers, think about capacity. Um, you know, the demand for computing is virtually endless, and they're very well positioned.
So I think Brookfield is definitely a winner in the AI space. Virtually endless demand. If this is true, then the additional interest on the loan will not stop anyone from creating AI, because the prize is too great. And that's the point of Bill's warning, because if higher rates don't slow spending, they'll just make everything that spending touches more expensive.
But here's the interesting part, because Katie actually agrees with Bill about the demand. She's not saying that people will stop creating AI. So the real fight between them is what all this demand does to prices . So does Katie have an answer for Bill? She has, and she has a chart that shows interest rates rising for a completely different reason.
If you find this content informative, please click " Subscribe" because your support helps me improve these videos. So, here is her diagram. This is the yield on five-year Treasury bonds, which the U.S. government pays for a loan for five years. So, listen, what part of this factor is exerting pressure, inflation or real growth?
This is the yield on the five-year Treasury bond, and you can see the real yield. So, if you break down inflation and real yields, you can see that the recent interest rate change is much more dependent on real yields than inflation. So, you see, I'm very interested in this. I know fixed income investors pay attention to every basis point, and of course we in the equity market are probably more focused on percentage points.
So, 100 basis points. But you might think that sometimes, because fixed income investors and journalists talk about inflation, it means that inflation is getting out of control in the bond market. Uh, we look at this and say, “Okay, it looks like the bond market expects this 5-year inflation rate to, uh, behave pretty well.
Uh, yeah, it's above 2%, but it's been kind of stable over the last, uh, few years. "Uh, we actually think it will break, uh, once these, uh, the impact of technology, uh, is more and more understood by the markets, their impact on inflation." So, Katie's point is that there are two very different reasons why rates may be rising.
If they rise because inflation is out of control, that's bad news for everyone. But if they are growing because the economy is growing faster, that is good news, and she says her chart shows that. And the bond market isn't panicking about inflation either. People who borrow from the US government for 5 years expect inflation to behave just above 2%, but steadily.
So, if Bill was right, one would expect them to be the first to demand more for the risk. But what is inflation actually doing now? Katie raises Fed Chairman Kevin's score. Varsh says he is watching, and then there is a private indicator that is moving much faster. So, look in which direction the person is moving.
Now that's PCE inflation. We are focused on this because Fed Chairman Warsh is focused on the overall PCE inflation measure. So, we got an update when we first created this chart, it was 3.7% with the update, and there was a slight decline. We have decreased to 3.4% year-on-year. You can see that the baseline is a little lower.
The problem is that I think Fed Chairman Warsh is fixated on the headline figure , or at least that's what he says. And this is far from 2%. So we've heard a lot of "hawkish" talk about the Fed continuing to tighten monetary policy. The Fed will continue to tighten monetary policy until this PCE inflation report is released.
It's still very high, but if we look at the private inflation measure, which we think Fed Chairman Warsh is also focused on private measures . Here you can see Trueflation. We have already written about this before. You can see that if you look at the headline, you 'll see that it has grown. This is purple, up to about 2.7.
Oil prices likely drove much of the recent growth. But look what happens to the kernel. The core continues to fall. And we don't think this has escaped the attention of Fed Chairman Vorsch. And you can see that if you look at the numbers before the COVID pandemic, we are really moving towards the average values that we saw then.
So, the official number for August is 3.4%. This is well above 2% , which Kevin Warsh calls the Fed's hard-fixed target. But the private Trueflation indicator shows that its underlying indicator is still declining. And Katie believes oil is a big part of the recent growth. This is important because the Fed has to decide which of these numbers to believe.
And Katie thinks Kevin Warsh noticed the private indicator too. But Katie has already said she expects oil prices to fall . So listen to how low she thinks they can go. 57% of oil consumption is accounted for by transportation. And we think transportation is moving away from... Gasoline cars and engines elsewhere, and diesel, uh, trucks to electric.
Uh, and so we think that, uh, oil demand is in the process of peaking since the beginning of this, uh, transition. And it began in earnest in China. Uh, and Europe. Uh, the US, uh, follows. Uh, but, uh, we think the price of oil could fall to the $30 to $35 range. Uh, and the reason we say that is because as you can see during various crises, it has fallen into that range.
In fact, during COVID it fell well below this range. Er, and therefore, er, it won't be a crisis. It will be a technological shift that will reduce it. And this will certainly help inflation expectations. But the part that really resonates with Bill is how quickly AI is getting cheaper. And the number that Katie shows here is absolutely insane .
And based on that, we don't even use Wright's law. I'll give you the annual, uh, number. Er, the cost of AI inferences, that is, how much does it cost Chat GBT to answer your questions. Uh, and that's at fixed productivity. level. Uh, costs are falling by 99.99% per year. Er, and, er, and therefore any activity related to artificial intelligence will experience a significant reduction in costs.
Now, the Jevons paradox, uh , leads us to believe, and we believe, that as AI prices come down , AI inferences come down , volume will skyrocket, and revenue growth, I think, OpenAI, Anthropic, uh, I think, SpaceX AI, uh, and Muse, uh, these, the volume growth associated with that kind of price reduction, uh, will be explosive.
And so, uh, we saw, uh , OpenAI's revenue grow from 20 billion, which is the annual revenue growth rate at the end of last year, to uh, 70 billion. They just reported, uh, that's over the 65 billion reported for July for Anthropic. So I'm sure Anthropic is over 65 billion. So they probably go hand in hand.
Uh, but despite this price decline and, uh, income growth is skyrocketing. We've never...seen anything like this. And this is kind of a microcosm of what we expect for the economy as a whole. August saw a 99.99% drop in just 5 months in one AI coding test and about 99% for 2025. So, getting the same quality of response from AI now costs a tiny fraction of what it did a year ago.
And when something gets that cheap, people don't just use a little more of it, they use a lot more, which is why OpenAI's annual revenue has grown from about $20 billion at the end of last year to almost $70 billion. This is what Katie means by Jevons' paradox. When coal engines became more efficient in the 1800s, Britain didn't use less coal, it used a lot more because suddenly it was worth using everywhere.
And Katie believes the same thing is happening with AI. So, the price per response is falling, but the amount of AI used is increasing dramatically. So who really wins when AI becomes so cheap? Mostly businesses that use it because work that used to be too expensive, like responding to every customer email immediately or checking every line of a contract, suddenly costs very little.
And when a business's costs fall so much, competition drives down its prices, which is exactly the deflation that Katie is talking about. And I myself feel this demand, because I use more and more artificial intelligence tools in my own business. So I understand why Bill thinks the demand for computing is so high, but that's why they can both look at the same boom and see completely different endings.
Wait a minute. If AI is getting so cheap, why have the companies selling it simply raised their prices? This is Steve Eisman's opinion, and he talks to the New Money podcast here. So listen to what happened at Uber. Last year, OpenAI and Anthropic sharply lowered the prices of their services. Then they raised the prices because they were undervaluing the tokens so much that it was killing them.
So they raised the prices so that the customer would bear most of the cost of the token. So you attract people, once you... attract people, raise the price. Uber seems to have spent its entire budget on artificial intelligence in about three or four months. Oh, oh. OK. OK. And then it was some other company that I read about that spent $500 million, before I even knew it spent $500 million.
So, people have become much more conscious about spending. Token maximization is gone and people are much more into these open- weight models. You don't need a Cadillac for everything. You know, that's why people will only use Anthropic and OpenAI models for extremely important tasks. For everything else they will use Kimik A3 or something like that.
So what I want to say about this is that this industry, despite all the hundreds of billions of dollars that have been spent on it , has no obstacles. Hmm. There are no obstacles. The Uber story is confirmed because Uber used up its entire annual budget for AI by April, but no one has ever called the company worth $500 million .
So I take it with understanding. a pinch of salt. So, OpenAI and Anthropic were undercharging. So they raised prices and many companies started using cheaper open models. But note that this doesn't hurt Katie's argument, because these cheaper open models are AI that is also getting cheaper. Where this hurts is the companies that spent hundreds of billions to build all of this, because Steve says none of them have a moat, which essentially means there's nothing stopping a customer from switching to someone cheaper.
And that brings us to the biggest problem in this entire video. So listen to the story that Steve uses to explain this, because it's about a company with only two customers. This is how I look at it . The concentration risks here are staggering. You know, so if you take a step back and you and someone say to me, " Why don't you analyze this software company?
Forget about what it does. It's a software company, you know, an established software company. And if it turns out that the company has thousands of customers, that would be great." If it turned out that the company only had two customers, you would say, "I don't want to invest in this because if something bad happens to one of those customers , this company is dead." Something similar happens throughout the history of artificial intelligence.
Now listen to two clients. Steve says everything depends on them, from Nvidia chips to giant cloud companies. And I'll be frank, I own Nvidia, so I'm referring to this particular chain. And my point applies to the entire chain from Nvidia to hyperscalers all the way down. The entire chain relies on the future health and growth of Anthropic and OpenAI because they are creating a commitment to hyperscalers.
If they don't grow and don't have the money to pay these obligations, then the whole chain will slow down. Yes. So, it's a risk. I would say that between the two, OpenAI is probably the weaker side, but it's unclear. Because in reality we don't have enough numbers. All we know is that The Wall Street Journal reported it, so I assume it's true.
In the second quarter of this year, OpenAI's revenue was about $6.5 billion . Good. Anthropic was at over 11. But that's just revenue. Just income. Good. OpenAI lost something, the costs amounted to about 12 billion. So the math worked like this: in 3 months, OpenAI's revenue grew by a billion, and its expenses by 3 billion, which, as we like to say, was turned upside down .
You want the opposite, not the first. They want a different path. Great business model. And their revenue grew by 18% in 3 months, while Anthropic's revenue grew by over 100% in 3 months. So, they are a weaker company at the moment and that could change. My only thought is that this whole industry makes me nervous, because let's just imagine OpenAI failing.
This can happen. All you know, the entire United States of America will fall into recession overnight. If that's the case...And now eventually there will be a lot more diversification, and there will be a lot more companies, but that's over time. So, you know, over the next year or so, these companies should remain healthy .
That's where I think we are now. So, there is a concentration risk. One number that needs to be corrected because it shows that $12 billion was OpenAI's loss for the quarter, not its expenses. So they actually spend even more than Steve said. I listened to Steve describe the money they were losing. While Anthropic told investors it had a small adjusted profit in the second quarter.
To say they are losing money would be a good thing. They are bleeding. Yes. What's the nicest way to say this? They are losing a lot of money. We are up to...Yes, we are up to... profit. To...profit. Something like... They make money, excluding all expenses. Yes, exactly. That's how they like to think about it.
That's what we should start reporting on this metric. Yes, that's how it should be. Um, so where does the money come from? Well , that's actually a very interesting question. I thought most of the company's money came from venture capital. But this is not true. Most of the company's funding comes from Amazon, Google, Microsoft, and Nvidia, which invest in these companies.
And SoftBank. OK. So, taking shares in... taking shares as capital is raised, and these guys put money in. Whether they want to continue investing, I don't know. It looks like a circle. It sounds like money or obligations go in one direction and then obligations come back in another. Yes, it really does have that tendency, doesn't it?
It reminds me a little bit of... It's actually a scene from the movie "The Great Depression" with your character, to work it in there, right? I remember him saying, "Oh, this is something like CDO A, and then they put parts of that into CDO B, and then those two put into CDO C." There are some similarities.
But, but, in their defense, there is a cyclical nature to funding. But as long as Anthropic and OpenAI continue to grow very, very quickly, and people continue to give them money, the chain will survive. That's when, if one of these two companies really fails and people withdraw their money or no longer want to invest , the chain won't survive.
This is where the problem of concentration risk arises. If the industry was much... If... If ... If... If... If... If... If I said that instead of 70% of hyperscale AI revenue coming from Anthropic and OpenAI, if that number was 10%, we'd be having a very different conversation because, uh, obviously there are a lot more large-scale customers.
This 70% is an analyst estimate, not what the companies have published, but it is a valid concern. OpenAI and Anthropic have promised to spend huge sums on leasing computing power, and cloud companies are building data centers based on these promises. So, if OpenAI ran out of money, these data centers would be left with fewer customers, and that would hit everyone from Oracle to Nvidia.
But Steve is honest about the other side. He says that as long as these two companies continue to grow very quickly and people continue to fund them, the chain will hold. And Katie's numbers show they're growing faster than ever. So, for now, the chain is holding. So, here's what I'm getting at. This is a risk that needs to be monitored, and if these companies fail, the house of cards will fall on them .
But I think that's becoming less likely every day, and I don't think it's fair to say that the chain is relying on these two unprofitable customers, because it's relying on a lot more things as well. It relies on the laws of AI scaling that remain relevant, i.e. the idea that larger models trained on more computing power become smarter.
It relies on the value of the application layer, so applications built on AI generate real revenue and actual profit. This is based on continuous implementation and the assumption that there will be no disaster. Bill doesn't ignore risk either. Chamath Palihapitiya asked him directly if we were in a bubble, and listen to the two things he said at the same time.
Are we in it now? We are definitely in a transition period. You could say that AI is absolutely transformative, and it's the most transformative technology, in my opinion, in our entire lifetime. There are elements that look like bubbles. There are a lot of crazy things going on in the private venture capital world, I think.
We're starting to invest in some... Speaking of the bubble, we at Virgin Galactic are investing in some venture capital. You know, I met with one company, let's say, 2 weeks ago, and they weren't raising money. They closed the round 2 days later. It's as if they were outbid by some investor who valued $ 50 million at $400 million.
And then 2 weeks later they raised another $50 million into a billion dollars. You know, a lot of people are afraid of losing, you know, the future. And that 's why there's a lot of capital invested in venture capital, a lot of competition, and a lot of preemption. And this leads to very high ratings.
You know, I saw a company, you know, a Series A round with a $5 billion pre-funding. So Bill calls AI the most transformative technology of our lifetime, but he also saw a startup raise money at a $400 million valuation and then at a $1 billion valuation 2 weeks later. And this happens in private, venture-backed startups, not in the big companies you see on the stock market.
I agree with Bill on the first part, because I don't think AI is a bubble. I think the demand is real because people and businesses are actually paying for these tools and using them every day. But listen to what Bill thinks will happen to such startups next. What advice would you give to founders right now who may be struggling to raise funds?
Now is a great time to raise capital. And if you raise capital, don't spend it so quickly, expecting the money to stay there forever. This is reminiscent of the fact that in the Internet bubble there was virtually unlimited capital available for rather dubious business plans. And then one weekend there was an article in Barron's magazine.
It listed all the Internet companies on the public markets and how much cash they had left, based on their burn rate and the number of months before they would fall to zero. This was probably a week before the market exploded. With a private company that has no capital, there is a certain discipline associated with self-development.
And then you take 50 million, or 100 million, or completely change the way you allocate capital, betting that when you need money, you can just take advantage of the markets. The time will come. There will be some kind of explosion, quite loud, when people lose a lot of money. And that will lead to a reset, and I think the companies that are disciplined in spending capital to have years of success will survive, and the ones that have to raise money in three months will disappear.
My advice to founders: treat every dollar as if it were your own money, spend it very carefully, and don't expect such viable, freely available capital to exist forever. A loud bang, then a reboot. So Bill believes that some startups spend money as if it will never end, and that some of them will disappear.
So even Bill, who believes the demand for AI is virtually endless, expects some of the money chasing it to burn. But Bill's warning has an even bigger second half. The US government also has to pay these higher rates. Steve was asked exactly that. So listen to why he doesn't lose any sleep over this.
US debt. This is such an interesting topic, and it is covered very, very, very widely . So now it is $40 billion. The average interest rate on it has increased from 1.77% in 2020 to 3.45% today. Interest expenses rose from $523 billion a year to $1.22 trillion . Is this something investors should really worry about?
Is there a real risk of a debt spiral in the future? I mean, all else being equal, I would like the deficit to be smaller. Yes. Hmm, I have my doubts about the debt spiral. First, we are the world's reserve currency . But perhaps more importantly, U.S. Treasury bonds are the financial system of planet Earth.
So, for example, banks around the world do something called repos, where they lend to each other overnight. They do this through US Treasury bonds overnight. So, as long as the US Treasury is the backbone of the world's financial system, I tend not to worry too much about the deficit, although I would like it to be smaller.
If there was an alternative, we would have problems. Good. But at the moment there is no alternative. So, the interest the US government pays on its debts has more than doubled since 2020, exceeding $1.2 trillion per year. It is these kinds of costs, according to Bill, that are reflected in inflation.
But Steve's point is that the entire world lives in debt to the US government. Banks even lend to each other overnight using it, so that money has nowhere else to go. Until something replaces them, he simply doesn't see the spiral that is happening. So, with all these loans, what interest rate should we really be watching?
Steve says the 10-year, long-term rate is important because that's the rate at which people actually borrow. And he calls it a number. For example, today the markets rose because, despite the Fed raising rates, the yield on 10- year bonds fell. I have the impression that 5% is a kind of Rubicon for the market.
And, um, as long as we're below that level, we'll probably be fine. But if something happens and we get through it, I think we'll get a correction. Good. Because that's roughly my assumption. I actually thought the number was 4 1/2% at first, and I was wrong. Good. But five seems better. Do you know where it is now?
This is about here. This is 4.98. 4.98, good. But yesterday it was over 5% and it has gone down again. Good. People are watching every tick. 5% on a 10-year basis, that's what the US government pays for 10-year loans, and that's important because mortgages and loans from large companies are priced based on that.
So when it rises, loans become more expensive for almost everyone. By the time Katie recorded her episode, it had risen above 5% again, which is exactly the line Steve had observed. And Katie has her own schedule for this line. So pay attention to where she says today's bet actually is. But what I would like to point out in this graph is that we are seeing a significant spike in interest rates.
And, boy, does it feel like a big splash. We dropped below one, I think, during COVID. And now we've surpassed five in 10- year bond yields. But look at it in the context of history, and you'll see that today's rate is actually in the middle, and perhaps in the middle of where we've been on average since 1790, although certainly higher than we've been for much of history.
Hmm, and one more thing I would like to point out is that the anomaly from a historical perspective is not what is happening now, but what was happening in the 70s, 80s and 90s. And indeed, before COVID. A significant spike in rates is the number where Bill's anxiety and Katie's optimism truly meet.
If the 10- year bond continues to rise because the economy is booming, that's a good move for Katie, but if it rises because people who borrow money are afraid of inflation, that's Bill's warning coming true. But even Katie sees some burden on borrowing. So listen to what she has to say about giant cloud companies and the cost of insuring their debt.
And there's this concern about credit default swaps, this concern about hyperscalers. And interestingly, look at the green to see that these, in the context of a short history, uh, they were monetary fortresses for many years, and they did n't need to go into the debt markets at all. And yet, uh , they're entering the debt markets, and it's a pretty unusual moment.
So, you know, their credit default swaps are getting worse, the rise means they're getting worse. Uh , and, uh, we see the investment grade growing, it's in black . Uh, high-yield, junk bonds, in purple, they're going up. Um, and I'm sure non-cloud service providers are largely responsible for this.
So Katie doesn't pretend this construction is free. Someone is paying for all these data centers now, and more and more of them are borrowing to do it. And so I think Bill is right in the short term. The cloud services she mentions are newer companies that rent out AI computing power, and they took out large loans to buy the chips.
So if anyone is feeling the pressure of higher rates first, it's probably them. But this puts pressure on companies that build, while AI itself continues to become cheaper. And that's where, according to Katie, it all ends, and I do n't think anyone is prepared for what happens. We expect that as this price decline penetrates deeper into the economy, and it will, uh, that uh, the volume growth associated with any activity will accelerate quite significantly.
And this is how we get the forecast for external GDP growth for the coming years. So, here it is , uh, in action, so to speak, how we look at the world. We look at the world, as you know, through the prism of technology. And here you can see 15 technologies participating in these five major innovation platforms.
And you can see them getting closer. And you can see how AI, highlighted in purple, has the biggest impact on other technologies. So the cost reduction that affects AI will affect these other technologies. And I often say, when we first, and I have to give credit to our chief futurist Brett Winton for drawing this, when we first drew this diagram, I think there were some boundaries between the golden ratio, blockchain technology, and AI.
But now they are thanks to agent payments, agent commerce, Grokbot, Muse, and now bots from OpenAI. So we think we're going to see a lot of deflationary pressure, but it's good deflation. That is, it causes a growth boom. Good deflation. Prices fall because things have become cheaper to produce, not because no one wants to buy them, and the economy grows faster because of it.
This is very different from the bad things like in Japan for years, where prices fell because people stopped spending. So what will happen? I think Beck Mann and Katie Wood might be right, one after the other. While AI is being created, the costs are huge, and that drives up costs and interest rates, which is what Bill is worried about.
But once it's created and used, everything it touches becomes cheaper, and that's a good argument for Katie about deflation. And if Katie is right, then every business that uses AI to do more for less is among the winners. Customers get cheaper products, businesses get higher margins until their competitors catch up, and the entire economy grows faster.
This is the boom that Katie is talking about, and that's why Arc's growth forecast is so much higher than almost anyone else's. But for that to happen, the chain has to endure, and that means more than just OpenAI and Anthropic paying their bills. This means that state laws are in place, AI-based applications are profitable, and people continue to use them.
So I'll be watching the next inflation numbers and this 10-year 5% mark, because they'll show us where we really are in the story.
Comments 0