From World Simple Media in Toronto. This is the trade show. I'm Swapnel Periq, VP of product at World Simple.
And I'm Matthew Carris, another VP of product at World Simple.
So, Matt, we're alive. We're here. We didn't die. It's been crazy the last month where everyone just talked about AI doomism and how the world was going to end and we were all going to die.
Yeah. I mean, I do have friends where it's like kind of painful to go out to dinner with them now because they're like mad, like we might still be alive right now, but you don't know. The future is going to be absolutely crazy.
And at the same time, these people are using AI every day and they're playing with it. They're talking about it, but they're also going to die from it.
Yeah. Well, I mean, this is the discourse. I mean, I think we should just get right to the charts. Like this chart I think is probably the single funniest chart I've seen in the last 10 years of loving charts and watching markets. And so, you know, if you look at it, it basically shows long-term real GDP over time. And actually if you squint at the x-axis it goes back all the way to 1870. And basically like, you know, real GDP, you know, production and consumption has been going up pretty steadily over time. You can see the little wiggle of the Great Depression in there. And, you know, at the end you see sort of the range of outcomes from AI based on what people are talking about here. And this is where it just gets, it's just hilarious. But this is like what people are saying, like, you know, the base case is like, you know, ongoing maybe like, you know, a little bit faster growth. We got the AI accelerationists like Marc Andreessen, you know, like thinking that we're going to have the singularity and it goes up up there. And then, you know, you have the doomers like Scary Potter thinking it's going to go down to zero. It's, Elon Musk I think said a few months ago, he said something like we don't have to worry about money. We don't need money. It's all going to be abundance. And so clearly there's people that believe that AI is going to take us to this land that we've never been in.
Yeah. And look, and I don't know, people are probably going to say that I'm like being flip about this and my friends actually in fact like say that I'm just completely naive, but like I don't even know, like what are you supposed to do with the line going down to zero? Like it's like how is that supposed to impact your trading? And this chart was published. It was like in the newspaper.
Oh, that was, it's even, and it's not a joke. Like this was, this is actually, this chart was originally published in the FT in December 2025. And it's from a Dallas Fed article where they're talking about the potential impacts of AI.
It's incredible. And you think about like the discourse on the internet, like it's wild. So, like Labor Day weekend's happening and there's this paper put out by Dario from Anthropic about the evils and like the doomism of AI. And literally two days later, he's at Dreamforce doing the keynote telling everyone how AI is going to be the future and it's going to be the best thing ever. And then two days later, Muse comes out and Instinct launches and now everyone's got an agent in their pocket. And it's like it moves fast between like you're gonna die to everything's going to be amazing and you're gonna have the best tool available to you.
Yeah. I don't, anyways, I don't even know what to do about it. I mean the main thing I keep going back to is like if civilization shuts down, like your entry in human society's great big spreadsheet is not going to be worth anything anyways. And so you should just count on that not happening and keep going. At least that's what I'm doing.
Exactly. You should use the tools. They're really interesting. All right. Well, since we're still here, let's talk about something completely different. Let's talk about bonds.
That's the other big thing going on in markets right now.
Everyone's freaking out about bond yields. They're up 50 basis points in the last couple months. The Fed started raising rates for the first time in a few years in the US. The Bank of Canada is now rumored to raise rates at their next meeting. Investors sort of believe that rates will rise over 1% over the next year or so. And this sort of makes it seem like rates are just going to keep going up. And, you know, like people get very scared when rates are really going up on the long end of the curve.
I mean, there are these people on Twitter, I mean, I forget some of the data, right? There's some of these people that like they just post it and they talk about like how like the government's going to go into, you know, going to default on its debt and like it's like, you know, they show like charts of like the public debt over time and the interest costs, you know, for servicing it. And these people have been having a field day basically like they're predicting like the great big crash, the crash to end all crashes. The best chart about this like from the doomers was the last time the US rate was here the S&P was down to like 4,000 or something and they draw lines, flat lines from here to there and they're like that's where we're going and it's like what are we talking about? I mean, it's not as crazy as like or as iconic as the like Markets in Turmoil from CNBC, but I did see like a tweet from them this weekend where they were talking about like when rates rise this fast, it's like calamity in markets waiting to happen, but then like nothing happened.
Yeah. So what do you actually think is driving it?
Well, I mean, you know, there's some people that are saying that it's like really like the government borrowing. There's some people that are saying like, you know, inflation expectations have gotten completely unhinged. I think probably like the main, you know, the best take that I've seen is one that was sort of popularized by this markets watcher and economist Matthew Klene that is really just saying that like it's a response to, you know, stronger growth on a backwards looking basis and expectations that growth will remain stronger and rates will need to be a bit higher.
Yeah. Like I sort of see it as like if you actually look back over the last, I don't know, if you go, I think we have a chart that goes pretty far back, but if you go back to like even the last 20 years or 30 years, rates actually like kind of move with GDP. So like as you see GDP going up, rates go up and as a result of it, like rates are just a response to like growth. They're not a response to any of these other like other noise we hear in the market.
Yeah. These charts just, so we're all clear, we're looking at, for the US and Canada, for each of them it basically shows back to, you know, 1958. I can't believe we got the data back that far, but it's easy to get with the AI. You can see basically the blue line is nominal GDP growth and the orange line is the 10-year government bond yield and you can see that, you know, while there are lots of squiggles and it's not sort of a perfect relationship, you know, basically the big increases and decreases or the trends in the bond yield follow from, you know, trailing growth. And so when the economy has been running stronger, when people have been increasing their production and their spending at a faster rate in nominal terms, you get a higher bond yield and on the flip side, you know, when it's lower and it's, you know, people were not doing all that much like in the 2010s and there was basically no inflation, you had bond yields that were a lot lower. And what this chart shows is that the rise in bond yields we've experienced, well, it's felt like a lot up close, is actually like even less than what might have been called for just by the change in the economic conditions that we've had over the last couple years.
Yeah. And so one of the other things that I think this JP Morgan analyst put out, Michael Cembalest,
Cembalest, if you think about AI growth and what the AI hyperscalers are doing, that it's really a supply and demand question. They're demanding a ton of capital that they got to fund some trillion dollars of expense over the capex over the next few years and so they're borrowing money and so when you borrow money you're willing to pay at a higher rate on an ongoing basis. And so I think if you think about it, like in the 2010s, the biggest borrower of long-term yields was the US government or the Canadian government or just governments in general. And now you have private companies spending a lot of their time selling debt.
Yeah. So like the biggest pushback that somebody would have to like these charts is like, Matt, like, you know, the lines like sort of move together but like it's not that big of a sample size and like when you look at things in level terms like this, like, you know, actually, you know, they don't have that much, you know, relationships on a near-term basis. It's really just, you know, a couple big moves that are making it seem like there's a close relationship. But I think the answer to that is like, hey, like there's actually a logical relationship here too, like the one that you're saying. And when you actually think about it, like AI has been a, you know, the shift to AI has just required like a lot of new building of stuff and all that building has required a lot more money and there's so much, you know, at stake, like so much profit at stake that people are just willing to pay a whole lot more to get that money. And I think the thing that you're saying of like to put the size of the investment in context, it's like we have an additional US government going out there and borrowing every year. Like that's basically what it is between the hyperscalers and Nvidia, they're borrowing 70% as much as the US government this year.
On top of what the US government's already borrowing. And when you have that many more borrowers out there, it's like hard to imagine that it's not going to cost more to get that money.
And if we look back, like let's go further back. I know you love going further back, 1870. I think we have a chart here that goes back to 1703.
So like the question I ask is like are rates actually even that high?
Yeah. So just pull this. I love charts like this. Again, like if people just want to know what the power of AI is, like it might kill us all, but it will make some great charts along the way. And what we got here is we have Canadian bond yields, US bond yields, and UK bond yields as far back as the AI could possibly find them. And I did check the sources. These are from government sources. Apparently, it is right that in like the 1700s in the UK, there was some young clerk that was like writing stuff down on parchment, keeping track of how much the UK government was paying to borrow. And, you know, basically the way to look at this is like the 50 basis point move that we've seen over the last year and even the move since like 2021 into 2021 feels big. It's like we went from like, you know, very low rates to what feels like high rates. But if you look back since 1980 we're still at the low end of what happened there. And if you look back to, you know, 1950 or 1850 or 1703, it looks like we're right at the average.
Yeah. And if you kind of put this now into context of like our lifetimes, like in the '50s, '60s, '70s, there was a lot of build going on. Like we had the industrial revolution, we had World War II and there were a lot of growth in GDP that kind of led to higher rates. And so maybe you do get higher rates, but you also get higher GDP.
Yeah. And so like what one of my old bosses used to like basically beat into us when we were analysts and trying to trade was like if you're trying to use history to trade, like really, you know, the question is like are you looking at a representative sample? And because it really goes back to like this question of like what is normal or what is crazy because a lot of what you're trying to do is bet on like some regression or reversion to normal. And here like what you think of as normal is just really dependent on what period you're looking at. Right. And I think this, you know, sort of plus like the logical incentives for more borrowing makes me think that like the move in yields is not that crazy and it doesn't necessarily mean that something has to break. It seems like this could actually just be the new normal.
And so do you think, I guess one of the questions that we have to ask ourselves about what's happening is higher rates, does that mean that stocks are going to collapse? Do our customers feel that stocks are going to collapse? Like I don't really think that. I think the way I would kind of think about this is that rates are going up, but earnings are going up faster and earnings are outpacing inflation and the increase in yields.
Yeah, there's some great stuff out this weekend about that. Our friend at Random Walk has some really good stuff. We should have pulled some good charts on this. We'll do, you know, the next time we have one of these shows, we can do a whole thing on whether stocks are expensive or cheap. But yeah, I mean I think that I think you actually look at it like a lot of people would come out and say okay because the bond yields are rising and the rates are rising that means that like stocks have to fall in the future. You know, the flip side is that like hey, like if the bond yields are rising or the rates are rising because the economy is so strong and because companies want to invest and spend so much in a way that's other companies like profits, you know, it doesn't necessarily mean the stocks have to fall. And you know, I think it's probably too soon to sell but if you look at what our clients have been doing it doesn't really seem like they're reacting in a way that would, you know, make them think that there's a big blast coming.
Yeah. Clients are buying fewer single name stocks, but we had highs just as rates were going up and yields were going up and there's a little bit of a shift in AI stocks. I mean, I think customers are selling some semiconductors, some AI stocks generally, but I think that's more of the ebbs and flows of how people are thinking about AI and earnings season, but generally speaking, customers are not selling their stocks because yields are going up. They're kind of in a wait and hold pattern as we all figure out what this actually means.
Yeah. Well, so just to make sure everyone knows what we're looking at here. So, what we did, and you know, I'll take credit for this one. I pulled these charts, but what I did over the weekend was pulled the data on how much our clients have been buying stocks. And on the left, you basically see the net purchases per month of stocks. Up is positive, so they're like putting in, you know, hundreds of millions or billions of dollars into stocks, and down would be negative. And you do see like a huge spike in May and June. And you see, you know, a bit of a falloff since then. And so, you know, one way of reading this is like, man, they're coming like a lot less bullish than they were before. And that could be like, oh, like are they being spooked by the rise in yields in a way that would, you know, suggest that they're going to buy less stocks, there's going to be less upward pressure on stocks and there, you know, might be easier for some other risk to come in and like overturn the apple cart. On the other hand, if you actually look at it, like the main reason that their buying behavior has shifted is really just, you know, how much AI stocks they're buying. They like loaded up on the semiconductor stocks in May and June and then, you know, bought less of those over the summer and even sold into September.
Seems like they're just holding. I mean, it's not, they're not selling. And I think cash is at an all-time low, I think, in the market. And so, I think it's a wait and see situation.
Yeah. Yeah. I mean, I guess that's my take as well. I mean the other chart that I find really interesting to look at is the one that you just mentioned which is like, you know, are you seeing them instead of like just selling stocks also going into cash because like I think that like the real bear case of the rising yields is like oh if like rates are higher it's much more attractive to hold cash. Are you seeing people shift from stocks into cash and that is not something we've seen and it's broadly in the market we haven't seen it. Cool. Well, before we bring out our guest, we have a quick word from our sponsor, Wealth Simple.
Matt, so we both work on product at Wealth Simple. We build products for our customers. Maybe we should talk about the two things that we're most excited about that we're about to launch or we've launched in the last few weeks.
Yeah, I mean, we were talking about this earlier. I think the one that I'm most excited about is bracket orders which I think has been our single biggest client request over the last couple years and, you know, mostly I've seen like the comments people don't think we're listening to them but we have actually been listening to them and we've actually gotten it to a place where we're using it internally.
Yeah. And so just for everyone like what a bracket is, it allows you to set exit conditions as soon as you buy a position. So you can set a price in which you want to sell and also limit your losses. And if any of these orders execute, the other order is automatically canceled.
So it makes it like really good to trade in very fast markets. So like meaningfully decreases the amount of time you have to spend like putting in, you know, your conditions over and over again. And we're hoping to get this out to our clients in the next couple weeks.
Yeah, I'm super excited about it. The other thing I'm really excited about is we made significant changes to our crypto pricing. We made it a lot easier and a lot cheaper for people that trade a lot of crypto.
Yeah. I mean, the thing that I'm most proud about this one is that for people that trade a lot of crypto, like the biggest traders, those people, you know, there's no better place in Canada to trade crypto than Wealth Simple after these changes.
Yeah. And that's super exciting. I mean, I think crypto markets are on the up and up and I'm super excited for our customers to try it. Should we get back to the show?
Our next guest is Raquel Urtasun, the CEO of Waabi. She has been working on AI from before you were born.
She probably has all the answers to every question we've talked about on this podcast for like the last five, six weeks.
Yeah. I mean, that's what really sold me on this interview, which is that, you know, we've been talking about AI. We've been talking about how it's going to impact, you know, the price of this. It might kill us. Like we have a compute shortage. Memory. I mean, she actually might know what those words mean.
Yeah. Yeah. And the coolest part is that like Waabi has taken a completely different approach to how they build self-driving cars, autonomous driving for trucks. They don't build cars. They build software. And that software trains hardware. And they put the hardware on a car.
Yeah. This also gets to like one of the big questions I've had is like, okay, maybe AI is a thing. Maybe it's going to impact different parts of the economy in different ways, but where's the value going to accrue? And this is like a very contrarian take. Very different from like, you know, Google, very different from Tesla. And the things that I want to know are like one, am I ever going to have to drive again? Am I ever going to have to sit in traffic? Is this it? Have we reached that top of that chart?
Well, let's see what she has to say about it from FT. Yeah, let's bring her on.
Hi, Raquel. Great to meet you and thank you so much for joining us here on the trade show.
Really a pleasure to be here today with you. Yes.
So we're going to get into it. Matt here is about to go on parental leave. And we were wondering, do you think his kid is ever going to need to learn how to drive?
Really going right for the hard-hitting questions.
Yeah. Yeah. We need to know the answers.
Yeah. It's the question that everybody's wondering. Well, absolutely not. There will not be a need to really learn to drive and actually much earlier than when it will be needed for your kid to do. So think of it as it will be really like a curiosity if you want to drive a car versus, you know, something that you need to do every day.
It's kind of like when I like to go to a racetrack and drive a car just for fun. Like I don't have to do that but I do find it fun to drive an F1 car.
What's the, just so Raquel, like I've been hearing about this idea a lot increasingly and there are a couple of cities that have, you know, toy examples or small scale programs of this. Like how soon do you think it's going to happen that like autonomous driving is going to be widespread?
Yeah. So I think we can also talk about what type of autonomous driving is going to be widely spread. I think that autonomous trucks will actually get to a scale earlier than robotaxis where when you think about long-haul trucking, a job that humans don't want to do, is really ripe for this technology. So you will see tens of thousands of, you know, by 2030, of self-driving trucks on highways as well as, you know, around city areas. And then when it goes to robotaxis, we've seen deployment today, but you will see increased deployments starting in the US hopefully soon, Canada, and then expanding, you know, to Europe etc. But it's a bit harder to disrupt the personal ownership of vehicles in North America. You know, oftentimes a car is a symbol also of freedom and identity, right? So it will take time to disrupt that versus if you live in a city, you know, taking a robot taxi is definitely something that will be, you know, part of your day-to-day very very soon.
Do you think, I guess one of the questions that I've sort of thought about with the self-driving trucks is do you think that these trucks that are self-driving will coexist with other trucks that are being driven by humans? Do you think there'll be separate lanes for them? Like how do you sort of see like a world where there's 354,000 trucks on the road that are self-driving? How do you see the coexistence of both human and autonomous driving?
Yeah, they absolutely need to coexist. You can't wait for infrastructure to be deployed so that self-driving trucks run in their own lanes. Instead, you know, the product that we are bringing to market is a product where it can drive simply than humans and it can do in coalition with what human drivers do today and that's actually part of the challenge because humans are very unpredictable and they're not always driving really well. But you still need to handle other uncertainty and then provide, you know, an extra level of safety. But yeah, you can't wait for, you know, different infrastructure that technology will never really truly be deployed at the scale.
So can we talk about that prototype? You know, one, I heard about it. I was a little bit disappointed, just a little bit disappointed to learn that you didn't want to drop the announcement on the trade show, but we'll forgive you for that. If you would just tell us a little bit more about what you're building and what the news is.
Yeah. So if you look at, you know, we're self-driving trucks for almost, you know, for four years or so, we've been doing commercial operations in Texas. So if you come to, you know, Texas, you know, highway as well as surface street, it's very likely that you will see one of our self-driving vehicles on the road. What they still have today is a safety driver that is there really to monitor the vehicle. And what is coming really early next year is that you will be able to see our trucks without a human onboard. And what we are waiting is on the final validation of the vehicle platform by Volvo, which is our OEM partner that provides a special type of trucks where our technology, including the AI driver, goes on top. So that's, you know, we are really waiting for that moment and then just launching, you know, commercial operations without a human and then scaling that product very very rapidly. So it's a really exciting time for us. And, you know, the announcement that you're referring to was that, you know, in self-driving particularly in trucking, is being tremendously difficult to expand geographically. The ability to, you know, for trucks to drive and typically to open a new freight lane, it takes, you know, many many months, typically more years than months. And what we saw last week is that well it turns out that for Waabi because of the amazing abilities of our AI driver, we can actually open lanes instantaneously without requiring any retraining, without requiring engineering, without even requiring simulation. We built really a physical AI platform that has, you know, it's the first instantiation of general intelligence where it can actually drive by learning how to drive, not by memorizing a route, which is typically what you see with the rest of the industry. And that's a massive milestone when you think about we are not here to bring trucks for a, you know, I guess a small geography. Our customers have nationwide as well as, you know, global networks and they want to buy our trucks and deploy immediately everywhere, right? So that's kind of the technology, the scale that you need in order to really commercialize this product. And Waabi is absolutely leading the industry. It lives in a very different dimension in terms of the speed and scalability of the approach.
So maybe one thing I think our listeners would love to hear you maybe go a little bit deeper in what is physical AI? We obviously hear a lot about AGI. We hear about the doomers around agents and bots. Could you tell us a bit more about what physical AI really means?
Yeah, over the last maybe four years, you know, I think the world has had a lot of conversations about AI. But when you think about what are they talking about, they're talking about the digital version of what AI can do. That has really, with, you know, large language models, it has really transformed the way that we interact with computers, but everything is in the digital world. Now think about bringing this technology to the physical world, right? The way that today, you know, we do tasks, you know, a human doing them. Now we're going to have robots that can automate, you know, many of those tasks so that as humans can actually do much better jobs, right? That's kind of where this technology is going. And it's really about, you know, AI that has a body, whether it's a vehicle, whether it's a drone, whether it's a humanoid or any other embodiment out there, is really about how do we build intelligent AI systems that can help humans in our day-to-day tasks. Now, self-driving is the first application where scale, you know, we are really at the ripe of that, you know, commercialization at scale. So that's the first where really, you know, will be deployed really everywhere. But other instantiations are also going to come, you know, over the next 5 to 10 years you're going to see humanoids and other things really popping up in your houses, in hospitals, right, helping take care of the elderly in nursing homes etc. Places where, you know, we require an alternative to humans because there is not enough humans to actually do those jobs, right? And when you think about that, 85% of the GDP is actually tied to physical assets, right? So think about the disruption that this technology is going to provide and the, you know, the amazing benefits, right? So if you thought that the digital world with LLMs was big, physical AI is a different dimension as well.
