Mohamed El-Erian on how AI is changing investing

View Full Transcript

Joe Cass   00:00:00

Hello, and welcome. My name is Joe Cass, Senior Director at S&P Global Ratings and the host and the creator of the FI15 podcast. On this episode, we have Mohamed El Erian, President of Queens College Cambridge, and Adviser to Allianz and Gramercy; and Sudeep Kesh, Chief Innovation Officer at S&P Global Ratings. So, a quick reminder that the views of the external guests are their views alone, and they do not represent the views of S&P Global Ratings. Mohamed, welcome back. You wrote a piece in the FT about six months ago now, and it was called 'how Gen AI will change asset management'. What are your current thoughts on this really broad-ranging topic?

Mohamed El-Erain   00:00:38

So I'm a believer that AI is a transformative innovation that will impact a lot of what we do. I think asset management is among the sectors that will be significantly influenced by AI. There is low-hanging fruit that is already being taken advantage of, but it is the higher up fruit that is really exciting. And that is to go from reporting to actually enhance analysts and enhance portfolio managers in achieving their missions.

Joe Cass   00:01:18

Perfect. Thanks, Mohamed. Sandeep, your research lab at S&P is primarily concerned with things like AI adoption, governance and risk in a whole heap of industries really. What are some standout examples that you're seeing in your own research?

Sudeep Kesh   00:01:35

Yes. I think it's a mix of actually exactly what Mohamed described. I think about two-thirds of companies, publicly traded companies now tend to be just kind of early in their adoption. So, they're really thinking about how do I use AI to teach me about AI. And what I mean by that is a lot of industries are basically looking like in banking or asset management, things like fraud detection and any kinds of aberrations for risk management and things like that, where we've actually had quite a history of using machine learning in these spaces. So now it's basically taking existing processes and looking to AI to lend some level of automation. Now I think similar to what Mohamad articulated is I think there's great potential if you really think about what your goal is and focus on that goal and see AI is basically kind of a transformative agent to mobilize your assets into something as of value or something of value to your stakeholders. So one of the standout examples is really actually at Louis Vuitton Moet Hennessey, one of the things I was reading that they're doing is they're actually looking at leveraging AI, not necessarily Gen AI, but AI more broadly on mimicking human old factory senses. So basically, the human senses smell and then looking early on in the supply chain, if one of these ingredients for a perfume is about $100,000 for a drop. If you're able to detect aberrations in your production early on, you could actually save millions and millions in products, not waste resources and really be able to get at this level of customizing products for distinct customers. Similarly, there are companies that are basically going -- and that's what I call back to front, by the way. So there's also companies looking front to back, and that becomes an intelligence gathering phase where you're basically able to say, okay, if I was to have an artificial beauty consultant, for example, I'm clearly not their target market. But if you were to create this kind of thing, you would start to get lots and lots of informatics around customer preferences, different sort of skin tone pallets, all of these other kinds of things. And that becomes another thing that unlocks the value of the existing assets. And I think that's the real benefit. So I think a lot of companies aren't there yet. They're really early in the journey, and that's okay. But I think where you're really going to see the rubber hit the road is those more novel use cases of doing things that haven't been able to be done before as opposed to just doing them faster, doing them.

Mohamed El-Erain   00:04:05

Sandeep, I'm really interested in what you just said because I get a sense that the majority of companies understand the what and the why of AI. but the majority also struggles with the how. And there is a whole range of views from let's educate ourselves, let's have an add-on to let's bring in someone who will disrupt us from inside to let's rethink the whole company as if it was AI native. What would you tell companies in terms of the how?

Sudeep Kesh   00:04:40

Yes, how -- so the way I think about it, and I should warn that I'm prohibited from giving advice. So, this doesn't constitute advice. But the way I think about it is that every business has assets. So, they could be the employees and the talent networks and things that they have access to. It could be data, it could be processes to unlock these things. Every company has customers, right? And then the secret sauce, their business model is actually a scientific hypothesis. -- on how they can unlock value from those assets to those customers. And the reason I say scientific hypothesis is because it's testable, right? So my advice, not advice would really think about the scientific method to say, okay, how do I treat all of these different processes as essentially science experiments and be able to kind of have this counterfactual and kinds of things. And then I think that sort of unlocks, a, what you're thinking and then how to think about basically meta cognitively, how to think about what you're not thinking about because I think that's the part of the how that's really difficult with AI is it's very cold and calculated. It's essentially sort of a pair of a Bayesian tree with a bunch of sorts of decision science applied on everything you could sort of think of, but it doesn't have feelings, it doesn't have emotions. Humans, I think, sometimes take for granted how intuitive decision-making can be when you're basically -- you're trusting your gut, these kinds of things. You can't -- unless -- I mean, we can sort of go down a rabbit hole on that with this branch of AI called causal AI that gets into that. But Gen AI and things like that doesn't really have any sense of causality, doesn't have any feeling. So that's why we kind of have to take this very sort of cold-hearted perspective and then I think you can unlock the how that way. That was a long-winded answer to your question.

Joe Cass   00:06:32

Great. Thank you. That's great. Mohamed, I've heard generative AI described as a combination of technologies and capabilities that help drive innovation in business. Do you mind sharing a few examples of what types of innovation Gen AI could help drive in the asset management space?

Mohamed El-Erain   00:06:52

So, we've already seen it impacting attribution analysis. So, your clients would like to know where did the returns come from, be they overperformance or underperformance. That is actually quite a labor-intensive exercise, and there's lots of shortcuts that are done. I've seen AI enhance this attribution analysis. They've made reporting a lot easier. For portfolio managers who have to go through a vast number of documents looking for particular things. So think of a basket of mortgages or think of when you're trying to assess how correlated are the risks in there, AI can really enhance your ability to analyze lots and lots of data and focus on a few things. So we've already seen these application happen. Where I haven't seen yet AI utilized is in secular asset allocations and in doing a better job at optimizing that combination of returns, volatility and correlation. And that every long-term investor will tell you that they start with these very ambitious plans to have an optimization program. And by the time they force all the constraints, they end up with something really arbitrary. And the hope is that AI will help you deal with some of these situations. I'm pretty hopeful that this will be the case, but it's still very early on in the process.

Joe Cass   00:08:30

Great. Thanks, Mohamed. Sudeep, I know you think a lot about AI, not just what it means to S&P or a particular industry, but more broadly, what advice would you give people when thinking about AI or Gen AI to help them use these tools to improve their lives?

Sudeep Kesh   00:08:49

Yes. I think there's a lot of directions you can kind of take. I think it's one of those things that be who you are, right? I think that's the main thing. So, if you're afraid of these tools, that's okay. But I would start to learn about these things and basically kind of revisit that fear. I think fear is healthy. Fear keeps us alive, right? But it's one of these things that we need to be sort of educated about what are the material issues and things like that. But I think one of the bits in the debate in terms of sort of humanity versus humanity plus AI is actually kind of AI in the arts and things like that. And it's one of these things that Ironically, that's how I got my start in AI twenty-five years ago was in music, like using AI to kind of be able to bring in this physicality in terms of emulating different spaces and sound design and things like that. That's how I came into AI. I recently met a friend; his stage name is King Willonius. He's an AI music artist. And he basically writes these comedy songs essentially, using AI, he does a prompt engineering technique to start the track. And then he's basically interacting, he's playing AI like an instrument. And I think that is really hugely illustrative of human creativity of drawing inspiration from anything that is of earth whether that's a traditional music instrument or AI. So, I don't think that these things are substitutes for each other. I think it's one of those things that we just have to treat the world with all of these technologies in it and then operate within their boundaries. Now I think with the fear thing, going back to the fear thing, like these things can be dangerous, right, in the wrong hands and so on. And that's fine like to recognize that, that fear is there, but we still need to do something about it. And I think that the key is education and being able to really be disciplined about it. Mohamed, I don't know what your advice would be. I think that would be really valuable for the audience as well.

Mohamed El-Erain   00:11:00

So, I would just quote a couple of situations. One was when someone from industry came to speak to our students and made a wonderful presentation. And then, of course, the first question was, how can I enhance the probability of getting a good job? And that's what the student asked. And the answer was to learn to speak to AI, learn to interact with AI. And I thought that's really interesting. And I didn't think twice about it. And then a few weeks later, I was in a meeting where someone from OpenAI was presenting. And then someone who was around the table, said, "Look, I just asked ChatGPT. -- what questions should I ask you? And they gave me this banal question to ask you. And the person from OpenAI said, that's because he didn't provide the context. So put in these four things. And they put in -- I am in a meeting with so and so and so and so. I'm listening to presentations with so and so who's back on and so on and so put in four little prompts, if you like, and then ask the question. Got a completely different answer and much more -- so I think this notion of experimenting, like you say, not be afraid of it, learn to interact with it because the more you interact with it, the more you realize like any other tool, you have to understand where its competitive advantage is. I also think, as Sudeep said, it's very important to understand this is an eighty-twenty proposition. 80% is good, 20% is bad. I go back from the U.S. to Europe quite often. The U.S. embraces the 80%, loves the 80%, wants to run with the 80% and tends to ignore the 20%. Europe's obsessed by the 20%, and they tend to ignore the 80%. I think the truth is, like Sudeep said, you have to embrace the eighty-twenty, embrace the whole distribution and try to manage the whole distribution because that's what you're getting.

Joe Cass   00:13:00

Fantastic. Thank you both. Mohamed, you spoke about it there briefly in terms of the risk. I mean, there's a whole host of risks associated with Gen AI from hallucination to data privacy or other considerations. What kind of risks keep you up at night about Gen AI or AI in general?

Mohamed El-Erain   00:13:21

So, hallucination doesn't keep me up because I think that as AI evolves, that risk will come down. It will always be there. I think what keeps me up at night is exactly what Sudeep said, which is a very powerful tool in the wrong hands. That's what keeps me up at night. And I can tell you, it's complex and it's making people think. I'll give you another example from education. People discovered that when they interviewed online, it leveled the playing field, that socioeconomically, you were having a much more inclusive platform. However, now there are tools that literally you put under your camera that listens to the question that's being asked and prompts answers for you. And if you're on the other end of an interview, you don't see that. So, you've got to evolve how you do things, understanding that the technology in the wrong hands can be used in a way that is counterproductive to ultimately what you're trying to do.

Joe Cass   00:14:27

Perfect. Thanks. So now we'll move on to some kind of off-topic questions. So, one thing I did in the last episode, which is kind of interesting, I'm going to replicate it here. So, it's quick-fire questions. So, it's basically the first thing that comes into your head when I say these words. And it's all about your favorite. So, what is your favorite X favorite Y. So, Mohamed, I'll go with you first, and then Sudeep will do your round second. So, Mohamed, what's your favorite movie?

Mohamed El-Erain   00:14:57

My cousin Vinny.

Joe Cass   00:14:59

What's your favorite season?

Mohamed El-Erain   00:15:02

Spring.

Joe Cass   00:15:03

What's your favorite drink?

Mohamed El-Erain   00:15:06

Sparking water.

Joe Cass   00:15:09

And what's your favorite restaurant?

Mohamed El-Erain   00:15:12

Chinese restaurants.

Joe Cass   00:15:14

Favorite gadget

Mohamed El-Erain   00:15:15

My computer.

Joe Cass   00:15:19

Favorite time of the day.

Mohamed El-Erain   00:15:21

Early morning.

Joe Cass   00:15:23

And last one, what's the profession other than your own that you would like to attempt?

Mohamed El-Erain   00:15:30

Being a lawyer.

Joe Cass   00:15:33

Interesting. Interesting. Okay. Sudeep, you're up now. So Sudeep, what's your favorite movie?

Sudeep Kesh   00:15:38

The Big Lebowski

Joe Cass   00:15:41

what's your favorite season?

Sudeep Kesh   00:15:42

Autumn

Joe Cass   00:15:45

What's your favorite drink?

Sudeep Kesh   00:15:48

Coffee.

Joe Cass   00:15:50

What's your favorite restaurant?

Sudeep Kesh   00:15:51

Pizza place down that way 

Joe Cass   00:15:56

What's your favorite gadget...

Sudeep Kesh   00:16:00

It's a tough one. AeroPress, I'll stick with the coffee 

Joe Cass   00:16:04

What's your favorite time of day?

Sudeep Kesh   00:16:08

Late afternoon.

Joe Cass   00:16:10

And lastly, what profession other than your own, would you like to attempt?

Sudeep Kesh   00:16:15

Time machine mechanic.

Mohamed El-Erain   00:16:20

May I point out Joe, that Sudeep had an unfair advantage.

Sudeep Kesh   00:16:25

I did have an unfair advantage, yes.

Mohamed El-Erain   00:16:26

He had an unfair advantage. May I also point out that his answers are all consistent because obviously, his love affair with coffee means that he needs a lot of it in the morning, which raises the one question. How many cups of coffee do you have before noon?

Sudeep Kesh   00:16:40

Three, which is too much.

Mohamed El-Erain   00:16:46

That's all?

Sudeep Kesh   00:16:46

Well, yes, it could be worse, but it's three and then one in the afternoon.

Mohamed El-Erain   00:16:51

So, I'm at five before noon.

Sudeep Kesh   00:16:54

I'm not sure who has to adjust.

Joe Cass   00:16:58

Mohamed, I wanted to talk to you just briefly about your newest book, Permacrisis, which I was actually reading on holiday a few weeks ago. Can you talk about the book, but also tell us about how your co-author, Gordon Brown, initially reached out to you in California and how that evolved into a weekly call during the pandemic between you, Gordon, and the Nobel Prize winner Michael Spence.

Mohamed El-Erain   00:17:21

So, I think it was back in 2016 or 2017, I got an e-mail from Gordon Brown saying, I've read your article. I'm coming to California. I'd love to get together and discuss. So, I showed it to my wife. I said, okay, someone's pulling my leg here. Who do you think it is? And she said, well, just respond and take it seriously. So I respond to that, of course, I'd be happy to meet. And then we organized it that he would come to the house for brunch. It lasted six hours. He had come actually to discuss the article, which was a bit of a because I hadn't read what I had written weeks ago. But we had this incredible discussion. And then we followed up two weeks later with another long session. Gordon is this incredible sponge really interested in public policy, really cares about economic well-being, about inclusiveness, and we just connected. Come the pandemic, early on, he and I were sharing concerns about this notion that people thought that when you restart an economy, it's like flicking a switch. That's not what happens. We are worried about what would happen to vaccine distribution. So, we decided to have a regular call. And then he started asking questions that I could not answer. So, I said, do you mind if we bring a third person in, a friend of mine, who is the nicest person in the world, but he also is the smartest person I know. And that is Michael Spence, who I had co-authored a couple of articles with a few years earlier. And he said, yes. And then we evolved where you had Mike being the precise content person, Gordon being the visionary policy person and I being the supplier of the Zoom links every week. And we did this for a year where we were talking about our concerns. And then there was a really important pivot where we came across three factors that we believed explained most of this notion of cascading crises. At that point, the conversation became really positive, and someone said, we should write this down. And that was an oh, no moment because no one had taken any notes for a year because this was more like a social event than anything else. And then that was the idea for the book. And then we brought in Reid Lidow, who is incredible, and he helped 3 very different ways of writing seem like seamless and one person. So, it was a wonderful cooperative effort among the four of us. Great.

Joe Cass   00:20:11

Thanks, Mohamed. Sudeep, high-level question for you now. If you could give your teenage self-advice, what would it be? And, if he could see you now, what would he be most impressed by?

Sudeep Kesh   00:20:26

I think I'm still a teenager, so it's a relative thing. I would say to my teenage self, yes, just kind of respect the fact that things happen. You can control yourself, but you don't control the outcomes of situations. It's like you just got to let stuff be. And related, I think it's one of these things that I'm starting to learn that actually through my children. I'm starting to learn that now. So, I think it would be like my teenage self in the future to the present. You might be impressed that you eventually got to figure that out because it's just like we don't hold all the cards. We actually hold very few of sort of changing the outcomes of things like that. So, I think it's just one of those things, be yourself, just get up, do it, stay focused.

Joe Cass   00:21:22

Yes, very cool. Mohammad, in business in corporations, in investing, - it's often spoken about the importance kind of the criticality of data. What's your view of the role of data in decision-making in 2024?

Mohamed El-Erain   00:21:39

Critical, absolutely critical. But it should inform your judgment. It shouldn't completely take over your judgment. Data is a critical input, but it does tend to capture the past and there are all sorts of structural changes that go on. So, you should also have an open mindset to what is changing. The mistakes that are made in policy is by sometimes not supplementing backward-looking data with forward-looking hypothesis. And that's the mix that needs to be done. So, data is really, really critical. And I think evidence-based policymaking, evidence-based decision-making in business is critical. I joked about the lawyers because I was told a joke when I was young. My father was a lawyer. I would have been a lawyer if I didn't want to come across as simply following my father's footsteps, so I chose something completely different economics. And I always -- I was once told the big difference between a lawyer and an economist is a lawyer can argue with 100% conviction and 10% foundation because that's their job. an economist needs 90% foundation to argue with 90% conviction. And the reason why I said law because I often wondered whether lawyers can't do a better job in getting the balance right between conviction and foundation. And I think data helps you get some of the way on conviction, not all the way, but some of the way.

Joe Cass   00:23:24

Great. Thanks, Mohamed. Mohamed, it really wouldn't be an episode with you as a guest if we didn't talk a bit about football/soccer. So, a reminder to Sudeep and also the viewers and the listeners that in the 1979 to '80 season, Mohamed was a captain of the football first team at Queen's College of Cambridge. And I've got the picture as proof. So, Mohammad, now you're the President of the University. Do you get to watch much football or just generally sports on campus?

Mohamed El-Erain   00:23:59

Yes. So, my biggest dilemma every weekend is how to watch four different teams that are playing on Saturday and Sunday and are playing in different places often at the same time. So, we have the two male football soccer teams. We have the female soccer team, and we have the one male rugby team. And I try to go out and see them. And they're normally playing different places. But absolutely, I mean, I love going out, I love cheering. And Joe, as you would know, you often do this in Cambridge when it's windy and rainy and everything else. The one thing I really worry about is there's a very high correlation between me turning up and the team losing. So, I have to figure out what is the right approach to that. Maybe AI can help me minimize that correlation because it's really way too positive.

Sudeep Kesh   00:24:56

You need more data.

Joe Cass   00:24:59

Excellent. And Sudeep, we spoke about this briefly before, and you also said you had some experience in this. But in your view, how could AI impact other sectors? So will we be watching AI movies in the future, listening to AI music artists, watching augmented reality, AI-infused football games?

Sudeep Kesh   00:25:21

Yes. I mean I think we already are, right? So, it's part of the answer to the question. And I would go back to what Mohamed said, not necessarily 80-20, but just treat these as a continuum. So, like 100%. What I fear is that if you were to watch a movie or listen to a music that was 100% AI, I worry about the implicit nature of data forming sort of this average and like you're not going to get exceptional music, but you'll just kind of get like 'meh' and maybe that's fine, I don't know. I call it - I don't mean it, but it's basically if you take an asymptotic relationship of as X approaches Fleetwood Mac, it's kind of like it's all right. That's all right. But it's just like it's not necessarily exceptional. So, it's just kind of like that's some of the stuff I worry about. With sports, I know in football, English football, when you started introducing cameras and replays and things like that, like everyone just went nuts. And I think it's one of those things I really enjoy the beauty of like watching like Messi in action where he just kind of has a slow start in terms of the physicality, but he's immediately studying everything about the pitch and having this calculation. That's kind of how AI works. So, it's one of those things that like if you do too much of it, it becomes less beautiful. So, I think it's one of those things that like it's happening now, like in terms of special effects, in terms of things with convolutional processing and music to allow some pure sonic space sound like it's from something different or create a lot of things that you can't even emulate on earth. This is some of the stuff that I was doing of trying to imagine what a guitar string made of glass would sound like. So basically, taking the physical properties of glass and the reverberations and things like that and then applying that using sort of computational techniques. And then that kind of influences my own imagination. But it's just like I, as a human have to do the hard work of imagining things. And when I'm consuming, I'm also viewing this thing through sort of my own sort of lenses of experience and preferences and everything else. So, nothing is ever devoid of the human, so I don't worry about that too much. But I think in the dialogue, we tend to overbias towards some of the stuff around AI. Hopefully, that made any sense whatsoever. I'm not sure that did.

Mohamed El-Erain   00:27:37

So Sudeep, let me push into this. Suppose you are the coach of a team of a football team, whether it's American football or soccer. And the other team has fully utilized AI to put in every data point about your players, strength and weaknesses, has observed your formations and everything else. And you're a coach of the team that's playing against the AI-enhanced team. What would you do?

Sudeep Kesh   00:28:06

I think it's that during Game Day, and we have this thing all the time. The analogy is funny to me because I'm the least athletic person like not in just this room, but like maybe on earth. But I think it's one of those things, come game day, you got what you got, right? So you got to play. And I think that you practiced time and time again using sort of your human intuition using the sort of communication channels on the field. That's one of the things I always loved about like the Xavi's and Andre Iniesta's of the world of like basically reading the field and then using all of that information as sort of a communication strategy. It's almost like a dance. And I think that would still allow a lot of things that haven't been assessed or analyzed by the AI bot on the other team. like that sort of communication and teamwork as an art form, I think, would still hold true. Now I don't know how that's going to play out. But I mean, that would be an interesting match, right? Like if you say like if people knew that, and then you can kind of say, who's reading for the AI-assisted team, who is reading for sort of the analog team, like it would be interesting. I don't know who would win.

Mohamed El-Erain   00:29:11

So, I thought you'd answer differently. I would have thought you'd say on T minus-one, you would change your formation and you would take the risk of a suboptimal formation rather than an optimal formation that the other side has analyzed fully.

Sudeep Kesh   00:29:28

Yes. I mean, because I think in American football, they do this all the time, right? Like they analyze like every nuance of everything. And I don't know that it always generates an advantage because you always have - this is like all of the psychology and the computer science games of your two or chess. I mean this is chess, right? Like you have all of these probabilistic and deterministic kinds of things that could sort of happen. But ultimately, you have to make a choice. And that choice is unknown until you've made it. So I think that it's still the same sort of dialogue in a different context.

Mohamed El-Erain   00:30:07

Essentially because I grew up at university playing the game of risk every single day. And the game of risk is completely probabilistic led. But when people start understanding your biases, you have to act irrationally once in a while.

Joe Cass   00:30:21

Right. Mohamed, as President of Queens College at Cambridge, you must engage. In fact, I know you engage with the students regularly. In fact, as part of the research for this question, I did some mining online. And I think it's a Facebook group, which is called 'Mo's bros', where there's a group of people at Cambridge, guys and girls who are part of this group and just value just speak to each other about what kind of insights you shared. So I know this happens. So I'm interested to know what kind of things have you learned from the students and also kind of the Gen Z generation more broadly?

Mohamed El-Erain   00:31:04

So let me just say this is the best job I've ever had. And one of the reasons why this is the best job I've ever had is interaction with students. They come to us from very different backgrounds, and they are wicked smart, as you would say, in Boston. They really, really are smart. And for me, it's just incredible talking to them and asking them questions and seeing how they think. And when you are with them for three years, you literally see them being transformed by the environment. You see the intellectual curiosity being satisfied and pushed. And in the process, they teach you a lot. So I try to spend as much time with students as I can. That includes not just meals. It also includes stopping and talking to them, having them over. And I think it's absolutely transformational for me as much as it is for them, this experience. Now I get the added advantage that I get to sit with professors and whenever we get someone who comes from a U.S. university who comes to Queens, the thing they love most is lunch because what happens in lunch, which is self-service, is you sit at the next available seat. You don't go sit on your own, you don't go sit in a group, you sit at the next available seat. So one day, you'll be talking to a physicist. The other day, you'll be talking to a humanist. I mean you get just incredible experiences from people who are at the top of their field. And it's just incredible interaction between the students and the professors there. That's awesome.

Joe Cass   00:32:45

Fantastic. Mohamed, I know you're a Board member at Under Armour, and I think - you can correct me if I'm wrong here -  I think I saw an interview where you had an Oura ring on. I thought it looked like an Oura ring. But interested to know, there we go. So with that kind of in mind, interested to know how you integrate kind of health, fitness, wellness into your daily routine.

Mohamed El-Erain   00:33:09

Not well enough, okay. So you asked me earlier about data. I love data. I love evidence-based. And I never understood my sleep. And then someone introduced me to the Oura ring. And the Oura ring has had two massive impacts in our lives. One is individually, which is every morning, the very first thing I do, even before I check the markets, the very first thing I do is look at my sleep. In fact, when my wife and I wake up and we said, how did you sleep? We said, I can't tell you, I need to look at my Oura score first. And I've noticed that it has changed my behavior that I've done things differently in order to try to promote a higher score of sleep. And importantly, when I do miss out because I fly to a red eye, I understand what it is that I have to make up for. So that's number one, it has changed my approach to sleep. Number two is like other things, you can compare and contrast. So we have a family group, and I have two daughters, and it's wonderful to me that we compete on the amount of sleep because every night we want to know who has the highest score. And I remember when I was in my 20s, which were where my daughters are right now, sleep wasn't a priority at all. I wish it had been, but it wasn't. Well, the fact that we are all competing for a high score has made it a priority.

Sudeep Kesh   00:34:39

Is that recursive then? Meaning that it's just kind of like, okay, like I wake up in the morning and I didn't get enough sleep. So I know I'm ill-equipped to make good decisions in certain context. So I'm going to refrain from doing that, but it also creates an incentive to say, look, today was a less than optimal day because I hijack myself by not sleeping. So I want to get more sleep in order for tomorrow to be a better day. Is that more or less how the thought process works?

Mohamed El-Erain   00:35:08

I wish it was. That's why you're the Innovation Officer because you think that way. No, mine is very simple is there are different metrics. They measure different metrics of sleep, your REM sleep, your deep sleep, your restfulness. So I look at what didn't happen. And then I try to figure out why is it that I was so restless during the night. What is it? And it's often because I read an e-mail before going to sleep or something like that, that my mind kept on thinking about. Is it because I didn't get enough hours sleep? Is it because it took me too long to get to sleep? So what I try to do is incorporate the data and try to change my behavior so that the next day, the next night, I'm better. But no, I don't do the advanced science that you do.

Sudeep Kesh   00:35:59

That I wish for. I don't do it either.

Joe Cass   00:36:02

It's interesting. I've got a similar one, again not an endorsement of the company, but I've got this thing called a whoop and it's very similar to the Oura ring in terms of sending the data of the sleep and recovery and if you can push yourself or not. And the one interesting thing, you're right in the fact that it changes your behavior. And the one example I have personally is that I found that if I don't eat or drink anything after 7:00 p.m., it incredibly improves my sleep score. So I did it kind of one just randomly kind of by mistake, I guess. And then it had such an impact that I just repeated it. And now whenever I can, let's say, five days out of seven, I've made that change in my life just because of the data based off the loop, which is kind of scary the power that it has.

Mohamed El-Erain   00:36:53

So will you not accept a dinner invitation at 8:30 at night?

Joe Cass   00:36:58

No, that's the thing that I try and kind of keep the kind of 70% to 80% rule with kind of health and fitness. So I think, okay, I'll be good in these areas, but I don't want to kind of sacrifice to an extreme. But it is crazy how kind of just tracking the data and being visible and popping up every morning on your phone, it changes your behavior as a human. So it's so powerful.

Mohamed El-Erain   00:37:25

It is. Absolutely.

Joe Cass   00:37:29

Sudeep, we've spoken about Gen AI throughout the podcast so far. I'm interested to know how you're using Gen AI in your personal and professional life at the moment?

Sudeep Kesh   00:37:40

Sure. Yes. So Gen AI, I use less admittedly than what they call discriminative AI and things like that in terms of machine learning for work. For Gen AI, so when my son was small, I mean, he's still small, but when he's smaller, He would always ask me to read him the story before bed and then we turn out the light and then he wants to hear another story. So I would make up a story. And I would try to incorporate different elements from his day. Now I think as he kind of ages, then things get more complicated. So I said, let me try asking ChatGPT to come up with a story using certain variables. And it will be like electric cars, bunny rabbits and the band Paramore. And then it was like, okay, what do you come up with? And what I found, I've tried this like again and again with a bunch of different variables. The thing is the stories could be pretty good, like in terms of like little plot twists and stuff like that. So like in personal life, I'll use it for that. Sometimes you could start to see a pattern pretty quickly after like you have about three. It's inspired me to kind of just try again just naturally of just kind of coming up with stuff. So that's kind of some of the personal life stuff. In terms of professional life, I think it's really good for summarization. Like so for example, if you have regulation, going back to Mohamed's issue about attorneys, for some reason, they can always generate these six hundred, seven hundred-page papers that it's like really, really difficult to read. So I think in terms of document summarization and things like that, a lot of the algorithms now are really good at getting to the heart of the material and give you enough of a flavor of it that you can essentially kind of use your sort of manual research processes to then fine-tune what I want to learn more about and things like that. So I think like that's a really good use. I think we have to be mindful of the biases and things like that, that are sort of brought about. But like Mohamed was alluding to, if you're not using this, then you're truly going to be behind. So I think it's learned how to use them well and use them. And it takes some practice. So I would also just kind of think about just what are the stakes of the outcome for doing it this way and so on and then build some governance processes around things that are going to abate something bad from happening. So that's on the professional life on the personal, it's glorified mad libs.

Joe Cass   00:40:14

Great, Thanks Sudeep. Mohamed, the last question goes to you. Over the course of your career, what's the best piece of advice you've been given? And who gave it to you? 

Mohamed El-Erain   00:40:25

I think the best piece of advice I got was from my father. I was thirteen years old. We moved a lot when I was young. And I remember that we had arrived in Paris, and we got four newspapers every morning. My father expected me to read the four newspapers. I had no interest in reading one, let alone four. And I remember trying to strike a deal with him. I said, 'look dad the news is same across the board, so I don't need to read four newspapers. I'll read one. You tell me which one you want me to read and I'll read one'. And he said, 'no, you don't understand. The interpretation of the news differs. And in front of you, you have four newspapers that go across the political spectrum. And unless you understand how different people interpret things, you will struggle in life'. And that, for me, was an incredible insight in terms of, you do need to keep an open mind. It's really important to have this cognitive diversity almost hardwired inside of you because it's hard, it's really hard. And I'm glad that my father was so insistent when I was thirteen. I can't imagine that my goal is now to read newspaper, let alone four. But for me, it was really important advice.

Sudeep Kesh   00:41:50

Do you find when you're talking to students, are they seeing the level that you do in terms of the need for cognitive diversity and just being able to just experience life with just different friends and different people from different walks of life. Are you finding that with the students?

Mohamed El-Erain   00:42:09

I find that you have to structure it. You have to let structure do the heavy lifting because in the world we live in today, you will tend to go to one point of view because social media, as we know, is a big driver of this. And it's understandable. They're just trying to curate for you what you're seeing. So they reinforce whatever it is. So you have to use structure to do the heavy lifting on this. And that's why we stress cognitive diversity so much.

Joe Cass   00:42:44

Fantastic. Well, listen, that's it. Thank you so much, Mohamed and Sudeep, for your time today. Everybody watching everyone listening, see you next time on FI15.