Catch the full episode: https://www.wealthformula.com/podcast/349-the-next-big-technology/
Buck: Welcome back to your show, everyone. Today, my guest on Wealth Formula parties is Avi Goldfarb. He’s the Rothman chair in Artificial Intelligence and Health Care and a professor of marketing at the Rotman School of Management, University of Toronto. He’s also chief data scientist at Creative Destruction Lab and Cedar Rapids Screening Consortium, which is a faculty affiliate at the Vector Institute and the Schwartz Reisman Institute for Technology Society. And he’s also research associate, National Bureau of Economic Research. Beyond that, AbbVie is also the author of multiple books, including the most recent one Power and Prediction The Disruptive Economics of Artificial Intelligence, AbbVie. Welcome to you all from your podcast.
Avi: Okay, Thank you. Great to be here.
Buck: You have a thing or two thing or two that tells us you know, about a thing or two, I guess, as they say. Let me start with this question because I think for, you know, people who are busy professionals, middle aged technologies just seems to be like hitting hit them over the head. Right. Constantly. And we keep hearing about new things. And I would put A.I. in this category where it’s something that it’s a buzzword. We hear it all the time, but I’m not sure that anybody that most people really understand what it is. Can you start by explaining in your own words what artificial intelligence actually is?
Avi: Absolutely. So when you hear people talk about artificial intelligence today, you know, when the hype in the press makes you think these are the machines that we see in science fiction, machines that can do just about everything we humans can do. Maybe they listen to us or maybe they don’t. And that’s what we get. Scary science fiction. That kind of artificial general intelligence is possible, but it has almost nothing to do with what we’re talking about. We’re talking about A.I. in business today. We’re talking with air today. We’re talking about a very particular branch of computer science. In fact, a branch of computational statistics called machine learning that’s gotten better. And so when you hear a, I think prediction technology and statistical sense taking information, you have to fill in information you don’t have. It’s a little less exciting to talk about computational stats instead of AI, but it can still be transformative.
Buck: Right. Absolutely. So can you give us some examples of how A.I. is currently used so that we can kind of put that definition into context?
Avi: Sure. There’s lots of places where, you know, prediction effects are your daily lives, so you want to get from point A to point B and the old days you would open up a map, would’ve been paper, and you would’ve figured out how to get how to get from point A to point B, Now you’ll open up an app, Google Maps or something else, and it will provide you with a prediction of the fastest route to get from point A to point B, That prediction will be largely informed by artificial intelligence, by a sense of where the traffic is going to be and how long it’s going to take.
Another example is the recommendation engines that you see from Spotify or from Amazon, where they recommend to you what they think you’re going to buy next time. So Amazon, think about Amazon as hundreds of millions of different products in the catalog. And for the recommendations, they tend to be right about 5% of the time, maybe more. That’s an incredibly great prediction about what you’re likely to want at that moment.
Buck: You know, that that’s I mean, it’s fascinating stuff. And I’m you know, I’m kind of going off a little bit of a non-economic tangent here, but I’m curious on I think I think I read in your bio somewhere that you were involved with some health care. But when you look at I as a physician, that’s very interesting to me, right? Like because the applications specifically when it comes to, let’s say, pharmaceuticals, vaccinations, that kind of thing, is that, you know, how far along are we with that? And maybe describe like how those things work a little bit?
Avi: Sure. So the biggest opportunity for Angel is around diagnosis. So what does your doctor do when they diagnose you? They take in data, information about your symptoms and they fill in the missing information to the cause of those symptoms. That’s prediction. And more recently, realize realizing the machine. But I could do that as well, or better than many human doctors, if not most, in terms of the reality on the ground. Those diagnosis predictions, the diagnosis is had very, very little impact. Where I had an impact in health care are on the research side. So if you’re trying to figure out what drug to produce, if you’re a pharmaceutical company, you might be using AI to help you search the literature you might be using AI to help you predict which which molecules are going to bind with which proteins to identify potential drug candidates you might be using instead of good old fashioned stats to do your stats in the background, you’ll be using A.I..
So there’s on the research side. And then a little bit, we’ve seen it on the operations side. So the for example, the hype in radiology has been about while radiologists look at images. Right. And really get a look at images. And so you’d think that I would start replacing radiologists hasn’t really happened yet, but AI has transformed workflow radiology.
So radiologists ten years ago we’re talking to a microphone as part of the workflow, and then that recording would be sent to some humans who would listen to it, write it down, transcribe it, and then send it back to the radiologist within 24 hours. Very few radiologists do that now. Now we have AIS doing the transcription for them. So it’s not it’s not what you think about in terms of the AI in medicine, but it actually has changed the workflow in a meaningful way. Yeah.
Buck: I mean, I have an interest in longevity science and it seems like there’s, you know, a lot of talk about using AI, you know, for measuring biological age and that kind of thing. Is that I don’t know how much you know about that, but I’m curious how far along that kind of stuff is.
Avi: It all remains in the lab, so it’s there’s lots of smart people thinking about it, like lots of other AI in health care. But my knowledge, clinical application has been pretty limited.
Buck: Let’s shift to the economy a little bit. I know your book is about the disruptive economics of AI and what makes it so disruptive.
Avi: That it can allow you to do things differently and you can serve your customers in a way that they’ve never been able to serve in that way before. And so once that can happen, you can transform industries. So I talked about the maps, the predictions about how to get from point A to point B, the original applications of that in business or what we call point solutions, which is that, you know, we look at professional drivers, we’re already professional drivers, truck drivers and especially cab drivers. And so they can do their jobs a little bit better if they have these real time predictions about where traffic is going. And that happened. And some of those had an impact on the taxi industry and other aspects of the professional driving industry. But then a handful of people realized, well, once you have good predictions about how to get from point A to point B, the expertise that many professional drivers had about how to get around a city or country was no longer an advantage.
People in the city of London used to be a three-year process of learning to become a cab driver to learn the streets of the city. And what what navigational AIS, what these predictive maps do is they mean that just about anybody can be the quality of a professional driver. And that led to the rise of Uber and Lyft and a totally new model of transportation that was incredibly disruptive to the old industries with professional drivers.
Buck: I’m curious about some maybe some other applications, you know, how about in your iPhone and digital? I can almost see like a hedge fund looking at this, you know, and looking at figuring out how to make money. Is that is that something that’s also a thing right now or could be or.
Avi: So there’s yeah, for sure. So there’s a couple in professional investing. There’s two parts to it. There’s one which is can you identify an angle that you can use prediction technology in a way that nobody else can. So if you’re you’re the first people who realized they could take satellite data and use it to predict how many cars were in the parking lot of the biggest retailers in the country, to get a sense of what sales are going to be before anybody else did, they had a leg up and they made money. The challenge in that kind of use of AI for professional investing is that once two deep pocketed people can figure out the same and they’ll sort of bid each other out. And so you have to be the first to do something and do it in a unique way.
Buck: And then the retail investors, one who loses.
Avi: And then the retail investors really want to was Absolutely. Yeah. The other way to think about it, though, is to think through which industries is going to affect and when and how do we think through the big risks, the big opportunities and the little wins.
Buck: To talk about that a little bit, you know, because which as you mentioned, truck driving, obviously, you know that that that certainly having better routes and that kind of thing. But some of the other application is obviously a it’s going to be transformative for some businesses, some maybe to just make them a little bit more efficient. But what what are the big opportunities that are out there right now?
Avi: So the big opportunities are thinking about industries where they’re where much of the structure of the industry is about failing to serve their customers well. And I’ll give you an example of an industry where they where much of what they do is about failing to serve our customers well. And then you can sort of speculate on what others might be.
So if you go to the airports, rated the best airports in the world Seoul, Incheon, Singapore, etc., they’re pretty spectacular. And the architects spent, you know, to spend billions and billions of dollars, sometimes private money, sometimes taxpayer money on these places that they say are great places to hang out. So Seoul, for example, it has a theater, it has great restaurants, great shopping, seems like a spectacular place. But then if you look at how the super rich fly, they fly through Schatz, the super rich, the private jet terminals are empty. There’s no restaurants, there’s no shopping, there’s no massage parlor, massage or theater. Why? Because no one wants to spend time at the airport. That architecture all exists. Really? Because those airports are failing in their mission. Their mission is to ensure smooth air transportation. When we see what the super rich do, we can see well, that’s what it means to ensure smooth air transportation you don’t spend any time with there. Yeah.
Buck: No layovers and no private jet.
Avi: Yeah. And no, you know, you know how long it’s going to take to get to the airport through security and fly where you want to go. Lots of industries have all this architecture, physical or virtual. A lot of what people do aren’t about delivering good service to your customer. They’re about the places where you fail to deliver good service and you try to make a point. And so the investment strategy or a more general strategy to think through who’s at risk of disruption is to identify those industries where a lot of what you do is based on not actually serving your customers, but making up for the fact that you don’t serve your customers well. And I would argue health care is a big part of that. I would argue insurance is a big part of that to some degree. Financial services, at least retail banking and some others.
Buck: To what degree is the the fear that some people have about AIG taking people jobs and, you know, rendering employees in certain fields useless? How serious should we be? Take that. Those kinds of concerns?
Avi: They are serious concerns, but there’s also reasons for optimism. And what I mean, so let’s talk about the taxi driver example. And I think it’s a much broader point, which is when Uber came along with their navigational aid, those people who had spent three years in school learning their way around the city of London, their income suddenly went down.
So I had the effect of taking some people who invested in skills and making those skills, those skills more irrelevant, but instead anybody could have those skills. So so the people who got some advantage relative to higher wages and they might otherwise have suddenly were competing with millions of other people. So the negative version of it, which you should think about, is those taxi drivers no longer had the you can they had before the positive version of it. Is it upskilled millions of people to be able to drive professionally? And I think on almost all of these disruptive technologies, we’re going to see both sides. We’re going to see there’s a handful of people who made their living doing that because they were better at it than anybody else. And there’s going to be lots of people who are upskilled.
So I don’t know if you saw it on November 30th opening, I released this new tool called Chat CBT and Chat CBT. Is it the halo? It’s the chat bot, but it’s an if you ask it to write a five paragraph essay comparing, I don’t know, Shakespeare to Michael Porter. It will. What is it called? Chat GPT.
Buck: Okay, that’s great. Maybe you can do some of my podcasts.
Avi: It is incredible. And writing lucid text. Really good. To the point where us professors, we have a real worry.
Avi: Got. We can’t do take on exams anymore. They’re just gone.
Buck: There is there’s actually I think just just is a slight diversion here. I, I had heard about some applications that we’re using using A.I. for marketing. They’re really tapping into, like, dopaminergic pathways. And what gets you going. Is that right? Maybe it’s just a marketing thing about marketing.
Avi: Whether it’s really A.I. or just, you know, generally. So I would argue there was a clever marketer who realized that, you know, if we measure very carefully what people click on and how big, you know, what makes people engage with the platform, then we can figure out which kinds of activities lead to more engagement, whether it’s, you know, whether it’s something deeply psychological or more superficial, I don’t know. But we can teach the AI or at least tell the A.I. what to optimize in a way that helps the company and may or may not help the customer. I guess is that how do you.
Buck: Think what are some of the things in daily life that you see? Like, you know, when people talk about A.I., they really. And maybe you can comment on this too, but people talk about it as if it it really is is transformative is the Internet for the world we’re going to be living in. First of all, is that true? And if so, on a day to day basis, what are we going? What kinds of things are we going to see?
Avi: Okay, So I believe it is true that AI is transformative, like the big technology, the past computing, the Internet, the steam engine, etc.. Okay. I a suite of technologies are in data science with the center that said that history tells us that it’s going to take more time than we think.
Buck: Right? We’re like intercom era right now.
Avi: It’s not even not even. Not even. So, with electricity, the light bulb was 1880, so it was clear in 1880 that electricity was going to change the way we lived and work. But it wasn’t until the 1920s that half of households and half of factories were electrified in the US. Sure. With computing, it was clear in the 1950s that this was going to be a transformative technology. And it wasn’t until the 1990s that showed up in the data. These changes take a lot of time, and so I expect the impact of AI to be extraordinary. But that doesn’t mean that we’re going to feel it in most industries tomorrow.
Buck: Yeah. Although probably faster than some of these other things by the very nature of the computers that are involved and how they get faster and chips get smaller and all that kind of stuff.
Yeah there’s reasons to be optimistic. It’ll happen faster. The only reason to be pessimistic that’ll happen faster is that you to your point about jobs, there as well, and technology, there can be people who benefit and people who get hurt. And in the context of A.I., I think a lot of the people who get hurt are the people who are currently in power. And so think about medicine. If A is doing diagnosis, maybe the doctors don’t like it so much.
Buck: The doctors get to proofread it, and so they’ll do. On radiology, they’re just replacing the residents.
Avi: So. Well, then maybe it’s fine or then maybe it’ll happen. Yeah, but but you know, there’s reasons to worry in a lot of industries that the people who currently benefit from the system as it stands but want a new system.
Buck: Right. No. You know, the last thing I sort of finish up with and maybe you can address this is, you know, there’s been some real fear mongering among some pretty big names in technology. You know, Elon Musk, Gates, Bezos, they’re all warning about, you know, what I could do and that, you know, the coming of the AI apocalypse, uh, what tell tell us tell us what your take on this is.
Avi: So that is not the technology we’re talking about today. So that’s an artificial general intelligence that has been 20 to 50 years away since the 1960s and 20 to 50 years away from terrifying since 1960s. And if you talk to experts today, it’s still 20 to 50 years away. That doesn’t mean never. And I’m very happy they’re smart political scientists and philosophers thinking about it and trying to understand what those consequences are. But in terms of the investment in business applications today, that’s not what we’re doing. Prediction technology is a long way away from a machine that’s about to take off.
Buck: Like singularity or something like that, right?
Avi: We’re not there yet.
Yeah, yeah. Fantastic. So the book again is Power and Prediction The Disruptive Economics of Artificial Intelligence. What, if anything, have we, you know, major points in this book? Have we not kind of addressed.
Avi: So there’s one thing we haven’t talked about. So we’ve talked about the big system change. We’ve talked about resistance to that change. The one piece we haven’t talked about is that there is no such thing as a machine decision. Okay? So as long as we’re in a world of prediction machines, as long as we’re in a world of where A.I. is prediction technology, the A.I. provides the prediction, but a human ultimately makes the decision. And so when you read Jobs at Risk by A.I. decision making or in I decided to fire that person or any I decided, you know, in military context to fire that that’s not what’s happening. And I think this distinction is important. What’s happening is the human who used to make those decisions on the ground at the moment of the action is no longer making the decision. And instead someone else typically more centralized, say, at headquarters, is making decisions for everybody else. So, you know, in terms of an AI, that’s for hiring. So instead of having individual HR Managers deciding who to hire, instead we are having some of the headquarters decided here’s what success looks like in our organization. We’re going to predict that and decide if you get over 80% chance of success, we’ll hire you. So this the idea that machines don’t decide is fundamental to allocating responsibility and identifying opportunity.
Buck: And very interesting stuff. Again, the book Power and Prediction, The Disruptive Economics of Artificial Intelligence are the Goldfarb RBA assumed the book is available anywhere Amazon, the usual routes.
Avi: Amazon and the Nobel where wherever you want to get it. Airport bookstores. It’s all around.
Buck: Wonderful. Well, great having you on the show and and look forward to reading the book.
Avi: Okay. Thank you very much. Take care.
Buck: We’ll be right back.