Transcript
Speaker: Hello and welcome to the Macro Brief from HSBC Global Investment Research, the podcast that looks at the big topics driving financial markets across the world. I'm your host, Piers Butler in London, and on this week's podcast, we're catching up on the latest developments in artificial intelligence.
Speaker: A sell-off of chip stocks saw billions of dollars wiped off stock markets this week. And with the so-called Magnificent Seven stocks in the thick of their reporting season, we're asking how investors and companies are feeling about AI given competing narratives and market jitters.
Speaker: To discuss this and more, I'm joined in the studio by Mark McDonald, head of ai and data science. Mark, great to have you back on The Macrobrief. Thanks, Piers. It's great to be back. So I was looking it up. ah We haven't spoken since December of last year. now so in the world of AI, quite a lot has gone on. ah And in fact, you've just co-authored a yeah report with Alice Sapinda, who's our head of global and EM equity strategy on AI narrative shifts.
Speaker: I thought it'd great to talk about these narrative shifts and you identify five of them. The first one is AI euphoria, which I guess encapsulates this idea that AI changes everything and we are at the foothills of a singularity. Are we at the foothills of a singularity?
Speaker: I think this is one of the things that at the moment, it it definitely doesn't feel like markets are trading like that. I think when when people ah envisage what they mean by a singularity, sometimes people let their imaginations run away with themselves. The market sometimes fully buys into that narrative um and at the moment seems quite sceptical of it.
Speaker: um So the the piece which we're referring to here like kind of tracks the prevalence of these different narratives through time. That one right now is not particularly dominant as you as you might expect.
Speaker: The next one is AI disruption. There's been a lot of discussions about the impact of AI on jobs. It feels like that narrative has evolved quite a lot, even since we last spoke. Yes, it has. I think, you know, with with any new technology, there's a tendency for people to ah focus on what might be taken away from them.
Speaker: And so it's very easy to see, as AI progressively does more and more tasks well, um that people begin to worry that it will begin to take over more and more jobs.
Speaker: Often this tends to underestimate the degree to which people are flexible and to which the global economy is is actually a complex system that will will react. And historical evidence suggests that new technologies generally, they're good for productivity, they're good for growth, and they generally generate more new jobs associated with the technology um than get get replaced by the technology. So we generally think that this um this fear of large-scale AI-driven unemployment is is an overblown fear. And often I think know what will happen is you know people's roles will change, but also who's going to get the productivity benefit. Often it ends up being the consumer that gets the productivity benefit through a sort of higher quality product.
Speaker: and ai kind of just raises the floor for what's acceptable and so it it might not end up replacing as many jobs as as people expect but there is this paradox isn't there that like that there was this fear when excel spreadsheets were introduced that accountants would be put out of jobs in fact it allowed them to do a lot more well exactly yes it's uh i'm sure when excel and similar products were released that people were saying it was terrible news for accountants and financial analysts But actually, there's been an explosion in jobs related to that type of calculation, because when you make something cheaper and easier to ah to produce, you tend to get a lot more of it. And this is the classic Jevons paradox.
Speaker: I think also as well, like if you look at the life of, say, people in this department as research analysts, you've got sort of three aspects to their life. Most important, which is like seeing clients. um You've got the writing the research report and then you've got the coming up with research ideas. Analysts are already using AI tools to help them do that research process. But what you're not seeing is people doing the same amount of research as before, ah just more quickly.
Speaker: People are doing much more thorough research because it's a competitive market. The critical resource of you know profitable investment ideas hasn't increased because of AI. And you've got everybody competing for that same resource with better technology. So people are still pouring the same amount of hours and effort into doing so. um It's just that it allows them to research more thoroughly unless the consumer of the research gets the benefit because they get higher quality research.
Speaker: The third narrative is hyperscaler overspend, which is clearly front and centre at the moment. So this is looking at the fact that CapEx investment numbers have grown exponentially.
Speaker: And the the core question is, will all this computing actually be required? Yeah, I think sometimes it feels like these companies can't win because when they weren't spending all this money on capex, people were complaining that they were wasting all this money and had all this cash and they weren't doing anything useful with it. And as soon as they do start investing for the future, then the people worry that they're over investing and overspending on capex. and And we go through periods like now where the market punishes them for doing so.
Speaker: Whether it turns out to be sensible investment or not, I guess time will tell. At the moment, it seems like there is vast potential demand ah for AI products and services.
Speaker: um And as these tools keep getting better and better, I think companies are only just beginning to find out ways to you know practically embed them into their operations in ways that really move the needle.
Speaker: um And as soon as companies start seeing the benefit from that, then there's going to be significant demand um from the business community for for spending on AI. So really very hard to tell at this stage. I guess the numbers are so huge and some of the companies are now burning through their free cash flow and actually borrowing. So the market is kind of saying, is this all really going to be justified?
Speaker: That is one of the narratives that's trending at the moment. um I think you know despite the AI trade having been obviously fantastically successful for several years now, often the market narrative is looking for a reason to be negative on AI.
Speaker: And that direction, the directionality of that commentary is always the same. People are looking for reasons to be negative, but the rationale for changes. And so you see these wild shifts in the the driving narrative as to why the AI theme is going to fail. It's just that it it flips rapidly from from rationale to rationale.
Speaker: And often these are quite short-lived little tempests and then then the market narrative changes. One negative, so to speak, is the the fourth narrative, which is China competition concerns. That deep seek moment was not a one off. And I guess it's kind of encapsulated in saying these U.S. companies are spending billions of dollars and actually China can do it much more cheaply.
Speaker: Yeah, it feels that the the US and the Chinese AI labs are competing on entirely different axes. So for the for the US AI labs, um when they release a new model or a new tool, and they need to be able to claim that it is the best, at least on some metrics. um the The Chinese AI companies have been competing more on efficiency. So they're not trying to come out and say that this new model is the best at this whole range of tasks, but they do want it to be good enough and dramatically cheaper.
Speaker: And it's going to be interesting to see how this plays out. um Often, particularly in a commodity business, being good enough but dramatically cheaper is a winning strategy.
Speaker: um I think it really comes down to how how much demand there is for increasing excellence in the in the model performance. I can imagine a bimodal world where you get some a small number of winners of the the enterprise AI race. So whoever can win that race to become the ah the default choice of most enterprises when they're looking for an agentic AI software tool.
Speaker: So whoever can win that race is going to be very sticky business. And that should be very easy to monetize and thus... you know that you know really does warrant spending lots and lots of money on capex today in order to try and achieve that prize.
Speaker: For the AI use cases where companies or people are accessing the AI tools via API calls, that is very commoditized because you've got lines of code that just periodically call an API, switching that from an anthropic model to an open AI model to a Google model or to a model ah from one of these Chinese companies that's been much more focused on efficiency. um That's very easy.
Speaker: And so I suspect that that part of the of the AI monetization race, that will be a commodity grind to quite low profitability. And so you're going to get this bimodal bimodal distribution, I think.
Speaker: Super interesting. The final narrative, which is which we're kind of living through at the moment, is ah AI positioning capitulation. Is that an unwinding of a momentum trade in in in market terms? Yeah.
Speaker: I suspect it's mostly that, yes. it's um Rather than something fundamental having changed, it really does feel like, as I mentioned before, that there's this long-running trend for narrative and commentary to be looking for negativity with the AI theme ah in the face of continued profitability and like a strong performance in the trade itself.
Speaker: and This feels like one of these short term periods of negativity, mostly like a positioning unwind rather than a fundamental change in what's happening with AI.
Speaker: So without getting too technical for our listeners, the report ascribes probabilities to these scenarios and what you call return signatures. How how do you go about doing that? So it's actually very similar to our other regime models like Eccles and Cockles and Tickles that ah look for look for regimes. How you come up with those names, I don't know. Yeah, I mean, they obviously coming up with the names is the most important part, so you've got to have good branding. But so these look for regimes in equity, commodity and rates markets, respectively.
Speaker: um And what we what we do here is first you you feed in data um to a clustering model, um And this data is generally based on returns and volatility and correlations from assets in the asset class. And the first thing is the clustering model tries to group periods of times together.
Speaker: And since these periods of time are much more similar to each other. So looking for patterns, I guess. Exactly. So we're not telling it anything at all. We're just giving it the data and saying, you tell us what the natural groups are in this data.
Speaker: um And then from those groups that the clustering model produces, um we then train um a classification model that given data today can allocate it to one of the same clusters.
Speaker: um And how you name the clusters and identify what's really going on is partly by looking at the returns of assets in each of these clusters, so these return signatures, and Partly looking at the drivers from this classification model. So the classification model, we can use modern explainability techniques like SHAP, which are kind of a cornerstone of of many of our data science pieces of research.
Speaker: And this allows you to understand why a model is doing what it's doing. And so you can say for this cluster, the sorts of characteristics that make the model confident that it's in this regime are X, Y, and And from that, you can then get a good understanding of what really is the natural hallmark of that particular cluster. And that allows to make these regimes. So a cluster would be, for example, in terms of what we've been discussing, hyperscaler overspend. Exactly. Exactly. What's striking in looking at the report is how frequently these regimes change. Is that likely to continue will it settle over time?
Speaker: For this, I think it's likely to continue. um in the tea the The AI theme has been such a successful trade ever since ChatGPT came out.
Speaker: that you know a lot of people are worried about it. It's new. The space, the the drivers in the space are changing all the time. um It feels like this volatility in narrative and explanation is just an inherent part of of this of this space.
Speaker: um you know when When it comes to our other regime models, you know the like equity regimes or commodity regimes, those are much stickier. Here, it really does feel like you know the the market narrative as to what's driving things in in AI can change at the drop of a hat. Because it's still so new, I guess.
Speaker: It's so new and it's so scary and so much of um you know the valuation of many of the companies involved in it are based in long-term expectations. And with long-term expectations in a rapidly changing area, you don't need to change your expectations that much in order to have a significant move.
Speaker: Closer to home, i wanted to ask you the question which I asked you last time is how is it impacting your work and that of other analysts? I saw that you had published ah recently a report looking at whether AI can update equity models, which to me seemed very far fetched for the clues in the name, large language models. How does that work?
Speaker: um Large language models for some time now have been very good at coding and they find coding easy and they find spreadsheets hard, which is the opposite of most humans. Most humans find spreadsheets dead easy and coding quite difficult. And so, I mean, obviously this experiment that we were doing, we were ah looking specifically at how well or otherwise and different AI tools can update the big complicated spreadsheets that single stock analysts use to model their companies. But I think the broader, bigger picture question is, to what degree can AI just safely and reliably um do spreadsheet-related tasks?
Speaker: Surprisingly mixed results, to be honest. um Like even, although there was a clear ranking amongst the different AI tools that we looked at yeah with the clawed suite of of AI tools performing best, ah there were still many surprisingly basic errors that it would make in this process of updating these models with new earnings release data.
Speaker: And the weird thing about this is that I think any individual cell that you are asking the tools to update in the spreadsheet, all of these tools will be able to do that and have been able to do that for some time now. The difficulty comes from, I think, two things. One, you're asking it to update all of these numbers at the same time.
Speaker: So there's a lot going on. um And the other thing that's difficult is you're asking it to understand um the structure of a complicated spreadsheet whilst doing this. So the sorts of mistakes that the tools would make is that they had a tendency to go in and update spreadsheets in a way that would break the structural integrity.
Speaker: So it would go over and you've like a column of cells where it's all the same formula that's been dragged down and it would just like randomly overwrite some of them with hard-coded values and not others. So kind of very dangerous errors that are hard for human to a human to spot.
Speaker: This feels like the sort of thing that the AI companies are going have no problem in really fixing. I suspect that part of what's going on is that, you know, all the people who are building these tools, AI companies, they're geeks like me. And so they're kind of just assuming, like, why is everybody using spreadsheets? Why can't they just get with the program and write it all in code? um And so I think it's been a bit of an oversight. And now these companies are really starting to realize the like vast amounts of, you know, business activity that is controlled by big complicated spreadsheets in enterprises all across the globe. um And I think now that they're focused on it, we're going to see rapidly improving performance there.
Speaker: More changes to come. yeah And indeed, there's another change that is about to occur, which is that you're relocating to Hong Kong. I am indeed. How excited are you about this change? and And what's your thoughts about how Asia is positioned in the AI revolution? I'm super excited.
Speaker: The interesting thing about Asia is I think ah over here in the West, we're so much more focused on what's happening with the AI labs in the US. yeah But as we discussed earlier with one of these ah one of these themes itself, is there is a very thriving AI scene in China as well.
Speaker: um And I think the when I was visiting Hong Kong for the for the yeah GIS earlier this year, um A lot of the conversations I had with with Asian investors, they were much more focused on you know this competition between what's happening with US AI labs and Chinese AI labs.
Speaker: um So i'm I'm very much looking forward to getting more of experience and exposure to that world. Well, I think that's all we've got time for today, but the best of luck with your relocation. Thank you. I hope all your boxes arrive in the right place. Fingers crossed. But thank you for joining us today. Thank you very much.
Speaker: Mark McDonald there on the latest developments in the AI space. And if you'd like to hear more on China's emergence as a force in chip making, then check out this week's edition of our sister podcast, Under the Banyan Tree.
Speaker: But that's all from us here on The Macrobrief. This episode was hosted by me, Pierce Butler, and produced by Tom Barton. Don't forget to like and follow wherever you get your podcasts. Thanks for listening, and we'll be back again next week.
