Raphael Douady - From Theorems to Wall Street
Pioneering finance mathematician Raphael Douady talks about his career, risk, and the value of teaching.
SUBSCRIBE NOW ON YOUR FAVOURITE PLATFORM
Episode Summary
In this episode we're joined by Raphael Douady, a pioneering academic who transformed his career from proving complex mathematical theorems to becoming one of the first academic quants on Wall Street. Discover how Raphael navigated the challenges of transitioning between diverse fields, from academia to the space industry, and ultimately into finance. He shares fascinating anecdotes about career serendipity and the evolving financial landscape, and invaluable insights of thriving in unpredictable environments.
Gain a deeper understanding of financial entrepreneurship and risk management with Raphael’s advice on industry challenges. From an unexpected career move to China, cut short by a global pandemic, to his current role advising hedge funds and family offices, Raphael illustrates the critical importance of mathematical techniques in risk management and the art of seizing opportunities. He candidly discusses missed opportunities and the lessons learned in preparedness, offering a poignant reminder to stay informed and proactive during financial crises.
Raphael’s passion for teaching and data science are also key highlights of our conversation. We talk about the creation of his mathematical primer courses that break down complex ideas for diverse audiences, and experiences teaching uncommon students, like Navy SEALs on the RWRI course.
We also explore the transformative impact of data science in industry and its potential to revolutionize finance, akin to past industrial shifts.
Raphael also brings a fresh perspective on navigating political ideologies and social media, sharing personal experiences and philosophical insights. This episode is a rich tapestry of stories and lessons that underscore the power of adaptability, learning, and risk management in an ever-changing world.
Transcript
PLEASE NOTE: Transcripts are AI generated and may contain slight discrepencies.
Sebastian David Lees
00:00
Welcome everyone to a new episode, and indeed a new series, of the Fat Tony's podcast. Before we get started, I just want to say an enormous thank you to all of our guests from the first series and all of our listeners. The first series exceeded my wildest expectations in terms of listeners, so really, really, thank you all so much. It's really made the second series possible.
00:25
So, moving on, our guest today is a man of many talents. He's held the Frey Family Endowed Chair of Quantitative Finance at Sony Brook University and, prior to that, was an academic director at the Laboratory of Excellence on Financial Regulation. He has more than 20 years of experience in the finance industry and 35 years of experience of research in pure and applied mathematics. If that wasn't enough, he's also an entrepreneur and has co-founded FinTech Firms, risk Data and Data Core, and his current research focuses on systemic risk and the anticipation of financial market crises. There's so much more I could say about today's guest, but it would take the entire podcast, I think, to list his accomplishments. So all I will say is Raphael Douady, it's an absolute pleasure to welcome you to Fat Tony's.
Raphael Douady
01:14
Thank you very much, Sebastian, for this kind introduction. Yeah, I like to say that I have several lives behind and hopefully ahead of me as well, but my usual joke is that I'm schizophrenic by at least three axes, since, first, geographically, I live between Paris and New York. I'm currently in London, actually, and currently in London, actually, scientifically, I've been, you know, oscillating between hard science math, cow theory and softer science economics, finance. And the third aspect is, yeah, of course, the division between the academic life and the entrepreneurial life, which again requires completely different sets of skills.
Sebastian David Lees
02:11
Well, I always think it's healthy to reinvent yourself every 10 years or so, and it certainly sounds like you've done that. So a life well lived for sure. I want to get back to the start and talk a little bit about the beginning of your finance life, and you were really part of one of the first generation of what I would call academic quants that came upon Wall Street People with physics, mathematics, PhD backgrounds. I'm curious to hear what the reception was like in the finance industry to that new way of thinking and what the environment was like, and if you have any anecdotes from that time.
Raphael Douady
02:50
Oh yeah, many. It was a bit by chance. I mean my first experience outside of pure mathematics. Really, I studied as a pure mathematician and there was a pivotal year in 86 where I proved the major theorem a pivotal year in 86 where I proved the major theorem a question said by a famous mathematician and physicist, vladimir Arnold, and so that made me famous. But at the same time I had the need to do something else and that's where I went to work for the space industry at that time. So for about seven years actually I've been working in that space, first as an engineer I mean, really dove into the pond of engineers, forget about the fancy mathematics, et cetera and with pretty hard problems. Some of them were really hard. It was the beginning of imaging, dealing with the capacity of computers, with the speed at which you had to react, you know, in the satellite control or spacecraft control. And then I, and then as a consultant, and then at some point there were some technical reasons but I had to move away from the space industry.
04:08
This one was not my choice, but I was very attracted by the, by the banks, and if you want an anecdote, I had really because my family was much more academic. My father was a very prominent mathematician in France, a member of Wolbecki, etc. And I wanted to do something else and I was very attracted by the banking world, which I had no clue of, truly no clue of. So my first reaction was to say if I want to know what's happening behind the scenes, I have to go to the corporate finance, so M&A. And then people look at me and said, well, why do you want to go there? I said, well, I'm interested. And then I had many interviews but never a job offer. And then I had many interviews but never a job offer. And then one day one guy told me you know, people will always take you to interview because they barely see people with your kind of profile. They've never seen that a pure mathematician like that, it's a complete foreign alien for them. So they will always give you an interview but they will never give you a job because in this world the only thing that matters is your Rolodex, and you have no Rolodex. Okay, sir, message understood, let me go and have an executive MBA. So I went to the only school that was offering this real type of executive MBA that's in France. And he said what do you? As a student. We usually hire people like you as professors, not as students. I said, well, obviously some line is missing on my CV so I need to fill it in, okay. Okay, if you want to sew, then okay.
06:03
So I was to start IN Sead in January 1994. And that's no joke, the Christmas Eve so it was the 23rd of December of 1993, a friend of mine who was a trader at SocGen called me Raphael. I know you want to go to the banking world, I know you need money, etc. So tomorrow morning 9 am, you have an appointment with the head of capital markets at Soggen. Wow. And then I said you know very well that I'm starting in SEAD in a few weeks, just after New Year. I can't do that etc. Don't discuss. Tomorrow morning 9 am, okay, sir, okay, I'll be there.
06:54
And I have got the most, the craziest interview. I mean it's like you know the reverse interview. I mean it's like you know I was explaining to the guy. Well, you know, I have no clue of capital markets. Even when I watch TV. I mean I see the CAC 40 has that on the channel. That's no problem, we'll teach you financial. We need your math. We'll teach you financial markets. Wow, you know people who pay you to teach you something, okay, and then I said I said you know a figure which to me seemed very good because it was just three times what I was earning in the space industry. So I just took myself in the space industry, I just multiplied by three. I said is that okay? Oh, yeah, no problem, okay, very good. So you start next. It's just basically the day I was supposed to start in Seattle. I was supposed to start SocGen.
Sebastian David Lees
07:44
Wow.
Raphael Douady
07:45
And then so I go to the trading room, see my friends. How did it go? Well, you know, I got the job. But how much? I said well, I said to the figure I said did you ask or did he propose? No, I asked. You're stupid, you should have asked for twice as much. So that was showing you the need for mathematicians. At that time you were asking. You know what was the ambience? It was 1994.
08:10
Very important People with a real math background. I'm not talking, you know, of people who know a bit or, you know, have a mastery math. I'm talking people who have a very strong math background. I was not the only one, but I was one of those who have this very strong math background or physics, etc. There were not so many at that time and with such a background you were at a high standard.
08:37
One of the reasons I realized afterwards is that it was a big time of Basel II. Banks had to implement all these value-added systems, etc. So there was a big time of Basel II. Banks had to implement all these value-added systems, etc. So there was a certain amount of work to be done. But also traders it was the big boom of options since the early 90s. I mean the option market really started in the 80s but the boom of the market was in the early 90s. So that was a big demand for people who were able to do all sorts of calculations of exotic options etc.
09:21
Mostly in close form, so very simple models compared to what we have today, because the computer capacity was not the one we know today and so the interesting part was two years later, in 96, I remember I mean, believe me, 94, 96, there was no competition. You arrive, you have some good training to do efficient algorithms, you get something that works. You know it comes to us slower. So an optimizer, you know that would take them. You know hours, you make them work in minutes or in seconds. People were completely blazed and you know saying, well, ok, how much you want, etc. So that was really the atmosphere.
Sebastian David Lees
10:00
Was there a suspicion of a pure mathematical approach at that time? Was there still a belief that to be a trader was instinctual?
Raphael Douady
10:07
It was split. It was split and jobs were actually pretty much like we have today with AI. I mean, we'll go to that. It was split because there was, you know, the quants were the quants, traders were traders. Some traders were quants. You have to know that France is very mathematical, so a lot of traders were from Polytechnique and et cetera, so they have some good math training, very unlike the. Even in the US you had more. Surprisingly enough, you have more quantitative trained traders than in the UK, where you have this idea of picking, you know, on the street, 16 years old kids, like you know. You pick them to be hackers, you pick them to be traders. You put them on the screen and the best one you know will make it, etc. And then eventually they will learn the math.
Sebastian David Lees
11:02
That's interesting, so like it's almost as if France's mathematical heritage allowed it to be slightly ahead of the curve.
Raphael Douady
11:10
I don't know whether they were ahead of the curve business-wise, but they were definitely ahead of the curve quant-wise. And that was even when I went to the US. So I went to the US in September 1995. And simply arriving there, there was a reputation as the french quant. The french quant was a reputation right, that's fascinating, okay.
Sebastian David Lees
11:30
So I want to move on slightly. I feel like I could talk to you for the whole hour about that, but I'm going to try and stay disciplined. Yeah, so these areas of kind of financial mathematics you work on and, uh, you know, indeed taleb and other people work on, you know, tail risk, hedging, systemic risk, anticipation of market crashes. One question I often hear uh in the community is is it possible to to use these strategies as a retail investor, or do you really need the scale of an institution to make this kind of stuff work?
Raphael Douady
12:04
I mean, of course, you can use it as a retail investor for the good reason that you know, I left studybrooke at the end of night 2018. There was some financial issues in the in the university that made it difficult to run the program, and I had a very good opportunity to work with a Chinese startup that was well-founded. The project was very attractive. It was in Shenzhen. Shenzhen is a bit Chinese Silicon Valley very enthusiastic people. The only thing that I didn't anticipate was COVID. So the project blew up in the middle of 2020 because it was evident that we could not sustain the restrictions led by COVID. Since then, I'm still having my PhD students.
12:55
I took back my position at Paris One on the academic side, but all my entrepreneurial life goes to be an advisor to small funds and actually I'm only advising small hedge funds, family offices etc. And really I wouldn't call that complete retail. But it's not established firms, you know, managing billions. It's firms managing, you know, in a much smaller amount of money, often family offices, who just want to improve their risk management or firms who have some cash and who want to improve their cash management. That's exactly the type of people I'm advising. So, yeah, it's trading strategies, risk management, better understanding statistics and often really going into their own trading and trying to improve it, to find ways. So yeah, that's pure retail and, believe me, there are some math. I mean there are some real math. We are doing simulations and understanding what's going on.
Sebastian David Lees
14:07
It's good to know. I know there's a slight mythology around some of these techniques that have been built and it's really interesting to know that it is possible at a retail level to say an anecdote about that, which which tells you know the, that that precisely you know.
Raphael Douady
14:32
That breaks a bit this a priori that people have about using math, etc. So it's a friend of mine. I know his family is wealthy. Uh, he was a some somebody I knew as a student very good friend of my brother, and so their family made some money, I think, in the real estate real estate operations. Himself is in computer science and his brother is managing the family money. So, basically, and I know the whole family and one day there was a party, and it was slightly after the 2008 crisis, and I remember his brother coming to me and so we're talking.
15:12
So how is the business going? How did you survive? Oh, you know what? I lost three quarters of our money during the crisis. It's terrible, et cetera. I mean really painful. I mean there's family money, et cetera. I mean really painful, I mean there's family money, et cetera. And I look at him and I say you know that. I mean we know each other. You know exactly what I do. Why, when you saw the crisis coming, why did you? Why did you? Did you even? You know, talk to me about it. It's not about consulting. I would have give you my advice for completely free I mean we're friends, etc. It's not you know, we would look at it give you some basic things to avoid this kind of catastrophe. Yeah, yeah, I realized afterwards when I saw what happened. I said I was stupid. I should have called Raphael.
Sebastian David Lees
15:57
It reminds me of something you said at RWRI. Yes, but the secret is to buy the insurance first, then the car, once you've secured the insurance.
Raphael Douady
16:07
Yeah, exactly Exactly, and you will see, so often people are in the. I mean, I can't blame them. I've done that mistake myself many times. You know you're involved in something I can say. You know I was involved, you know, in my risk data business, and then it was 2011. And a friend of mine, very, very good friend, very close friend I mean we were doing math together for years, et cetera and you know, he told me you know, there's this stuff, bitcoin. I'm going to put a thousand euros in it, but give me anything. I mean, give me even a hundred euros. I will just, I mean even 100 euros, I will just, I mean it cost me to simply, you know, wire him 100 euros to one of my best friends, and don't worry, I manage it, etc. He was telling me he was passionate with this. You know, byzantine general system, etc. And he told me, I manage it. Today he has a fortune and I missed it and just simply not because I didn't want to, because I forgot, I was busy with some other stuff.
Sebastian David Lees
17:09
And that happens, that's life right. So this is interesting Talking about advising against this kind of thing. So I've heard you say and I heard Taleb and a few others mention this as well that every model has an error, every single one. Every model has some sort of error, and whatever you estimate from a model must logically also have an error. So I suppose my question to you is how do you go about identifying model error? Or, if you know you can't identify model error, how do you account for it?
Raphael Douady
17:40
so I rather the second thing, because I. So there are two things. There is something that is today, people give it the word, you know, self-organized criticality. But basically, very often you will find that even a model that is on the spot, you know, for some time, the market will tend to drift to the edge of the model because the model is being arbitraged, because it's being played, etc. So we'll tend to drift towards the model invalidation and some other model.
18:18
I was giving the example of arbitrage pricing theory and equilibrium theory in economics, where they are almost opposite. One says you know, things will converge to equilibrium and the other says there is no such thing as an equilibrium because if it stays something it will be scaled up until. And, of course, the arbitrage argument, which says if there is an arbitrage it will be cancelled because it will be used up to a volume that makes it, you know, inexistent. Then precisely that volume is in fact an equilibrium volume. So you have a this is a good example where you have a balance between two theories and the actual trading of banks and funds and so on will happen at the edge of those two theories, where none of them is valid.
Sebastian David Lees
19:13
So it's almost like you have these theories pulling in different vectors and the reality kind of emerges out of kind of placed between these.
Raphael Douady
19:23
Yeah, so you will see this phenomenon often. So the question now is where do you identify, et cetera? I mean, I like this joke, you know, of this German box. You know who was saying all models are wrong but some are useful. And so the question is what is the difference between the useful model and the useless model? And a model for which you twiggle hypotheses and you get something that is still meaningful not completely wrong, at least signs are correct, et cetera and a model for which as soon as you change the least thing, then it drives you to the complete wrong decision. Then that makes a difference between models that's useful. So it's kind of a robustness of the model that matters to say that the model is useful, so the fact that the model can resist externalities, breakage of hypothesis.
20:24
The classical one that we use with Nassim is, of course, the fat tails, but there are plenty of things other than fat tails that may happen, and so a model that is kind of robust to all those different things is what you know makes it useful. So, yeah, by how much that's experienced, et cetera. So there is precisely, you know, that the difference between reality and the reality always come with, and that's something that you learn by doing engineering, you can do, you know beautiful math, et cetera, but at some point you have to stay fit on earth. Beautiful math, et cetera, but at some point you have to stay fit on Earth. And I was lucky enough to be kind of educated scientifically in this mood of having the head in the sky but the feet always on Earth.
Sebastian David Lees
21:21
It's interesting I'm a software engineer by trade and certainly in software engineering your mental model versus what can actually happen, because in software things tend to be more and more interconnected interconnected now so you have scaling errors of non-linearity where things can go catastrophically wrong if you're not careful. And we saw that earlier in the year, the, the security company who released the update, and then half the airports, half the flights around the world, were cancelled back in oct I can't remember the name of the company now. So it's interesting, I think as well. Something I've heard you say before is about the importance of information. So when facing a threat of unknown importance or an emergency, information has the highest value. So without it we tend to take paranoid measures. And what you were saying about all models are wrong and I think I've heard it paraphrased as all models are wrong, some are deadly. I think people can misconstrue that to mean models are useless, and I don't think that's quite true. Models can be useful, but you just have to be aware of what you're dealing with. It is a model.
Raphael Douady
22:25
That's a good point. The other point where I have a little disagreement with Nassim is more, I think, in the phrasing of it, because when we speak, when I see Nassim acting, he's not mad at all. But paranoia itself can be deadly and people don't realize that that is. It's not a safe. I mean the famous take a snake for a stick is better to take a stick for a snake than a snake for a stick. Sometimes taking a stick for a snake itself is deadly, and that's so. It's all a matter of excess. That is. The radical attitude is very often not the right attitude at all and it's a wrong thing. It's a wrong idea to think that you have a complete safe haven by taking a radical attitude.
Sebastian David Lees
23:24
It's interesting. I think that applies to so much in life, not just finance. But you know I've often said any. Any philosophy taken to its extreme is self-destructive. Oh so, you know, it's about that balance. Okay, I'm going to move on again, again. I could each one of these areas, I could spend an hour talking to you, but I want to move on a little bit to another recent thing in your life, which is your primers courses that you started doing, these mathematical courses I want to ask you about. Obviously, you are a man of many talents and you've got your fingers in a lot of different pies, but what? What led you to create this? What? What was the impetus and why do you think they were needed?
Raphael Douady
24:06
well, there's a history that made it like. I just told you about the history that brought me to finance, which was a bit unexpected. There are two things. First, okay, don't blame me for arrogant, et cetera. It's just a matter of nature. I'm a born teacher and everybody tells me that about it around me. Even my wife is joking that you, you know, if I don't teach enough, then I become pressure on the environment, etc. So so, yeah, that day, this thing, you know. I like explaining, etc. I like finding the right images. I'm very curious about the psychology of everybody to see what are the trigger that make them understand something, et cetera. I like conveying. You cannot invent it. You are or you are not. And then was again. So wait a second. Yeah, we saw RISData in 2014, the software company, because I have a software and a consulting company. It's my consulting company. It's still called RISData. The American company, the French company, was a software company. They were basically, at that time, the consulting company was essentially consulting for RISData clients, so it was a RISK Data Inc and so this one stayed. But we sold the RISData SA as a software, and it was just before I started.
25:37
Both Tony Brook and DataCorp and Nassim said, okay, well, you know we could do those teachings. What do you think of it? Okay, let's try it. Let's do a try. So we did the first session of this real-world stuff in 2014 and it was successful. It was a very nice people liked it. And then I joined stony brook. I met robert fry at that time very good tie and robert loved the idea. So we had a meeting with robert, nasim and myself and we decided to make it on a regular basis. So I think there was one session, trial session, in 2014. Then we did the regular. At the beginning was twice, three times a year. Then now it's back to one time once a year. Sorry, but the so. So we decided to to start this and again, you know, this is this little thing where I like teaching students, but I like teaching to people to whom it's hard to teach. And I remember, for instance, one of the session we had.
26:48
Nassim just gave a talk at the Pentagon and there was a group of 10-ish Navy SEALs Navy SEALs coming to mathematicians to learn about risk. Isn't that funny? It's just impossible. So they were here At that time. It was physical in a room, it was at the Harvard Club and we were having a drink on the.
27:12
So it starts on the Monday, tuesday evening we're having a drink and you could see the guy. One of them was a giant like you know, like the, the Dwayne Johnson and his kind of caliber, and and the others were all you know, they're pretty solid, etc. And they were always all you know they're pretty solid, et cetera, and they were always making remarks. You know, like guys, when you're on the ground what you need is a knife and a lamp. The rest is just secondary. You know you're teaching that.
27:47
And then we're on the Tuesday evening evening and having a drink, you know, with the guys, with the group, I mean not only those navy seals, but I was talking to them and say, guys, I'm a bit bothered because I mean how us mathematicians, you know, are going to teach you anything in terms of risk? I mean, when we teach traders, we can understand, we have maybe some edge, but respect to people like you who are on the ground risking their life every day, does that make sense why you're coming here? And then they say, raphael, you don't realize you're giving us the language we are speaking to our hierarchy and the hierarchy that giving us, you know, principles and things that make no sense for us, and now you're giving us the words to answer them.
Sebastian David Lees
28:28
You know, I've heard that before, not with rwi, but with nassim's writing as well. I've heard people say that it's giving words to things they've intuitively haven't been happy elocution to be able to put it into words. Yeah, exactly, really thankful for that. And you know, I I was on rwri this year and it was an absolutely fantastic experience and, yeah, just just amazing. And you're wanting to teach, not the impossible, but teach says it's difficult to teach.
28:58
That's fantastic as well yeah, exactly so I'm going to move on again. Normally what we do in these is I ask the community if they have any questions they want to ask our guests, and normally we get you know a handful of questions. You have any questions they want to ask our guests? Normally we get you know a handful of questions. You have had the most questions of any guest on this podcast ever. So I'm going to devote a little bit more time than I normally would to asking these questions. So these might be a little bit disorganized, a little bit rapid fire, but this is always fun because you never know what people are going to ask's a little bit random. So the first one I'll ask what are you long and what are you short on for the next four years?
Raphael Douady
29:36
they've said it doesn't necessarily have to be financial yeah, at a point in my life where I have basically no choice and on the one hand, I'm I made many mistakes that made me today still needing to work to pay the rent etc. I'm not sitting on a big financial cushion and I think things have changed. I mean, I've been trying to manage different horses. Probably that was by itself maybe a mistake. I should have focused on having, you know, more pure phases where I was doing one thing and then another thing. So I would put aside the risk data period where really I was completely in risk data. Then, you know, I was again, you know, split between academics and Sarah had the laybacks, had the, the, the thex, had the study broke. So I mean I'm trying to focus now. I have no choice to use, you know, those few years that I have to work hard to really make, you know, some financial cushion that I need for myself and my family. That's a bit, you know, while I'm long off in some way, it is entrepreneurial things where I need to make something that will work out with all the experience I have accumulated. So I'm advising a hedge fund led by a friend of mine who is hopefully will be very successful and, again, you know, really making all correct, avoid all the mistakes that we can do there will be, etc.
31:24
What's the risk? The risk, you know, same thing as we started, risk Data we took in the face of the 2008 crisis. We took the crisis actually because just when we started. It's like we started two years before 9-11. 9-11 was already a difficult thing. We made a plus out of a minus by moving to the fund of hedge fund business, you know, helping front of hedge funds, and we were at some point the leading risk management software for front of hedge funds. And that would have never happened if there was no 9-11. We would have stayed in regulatory stuff, et cetera, and we were very successful until 2008. And 2008 was both the absolute proof that our software risk approach was the best risk approach in terms of crisis, and yet the terrible hit to the company because our clientele itself was completely devastated and so we had to restart actually much more the regulations.
32:31
Like you see, it's something like this and that's actually what led me to the LabX. So you see, I mean again, you see I made a mistake by going to some entrepreneurship, because I'm always attracted by this slide, and then I took the COVID crisis in the face. You see, you do things like that. Now I'm again, you know, building a thing which I think, if it's purely entrepreneurial, etc. And the world stays normal, then the chances of success are pretty high. But if I look at what are the risks that I may face is, you know, complete mess up in the world because, for political reasons, the world is unstable today, etc. So that's the kind of thing I'm fearing. But what I'm long of, yeah, this kind of thing.
33:17
Now, unlike Nassim, I think that Bitcoins have a future. I'm currently working on the Bitcoin strategy, purely quantitative. That is many years now that I've been following it. Never again, you see, never had the opportunity to put some money on it and make it even a small track record. So it's still in simulation, but it's like paper trading, but the paper trading is very successful. So now I'm trying to sell that. So, if you're asking what I'm going to to below, it's very possible that one of the success I will get will be simply from cryptos it's interesting what you're saying about regret and about bitcoin.
Sebastian David Lees
33:58
I remember having a conversation with the late great jaffa ali before he sadly passed away. We were talking about regret because I lucky to buy a small amount of bitcoin very, very early and I cashed it in, and he was saying how he does not live his life with regrets and he was telling me an anecdote about in the early 2000s, before the dot-com crash, he had the opportunity to sell a dot-com business for an enormous amount of money I won't say how much, but he said no and then, two years later, it was worth nothing. And I said to him well, how do you live with that regret? And he said it doesn't matter. He had zero regrets for the path he's taken in life and it really helped me. It really really helped me.
Raphael Douady
34:40
No, no, I'm exactly in this state of mind, exactly, and it is a state of mind that you have to adopt. Yeah, no, I mean the the. To be honest, I had some very strong arguments with one of my partners at risk data in 2007 because I wanted I saw the crisis coming and we're in full success and say that's time to sell the company, and then he said, no, no, the company today is valued at 70 million and we want to reach 100 million before we start talking about selling it and say it won't happen.
Sebastian David Lees
35:10
Something is coming and and we didn't do it and that would have been the right thing to do- we all have stories like that, I think, and you can't, it will eat you alive, and you can't try and post-rationalize the things I think looking back. So, okay, let's, oh, let's go back to your teaching, because we've talked about teaching a little bit. How can education prepare students for the future rather than the teacher's path, and what is the role of mathematics in this?
Raphael Douady
35:43
So and that was another question which is deeply related to that what do you see in the future of science? Obviously, you need to teach people what will be useful in the future, not what will be useful in the past. Now again, we live in a world where the things are being redefined. We live first, we live through a revolution. The importance of it is as big as the industrial revolution in the 19th century. So the computer it started with the internet revolution. You are mentioning the dot-com bubble. Now we are going through AI, which is a natural follow-up on the internet revolution, which is a natural follow-up on the internet revolution. So what does that mean in terms of science? In terms of science, a new science is born. It is born. It's not, it's going to. It is born.
36:43
It's called data science, and data science is, you know, a word that encompasses several things. In the past, we were talking about statistics, information theory, decision theory, control theory, etc. Data science contains all of that. But why is it a science and not just a branch of math? Because it's not math. You see, first, statistics have never been really math. There was always a part of statistics which is practical and a part of statistics which is theory. Exactly like physics, physics has its experimental aspect and its theoretical aspect. So you have people who, in theoretical physics, will think like mathematicians Mathematicians, the kind of math that I've done in cow theory etc. Is obviously very closely related to physics. I've worked a lot with people doing mechanics. So, moving into the now, data science has again its experimental part and its theoretical part. There is a big thing.
37:52
I had some arguments with people you know about whether artificial intelligence itself is a science. Yeah, why not? You can call it this way. For me, I would see that as a part of data science, because I have this definition of data science. Again, that's wording, so people will differentiate. You know the I don't care. I mean because there is a lot of marketing in here and let's try to put ourselves aside for marketing and try to just, you know, put words that best describe the topic. So you have a new science and I think we need to teach people about this new science with all its methods. I consider that I've done that a lot, that there's some areas, that of this science that I master, some areas that I don't master, and no less, no more than any any you know physicists who've been really involved in the, in the, you know cutting edge research, are also mastering certain areas and not others, but definitely I have this dual experience in theory and in in experiments.
Sebastian David Lees
39:07
It's uh it's that got that intersection I think that you need between practice and academia and, uh, you know famously mandelbrot, who I know, actually when you were a child, oh, yeah, Mandelbrot.
Raphael Douady
39:19
Yes, he was a close friend of my father.
Sebastian David Lees
39:23
Yeah, Wow, just absolutely fantastic. And you know operating that intersection between theory and practice, and maybe that's what we need to encourage a little bit more. There's a few questions that related to what you just answered. I'm not going to ask you to repeat yourself because I imagine they're very similar, but people are asking you know, how do we train the next generation of scientists? People have been asking about what scientific disciplines do you think will shape finance the most in the next 10 to 20 years, and I can imagine I know the answer from that, just what you answered no, I mean yeah, I mean the, the.
Raphael Douady
40:01
You see that what happened is obviously, I mean, you have to. Really, I like the comparison with the 19th century industrial revolution. Thermodynamics is born from that. People were building machines and because they were building machines cars, cars, trains, etc. Thermodynamics was at the center of it. You had to find the way to convert one kind of energy, namely heat, into work and you had to find the optimal way of doing it, etc.
40:34
And then the other thing that you know, is less popular but it's noble, et cetera, is chemistry. There was a huge development of chemistry. You see that in both cases that went for the best and for the worst. That was, you know, we are very glad today. And I forgot the third aspect was electricity, with the famous battle between Tesla and Edison. And Edison had his ideas, but he was a killer. He did everything to make sure that Tesla's idea would not prevail, whereas Tesla was physically much more on the spot. And the fact that we're using an alternative current today and we're rediscovering even today, things that Tesla was having and was trying to do at that time. I mean, tesla was an absolute genius, it's interesting.
Sebastian David Lees
41:26
It's like we need. When you look at innovation and the evolution of science since the Industrial Revolution, it's been these rapid advancements and plateaus in different disciplines and then another one comes along and you need the combination before you can advance. And you talked about thermodynamics and computing. I look at charles babbage who tried with his analytical engine and difference engine to try and build a very, very early mechanical programmable computer. But because the electrical revolution hadn't happened yet, he didn't. He wasn't quite able to turn that into reality. So you know, you can be the most intelligent, amazing person in the world, but unless you have that, unless the whole system is at the state where you can be in the right place at the right time to take advantage of that, so that's, you see.
Raphael Douady
42:15
I mean basically you have. You will see, in the advance of science you will see two sources this you know industrial need that pushes, you know pressures for finding solutions and that pushes physics, you know ahead. And also, of course, the military. It's interesting to see you know that. And also, of course, the military. It's interesting to see that, for instance, the first usage of planes was essentially military and then eventually people went to civil aviation. Same thing for nuclear and eventually we came to nuclear plants, etc. So there is a lot of technology that starts military and then was used in the civil industry.
Sebastian David Lees
42:56
Yeah, nasa as well. Going back to your space industry, but I know NASA isn't technically military but kind of quasi-military organizations a lot of the innovations that came from the space race.
Raphael Douady
43:07
Yeah, but except that in the space there is something that was also hubris. There was this fight during the Cold War. Both the US and Soviet Union wanted to be the first one there, so they got Gagarin, then they got Saturn V, etc. At the Apollo program there was a competition which was not military. It was like the best you know, it looks like a football game that takes a political dimension, but except it's a football game that takes a political dimension, but except it's a football game with billions of dollars to do something completely crazy, etc. That's neither civil nor military.
Sebastian David Lees
43:45
It's a great analogy, because I think it's even called the nuclear football, wasn't it that the president carried? Yeah, exactly.
Raphael Douady
43:50
So, anyway, going back to what's happening today, we started with the internet and then now with AI. You see, it's something that is deeply in data science and that's why you get this emerging of this new science, because it is needed. You have an industrial revolution and it is the point of that industrial revolution. There's no chance that the companies that have the most value, the biggest market cap today they don't produce actual device, they produce data. When we click on Google, it's free, etc. When we use a search engine, we use Yahoo data, we use, etc. It's all free. The reason it's free is that, in fact, we are paying with information about ourselves. So you see that the economics have nothing to do with traditional economics. It has completely changed the economic relationship between consumers, clients, producers, employees etc.
Sebastian David Lees
45:05
I've heard it said before that you know if something is free, if a product is free, you are the product.
Raphael Douady
45:14
You are the product. That's one of nassim's jokes, yeah, but uh, the point beyond that I want to make is that all traditional economic models you know that, you see, with cost of labor and cost of, etc. Innovation, it's invalid. It's invalid because the the, the basis of the economic relationship has changed completely, because it's no longer a matter of I buy some item from you for a certain price and you will put some new things in it. No, it's data that I'm buying from you. And what is data? Now comes the data, now comes the real thing, and that's the difficult question.
45:53
And that's what people so you see, we're trying to from student, what is a good engineer? Someone who has the basic math that they need. Uh, okay, there will always be a very thin layer of people who are doing research at the top level, and they will, you know, they will push the stuff further, et cetera, the knowledge further, the understanding further. That will be always a very thin layer. Of course, I'm passionate with this kind of people, I think I'm a big part of it, et cetera, and that's what drives me. But, to be honest, if I look at my students, the vast majority are more like engineers. We would have called them engineers in the past century, call them quants, call them whatever you want. I mean that's people who have this role. They're more actually building the systems, the products, the thing that works.
46:50
And a good engineer is someone who has the math that he needs, has the basic understanding of the theory at different levels, and this capability to transform that into a system that works. Today it's by programming, before it could have been, you know, by actually engineering, et cetera, but it's making you know those things that work. And the thing that works, you know is that works is you can have a theoretical calculation of some parameter, but at some point the value of the parameter comes from the usage that you will have to do. I mean, if I'm telling you that the tire of your car has three kilos of pressure, you can make a theoretical calculation, but it just happened that you tried two and a half and it was not enough, and tried three and a half and it was not enough, and try three and a half it was too big, and three works. That's engineering.
Sebastian David Lees
47:40
It's like we have the. You know, you talk about the thin layer of the true pushing, pushing the paradigm forward and then, underneath that, a larger layer of transforming that into application. The engineers again going, again, going back to the Industrial Revolution analogy, but the people who built the infrastructure of the Industrial Revolution and building that out and it's almost like they operate at this maybe what I would call like a platonic fold, where they're pushed against each other and out of it emerges application and real-world application. Yeah, application and real-world application.
Raphael Douady
48:15
Yeah, so if you want to teach, you have to prepare the students for this kind of world, and that's why it's difficult, because you have to pick up in the traditional teaching. Now the interesting thing I noticed is that, compared to traditional computer science, ai uses a lot of math, actually much more math, and so people have to go back. I was in the space industry. In space industry you get a lot of precisely three mechanics, positioning, control theory, etc. These are very difficult maths, not exactly the maths of AI. It's more like PDEs etc. Partial differential equations, but there's some very difficult maths. So even engineers who went to engineer school, they still had to progress in their maths for their particular problem.
Sebastian David Lees
49:16
Yeah.
Raphael Douady
49:17
Same thing in here. They learn computer science, they learn the methods, et cetera, and then they go and work my area is finance, so they will work with hedge funds, with banks, et cetera and then they get all this data to analyze. And then they get all this data to analyze and what is the first thing about data is what is data? What is information?
Sebastian David Lees
49:40
Yeah.
Raphael Douady
49:41
How do you transform, how do you extract information from data? And I see it as practically as easy as extracting metal from ore, I mean from. So you get this earth. You see the pool of data, this big data, as being untreated earth, containing you know there is some information in it, but you know exactly. And then the whole thing is to go from this massive pool of data and extracting information that will be useful for the purpose you want to achieve, like a refinement, exactly, and that's exactly. That's a refinery thing and that's where the value is. That's why data you provide a lot of data. By being yourself, by using all the social network, et cetera, invisibly you're providing data. But that data is raw data. It's unmanageable and all those engineers are employed at least a large portion of them are employed at transforming this raw data into useful information and that contains very repetitive, stupid work that actually progressively becomes robotized, robotified sorry, how do you say in english two?
Sebastian David Lees
50:58
two things I want to pick up on there. Firstly, we're talking about the amount of math in data science and ai compared to traditional computer science, and that's absolutely something I see as a software engineer. There is this myth but you know, programmers you need to be really good at math, and it's not really true. And a lot of colleagues I see now want to go into AI because obviously it's a huge trend at the moment and they're hitting this roadblock of, oh, the upskill required in math is significant if you really want to understand it. Of course, there are tools you can apply and still do very successfully without understanding truly underneath, but there really is a difference. And the refinery you're talking about in data, that's so true Because when you look, even the terminology we use in our industry sounds like refinery.
51:49
We talk about data pipelines, we talk about data lakes, we talk about fire hoses. You know it's all this. So there is this strong analogy there with extraction and refinement absolutely fascinating. And so we're at an hour. I'm gonna, I'm gonna ask maybe one or two very more quick questions that we've got on this list and I'm going to move on slightly, uh, from academics and theory. Uh, so in your opinion, in your profession or in your life, what is one opinion or one position you have that very few people agree with you on?
Raphael Douady
52:23
I have political, provocative idea and I have an issue with social network software because they will tend to put me in a box where I don't belong to, simply because I'm criticizing another box. Know so those software? They're very simple. They put boxes a, b, c, d for call it right, left, controversial, whatever, and they they. If you start criticizing one box, automatically they put you in the other box and then they feed you with information from internet. Let me give you an information, something politically.
52:57
For instance, my entire family has been made on my father's side of resistance to the Nazis. I mean, they were really deeply involved in resistance. There was a famous resistant lady who just passed away. There was books published on her. She speaks of my grandfather, et cetera. I mean, there are really deep things. Traditionally, no, I couldn't say because I was right and left, no, that was not even so important. Some were following the goal, some were more on the left side, but it doesn't matter the deep thing that you know, the family, the DNA of the family, is anti-fascist. We hate totalitarian and myself I would say even among my cousins, I'm probably one of the most prominent passionate about individual freedom. I hate, literally hate, the idea that someone could have the idea of possessing another individual. The idea of possessing another individual or being possessed by another individual is something despise me completely. That's personal feelings you share, you don't share.
Sebastian David Lees
54:09
That's, that's how I feel it's interesting what you're saying about social networks. I think it plays to that human tendency to want to compartmentalize people and individuals.
Raphael Douady
54:18
Yeah, Guess what, guess what Today, when we have all these you know, people who like controlling, putting data etc. And these mainstream things, so I'm very hard. So I'm very hard. I mean, on that side, I've been, you know, laughed at because I was backing completely Trump, not Musk, in his, you know, defense of total freedom of speech. Okay, I'm a bit radical on that side, but that's my personal idea and, of course, as a consequence, especially in France, because I'm following the social network both in French and in English, and in the French side, they put me like, you know, I'm totally fed with the information from anti-vax, from the extreme right, le Pen and this kind of people on the American side, on the Trumpists, including the American side, on the Trumpists, including the very redneck type of Trumpists, et cetera. Because I was criticizing the other side and I cannot feel more different from those people who are collaborationists of Nazi, you see, and yet you know. So in that way, politically speaking, I cannot be placed neither right nor left, etc.
Sebastian David Lees
55:40
Because I will always be an extremely independent thinker box and I get criticized by friends alike on either side and the danger, of course, is that by the algorithm putting us into boxes or encouraging us into boxes, it's all to drive engagement right.
56:02
And you know, in the 60s and 70s they might have said sex sells, and now it's almost like conflict sells and media wants us to be, you know, friends and family, to be at each other's throats because it drives engagement and I see it. I just see it as crazy that I see it like football fanatics or sports fanatics. You get a certain small amount of people and whether they're extreme left or extreme right, and they're shouting at each other and I want to say to them you are the exact same type of personality. Say to them you are the exact same type of personality. You are on the ideological upper, you are the extremists on either side who wear a red shirt or a blue shirt and shout at each other across the football field and, other than the surface difference of your ideology, you're the same person exactly. And unfortunately, social media is driving the people in the center further out.
Raphael Douady
56:55
Yeah, that's one of the effects is radicalizing. Actually, Frederick von Hayek in his book, which I love, it's my preferred book and I was asked, you know that's one of the books I would really recommend very deeply is the Road to Servedom. It's a small book. It's a small book, it's maybe 100, 150 pages. It's a very small book compared to the big volume that he wrote and, as he says, it's not an economics book, it's a political book and he wrote that in 1942 in London. He was Austrian, lived in London, no-transcript, you know, as you say, supportive of two teams.
58:05
We are at the thing, I mean the, the. I think one of the thing, you know that I, for instance, that, uh, it's always the same thing. You see cycles and even Socrates himself was speaking of those cycles, about the autocracy, monarchy, occlocracy, etc. Democracy. He described the whole cycle and it's like a cycle you're a prisoner of, you cannot do anything out. And I would like to say but there is nothing better than capitalism, than the idea of freedom, etc. And yet it contains its own end, because that's what we're seeing today. What is the disease of the day is and you're asking where I'm I deeply disagree with a lot of people, in that we are living in a world where corruption has grown to a point where it's like you know, the frogs that are in a pan and the water is progressively heating, heating and there is no particular trigger point. So you know, a bit of corruption always happens in business. It's if you think that you're going to avoid it, etc. That's completely illusory. And trying to be radical and completely eliminating that's a bit like Lee Kuan Yew in Singapore. I mean, he made the success of Singapore thanks to completely eliminating corruption.
59:31
I can tell you from corruption my grandfather was a medical doctor who decided to fight tuberculosis and he managed to eliminate tuberculosis in France. How did he do that? He did several things. He did sanatoriums for students, he generalized the usage of Atascam, who was remifon, the antibiotic that was used to cure tuberculosis, and he also installed the mandatory. He was the first in the world because when people tell me I'm anti-vax etc. He was the first in the world to do a mandatory BCG vaccine for the kids and that eradicated tuberculosis. It was a joint action of several things. Was he blind on side effects? Not at all. We're speaking of it all the time.
01:00:20
He was watching the thing he attacked a story that was centuries long of history. He just decided okay, we need to get rid of that disease. It was not fabricated in a lab, it was not. We are going to save the world thanks to vaccines, etc. It was one particular disease he decided to get rid of. Was there corruption? He was still complaining. It's all led by finance, etc. I'm trying to do medicine for the day. He was basically the what do you know? The, the director de la sante, so head of health, basically the, the, the position of fauci. He was having that in france in the late 50s, early 60s, pre the goal, pre madastron, then pre the goal. This kind of person, the same guy who was hiding resistance in his sanatoriums during the war, just was talking about it's interesting.
Sebastian David Lees
01:01:18
I used to get quite angry about corruption and kind of grifters and things like that and a friend told me one interesting thing that you have to remember is that every ecosystem, every interesting ecosystem, will always have a parasitic layer. And if you are operating in an environment where there is corruption or drifting or even the mafia, things like that, you know you're operating in an interesting ecosystem, at least.
Raphael Douady
01:01:47
Yeah, that's what I was saying. You see, you always have some of it in here. Now, where it becomes more difficult is the following it's the amount. There's something. It becomes dangerous the day.
01:02:01
I mean the fact that, you see, a business does everything to develop its business. That's part of the game, and so the role of the state here is to control the appetite of the business. The fact that business people are predators if they are not, there's no business. So you are killing the business. The fact that business people are predators if they are not, there is no business. So you are killing the business. If you try to, everybody will be sheep, etc. Then there will be no business you need to have. The world is made of predators and prey, etc. That's laundering. But the role of the state is precisely so that we don't live in the law of the jungle. Because the law of the jungle, because the law of the jungle has two ends Either one of the super predator eats everyone, or the super predator is eaten by a colony of ants or viruses or bacteria, whatever you want, and then the story restarts.
Sebastian David Lees
01:02:55
Again, go back to the cycle.
Raphael Douady
01:02:57
Exactly those cycles. So when you see businesses, you know you will always have that and the role of the state is really to impose regulation If it becomes the interest of the business to buy the regulator. And it has a name it's called regulatory capture. I've worked on that. We're talking of laybacks. I've worked on that, the regulatory capture a lot in terms of in the banking industry. I know exactly how it works. I've seen it, I've described it. I've proposed solutions to control regulatory capture, et cetera. To control regulatory capture, et cetera, where they apply now, because precisely, there was too many interests to maintain this regulatory capture. And regulatory capture is like drugs for the business, it's heroin for the business. It makes it cool, the business is now nice, beautiful, et cetera, and nobody is ahead of you except that you're heading directly to the wall. So that's why I was, you know, when I saw the COVID thing, I saw the exact same mechanism of regulatory capture that I was working on for the financial industry. I saw that in the pharmacy industry. Did that exist at my grandfather's time? No, there was corruption, there was pressure, there was, etc. But it was not at that scale. It was not at the scale of regulatory capture. Regulatory capture means that, precisely, you get people like Fauci, who are now the actors of the corruption. They are not trying to resist the pressure, they become the actors the corruption. They are not trying to resist pressure, they become the actors. So those people who are in place, who are supposed to be, etc. They become.
01:04:44
The same story led to the 2008 crisis, actually to the 2007,. To the subprime crisis, and then they tried to blame the quants, etc. It was a banking story. They were giving loans to anyone because it was practical, it was just crazy, etc. And then and they were also, you know, banks were selling all these credit derivatives that were, you know, repackaging completely crazy loans into acceptable products, etc. So so it was a whole thing that you knew was a dead end, and we saw the end in 2008.
Sebastian David Lees
01:05:22
And do you see and I probably will have to make this the last question I wish I could spend hours with you. But do you see, are we reaching a tipping point again? I mean, we're seeing huge, enormous asset inflation over the last five or 10 years. We're kind of coming out of the ZIRP phenomenon hangover. We have a whole generation of property owners, we have a whole generation of finance professionals who have never known anything but ZIRP. Do you think we're heading for another shift, state change, crash, call it what you want, or maybe more of a slow deflation or business as usual?
Raphael Douady
01:06:01
First of all, I don't see any solution to the debt. I like finding solutions to great problems like this, unless, you know, I would say a massive market crash is the least damaging solution I see to the debt, because I see three exits to this debt problem. One is a big market crash and redefinition like a 29. Second possibility is the money becoming meaningless, so a total slow drift to systematically bailing out and saving, so printing etc. So they're becoming basically meaningless. And that is a slow drift from capital economy to Soviet economy. In capital economy, shelves are full but people don't have money to buy items on the shelves. In Soviet economy, people have plenty of money but shelves are empty and they can't do anything with their money. So the supply chain is broken because there is no incentive. So we see a small drift, a slow drift toward that.
01:07:14
And the third one, which is much, much more dramatic, would be a world war, you know redefining, because obviously all the East I mean whether it's the BRICS, russia, china, you know are going back to really manufacturing, producing commodities and manufactured products that are needed by the West. So we are no longer in the true data economy, pure virtual economy, and so they will want to fence their business. It's not only the US who are going to fence, they're going to be fenced on both sides, and so either it goes to a Cold War or it can go to a hot war, and that is. You know, there is pressure, surprisingly, on this side. You know where people want to go to a hot war because they feel like it's not sustainable, and so kind of a division, a re-division of the world, would be one of the least dramatic outcome.
01:08:21
So now, talking markets, the kind of direction in which I'm advising my clients who are asking what I'm long I'm trying to put in place strategies that preserve or even increase their wealth in two situations a slow increase or a big crash. But a sudden rise of the S&P by 20% in less than one month, I think it's unlikely, it may happen, but I don't think it's likely. And a slow decrease, a stagnation, possibly a slow decrease is something that I mean. I'm not saying the probability is zero, I'm just saying that in my designer strategy that would be the losing thing.
Sebastian David Lees
01:09:06
Wow, very, very sobering words to finish with, but important, and I think it'll be interesting to have listeners dwell on that final few minutes and really show some thoughts, and maybe we can have a talk on the Fat Tony Surfer about that at a later date. I must end it here. We've already gone over time. I've tried to keep you as long as possible, but you know, rafael Duarte, thank you so so much. Very final thoughts. Do you have anything you want to give a shout out to or a plug to? If if people are interested in the Primers course, is there a website?
Raphael Douady
01:09:38
Well, yeah, this Primers lecture that I'm giving now about three times a year, that actually initially were a spin-off from the RWRI, from the Rail World Risk stuff with Nassim and Robert. Now I'm doing that, you know it's more. It came from a demand from the, from the participants, to go a bit deeper in the calculation. So very practical. So there are three series on the practical probability. I call it processes, where we understand, you know, the dynamics, random dynamics. This one is more towards trading and the third one is purely on AI, with my friend, Patrick Skinner. So we are Patrice is really an engineer, so we are a good balance to understand, really opening the trunk and looking at how the engine works inside. Probably the next session in February or March, February, probably end of February.
Sebastian David Lees
01:10:42
We'll include links to you know, obviously, Raphael's social profiles, but websites and all that kind of stuff if people are interested. Whatever platform you're listening on, the links should be below when you're listening to this. So thank you so so much, Raphael. It's been an absolute pleasure. I really, really enjoyed this conversation.