Episode 30: R. Martin Chavez - Vice Chairman and Partner at Sixth Street

0:00This week on Generating Alpha, I'm joined by none other than Marty Chavez, Vice Chairman of Sixth Street, board member at Alphabet and one of the most influential figures at the intersection of finance and technology. Trained as both a computer scientist and physician, Marty's unconventional background has fueled a career defined by innovation, leadership, and a deep understanding of how data and software reshape industries. After pioneering machine learning applications in the early 90s, Marty spent nearly two decades at Goldman Sachs, where he helped build and lead the firm's most critical trading and risk infrastructure.

0:32He ultimately served as CIO, CFO, and co-head of the securities division, modernizing one of the Wall Street's most powerful institutions. In this conversation, Marty reflects on his journey from a tight-knit, education-focused upbringing in New Mexico to his current work at the forefront of markets. He shares how being a gay Latino outsider shaped his drive, the role of communication in his success, and how he sees AI transforming, not replacing human work in finance. I really enjoyed making this episode, and I hope you guys enjoyed listening. If you enjoyed the episode, I urge you to subscribe to my YouTube channel, follow me on Spotify, give it a five out of five stars, and share it to family and friends who might be interested.

1:11Thank you. Thank you for joining me, Marty. I really appreciate it. Happy to be here. Thanks for inviting me. I'd like to start where I always do. Let's take a step back and focus on your childhood. So tell me a little bit about your childhood and upbringing, and how do you think the kind of environment you grew up in shaped the person you are today? so I'm from Albuquerque New Mexico I'm from a family of five I'm the oldest New Mexico is an interesting part of the Hispanic world in that it mostly got it got settled in the 1500s and mostly got left alone for hundreds of years. And it's Spanish, it's Mexican, it's been a part of the U.S.

2:05since 1848, but it was settled a lot before then. And some of those original settlers who came up north in the 1580s are my direct ancestors, which is fun. I speak a version of Spanish that I thought was broken Spanish. It just turns out to be Spanish from the time of Cervantes. So it's a little bit of a backwater in some ways, a beautiful place, also heavily influenced by the National Weapons Laboratories. So famously, Los Alamos Labs is there, not so far away from where I grew up. And another one of the National Labs, Sandia Labs, that does the ordinance engineering, the material science for the arsenal.

2:59That is there. That's where my dad worked and my mom worked. Basically, all of my upbringing. I grew up in a very typical, I can see in retrospect, barely middle class home where education was everything. So the central myth, which may or may not be true, the central myth of the family is that my mother, when she was 10 years old, made a vow. Either she would be a nun or she would have 10 children and send them all to Harvard. why a girl from the the the barrio of albuquerque new mexico would get the ambition to have a lot of kids sitting to harvard is a bit mysterious to everybody uh as it happens there are five kids in my family as i mentioned i'm the oldest and we did all go to harvard and that generated a little news back when my baby sister graduated some years ago one of my professors said why there's a lot of you chavases coming through this place i've lost count how many is it and i said there's five and he said there's got to be some kind of a record do you mind if i do a little digging and it was some kind of record i'm not i don't know the exact statement but there if there have been some other families had sent a bunch of kids to Harvard.

4:32It's a small number. And the university did a little PR. And then the Today Show asked if they could film from my sister's graduation and have the whole family there. And the university gave my parents a certificate that looks like a diploma for getting all five kids through Harvard. And then the story really blew up because we were on the Today Show. And I got asked right there on the Today Show, wow, it sounded like you had a really hardcore childhood. And did you ever think that your parents were just mean? That was the exact question. And I said, it was intense. And the way I think about it now is no two people ever did more for their kids.

5:24So my parents really scraped to send us to a great prep school. We were always conscious of being just one step out of the barrio. And my mother insisted that we all have white collar jobs starting from when we were 16. And in New Mexico, your main choices were flipping burgers for the tourism business or working at the National Weapons Labs. So my very first summer job was writing FORTRAN programs to simulate the scattering of content electrons from neutron bombs. I came to be 16 just at the time when the government had signed the nuclear test ban treaty and wanted to explode atomic bombs in software simulation rather than actually exploding them in the atmosphere.

6:17And so writing programs on supercomputers to simulate those explosions was a thing. And it was a thing that I could work on. I've been writing computer programs since I was 10. I was really good at math. My parents sent me to the University of New Mexico to take math classes after my regular school classes. I discovered the Computing Center. I fell in love with computing. I knew some physics. and that started my career. So I got a very early exposure to stochastic partial differential equations. And I mentioned that because the math used to simulate neutron bombs is the same math used to price financial derivatives and to manage risk.

7:06And so I didn't know it at a time, but it was a perfect setup for a career on wall street and uh i also had the other skill that i developed that i didn't know wall street would want because wall street didn't want it in 1975 when i started building it at the age of 11 was the skill set of building digital twins of some business or scientific reality and the value of a digital twin is you can blow up the digital twin and so what it's not so bad terrible happened in reality and then you can inspect the what's left over after your digital twin blows up and ask what went wrong and run it backwards in time and ask what could we have done differently and this capability is really crucial in finance and i built a career out of it why do you think you were drawn to computer science what did you like about it.

8:03So initially, what I liked about it was that I grew up in a very strict Roman Catholic family. And the freedom that I found at the age of 11 was, if I went to the University of New Mexico Computing Center, and my parents knew that I was there, I could spend as much time there as I wanted. So from a very early age, computers and computing to me represented freedom to do my own thing. And that's how I fell in love with it. And to this day, I think of writing software as meditation. It's my favorite thing to do. I can work on a complex piece of software for 16 hours straight and not really know that any time has passed.

9:01And you obviously grew up in a household, like you mentioned, that emphasized education very heavily. How do you personally think about education? What's your perspective on it? What has it been and what is it today? So I was lucky in a very strange way. So So in the 1930s, if you were unfortunate enough to get tuberculosis, there was no treatment. So people who got TB back then were told your only hope is to live in a very dry climate. So there was a very wealthy New Yorker who got a TB diagnosis, was told he needed to live in a dry climate. He chose New Mexico, very dry, went out there and saw all this empty land and bought it, a million acres, something like that, for a penny an acre.

9:58And then he left it all to a nascent, non-denominational, boys-only at the time, prep school called Albuquerque Academy. So this school had an endowment that the big East Coast schools were always trying to get their hands on by merging with my school and making it their Western campus, right? So in Albuquerque, New Mexico, which didn't have a great public education system, there was this amazing school that belonged in a place more like Boston or New York or L.A., except it was in Albuquerque. And my parents just thought the kids are going to go there because they're doing backwards induction.

10:47If the kids are going to go to Harvard, they should go to the best school before Harvard that they could possibly go to. And that's Albuquerque Academy now. Strange dynamic. In a state that was 75 % Hispanic, the Hispanics were the underclass. And in this whole prep school, there were like three Hispanic students. Me and then the children of some of the teachers in the school. and that was it. It really was another time. It wasn't really designed for Hispanic kids and my parents didn't care and through a combination of scrapping and being very scrappy and also financial aid, I went to the school and it turned out I was really good at math and my parents really encouraged it and I saw math as the way out of our situation the way to climb the ladder the way to advance my prospects in the world to use an old-fashioned phrase and and I got a lot of validation doing that and so it became a cycle that fed on itself and then after that I went to Harvard.

12:11And then I thought, well, you're going to go to Harvard and you want to go to get a PhD in computer science. What's the best place for that? And that was Stanford. So I did that. So I've just always had in my mind the value of an education. I saw it in action. I wanted it to be a liberal education. I didn't just want to be good at math and science. I wanted to know other things too. And I want to pay it forward. So most of my philanthropy is about education and health and the arts. And I've got some children and two children. One of them is here in my office as we're talking. He's reading a book.

12:52And I want to do with my children my version, given my resources of what my parents did with me. So I'm always thinking, well, what's the next level of a great education? And I'm doing that. So yeah, it's been central to everything in my life. That's incredible. I'd like to fast forward a little bit. So you go off to school at Harvard, then Stanford, and you end up working on machine learning in the early 90s with a very interesting group of people, one of which is Reed Hastings. The other is the now CSO at Microsoft. What lessons did you take away from kind of the work itself, but then also the people who you worked with at that time?

13:38So I was an early fanboy of artificial intelligence. So, of course, I was going to get a PhD in AI.

13:52unfortunately the universe doesn't necessarily give you great signals on timing at least not for me and so I went I did the MD PhD program at at Stanford dropped out of the MD MD program to start a company which maybe we could talk about maybe not but did finish the PhD in artificial intelligence applied to medicine. And strangely, one of the pioneers of AI in those days, Ed Shortliff, Ted Shortliff, Edward, but everybody called him Ted, had written a program called Mycin that diagnosed blood bacterial infections. And it was an early expert system. And so he formed this program of medicine and computer science.

14:44And I was always interested in medicine. My undergraduate, my bachelor's degree is actually in biochemical sciences and microbiology and a master's from Harvard and MIT in computer science. So I've always been combining those two interests. So Stanford was the obvious place to go. This was the obvious program. And what an interesting collection of people. You mentioned two of them, But there were many, many, many others. And of course, we didn't know it at the time. We all thought Reed was super smart. And so when he decided to use these AI techniques, which were not successful or powerful enough to do a lot in medicine, certainly not compared to our ambitions, but he figured they were enough to help figure out what movie to watch tonight.

15:35And so all credit to Reed for finding the right early application for that time. We were early on medicine and medicine was very hard. We at that time and still today, but especially at that time, there were two areas of research in AI. One I'll call the symbolic approach and the other I'll call the connectionist or neural network approach. I was, because of the Stanford connection and Ted Shortliff, I was very much in the symbolic camp. We were building expert systems. And to us, it seemed obvious that we would want to start with human experts and somehow in software recapitulate what they were doing.

16:22And we knew that was going to be hard. And the systems we built were really brittle and didn't scale. But then at the same time, there were these other people, honestly, we thought they were nuts, who said, well, we're just going to start at the bottom, not with human experts, but with a neuron. And we're going to write a little program that simulates what a neuron does. It takes a bunch of inputs and then based on some formula on the inputs, it either fires or doesn't fire. And then it becomes that firing or the absence of firing becomes an input for another neuron. We're going to start at the bottom.

16:59And we thought that seems completely bonkers. Like that's never going to work. So it goes to show you what did we know, right? We turned out to be wrong. And now I would like to say I wasn't militant on this topic. I was part of the Stanford crew and we were doing what we were doing. And I always thought the connectionists were super interesting, even though they were the other crowd. And I was skeptical. But then I also thought exponential processes are very hard for human beings to grok and maybe enough generations of Moore's Law and they'll be on to something. Well, 30 years of Moore's Law, doubling every 18 months, takes you to some amazing places.

17:47It takes you to the proverbial second half of the chessboard. And that line of research proved scalable in a massive way. Now, the scalability, I would say, is something that a few people, some of the open AI and Google people, conned on to maybe five years ago. But it's really only in the last three years since the chat GPT moment of late 2022 that the fruitfulness of scaling is now obvious to the whole planet. it and there are also some limitations to it hallucinations and other things it's still not clear there's still not a consensus are these things reasoning are they forming symbols does it matter is just enough compute all you need do you need to go revisit some of those symbolic techniques back from the 90s and maybe fuse them with the connectionist techniques.

18:55So this is almost a piggybacking on the Kahneman and Tversky idea of two different processing systems in the human brain, the one that's fast and instinctual and the one that's slow and reasoning, right? And the one that's slow and reasoning is built virtually on top of the lower level hardware. Maybe we want to do something like that with these AI systems, or maybe more compute is all you need. That's the so-called bitter lesson of AI. Don't write any software because that's brittle, and it will just be blown away by the next generation of the scaling of the model. So it is the most exciting time to be a computer scientist since I've been alive, and I've been alive for a good chunk of the history of computer science.

19:46So I don't know, maybe it was more exciting between John von Neumann and when I was born. I just wasn't around for it. But now is an incredible golden age. And I'd love to talk about a bit more about kind of computer science a little bit later. But you're at Stanford. You get a letter in the mail from a headhunter from Goldman Sachs, New York. And you take a flight out there for an interview. Tell us how you ended up at Goldman. And if you can kind of give us a brief summary of your time there, any important turning points or kind of events that pushed you all the way to the top because you were eventually chief information officer, chief financial officer, and co-head of markets.

20:30So I once asked an extremely successful friend of mine, what is the secret of your success? And he said, oh, that's easy. I was in the right place at the right time. I thought, well, that's not actionable. And then he added with the right preparation. The bad news is that, of course, you don't control the place and the time. The good news is you totally control the preparation. The bad news inside the good news is that you only think you know what you're preparing for. The universe doesn't reveal that to you until until later. Right. And so but you do totally control the prep. And and I will assert that your fate is over time largely a consequence of the work and preparation you put into it.

21:29Of course, their initial conditions. But generally, my observation is by the time you're in the 30s, if you're fortunate enough to be born in this country, your own choices are a huge determinant as well of your of your outcomes. So why do I say all that? Well, I was preparing myself for the future of medicine and AI. My timing was terrible. Or maybe not, right? I missed one of my favorite formation stories, which is I showed up at Harvard at the age of 17. And I took sophomore standing and I had to declare a major. and I hadn't done any due diligence on Harvard and I didn't realize they didn't have a computer science major.

22:19So I went to the science center to go find a science to major in and a scientist is sitting opposite a table from me and he says to me, what are you? And I said, I'm a computer scientist, like the Harry Potter sorting hat, right? And then he said, the future of biology is computational. Now, that's an iracular thing to say to a 17-year-old in 1981, and not at all obvious. That professor, Stephen Harrison, was one of the forces behind the protein databank. And he said, in my lab, we are crystallizing proteins, and we're going to put the location of each atom in a database somewhere. And after we crystallize one protein, we're going to crystallize another protein.

23:08And someday someone's going to figure out what to do with all that data. And it might take 50 years before we solve the protein folding problem. So it actually happened to me. And of course, you know, the rest of that story, right? Demis Asabas and team came around and used that database that I was a minuscule contributor to as a kid and eventually solved the protein folding problem. all that came later none of that was helpful to me in 1993 in 1993 the reality of ai is that it hit a wall and all the businesses there was a wave of ai businesses they all failed they all got bought for scrap and here i was with a freshly minted phd in artificial intelligence in a time when artificial intelligence had such a bad rap that you could not get me to say the words artificial intelligence in public.

24:10I would say machine learning or I'd just say computer science because it was easier, right? And I was broke and had no idea what to do next. and one day I went to my box at Stanford to Stanford and I looked at my mailbox and there was a FedEx package in it and it didn't feel too special because there was a clearly identical FedEx envelope in the other students boxes too and I remember others got them and said what is this and threw it away and I remember it was from a headhunter I still remember his name He said, my name is Jory Marino. I don't know what Jory's up to. And I'm a headhunter for Goldman Sachs.

24:55And I've been instructed to make a list of entrepreneurs in Silicon Valley with PhDs from Stanford in computer science. And you are on my list. Would you come out to an interview? and I thought it's a joke. I seriously, I put no thought into it at all. I thought I'm going to scam these people for a free trip to New York and I'm going to see my college roommates who are living in New York and will have fun at the weekend and then I'll go do my interview and then I'll go back to San Francisco and I'll figure something out. So I went and I did the interview and I show up at Goldman Sachs, which I knew nothing about.

25:34Now, remember that my Harvard graduating class of 85, a large number of people went to Wall Street, but they mostly all got washed out of Wall Street in the 87 stock market crash. But I was dissecting cadavers at Stanford Medical School, and I barely noticed the stock market crash. I didn't have any stocks. I was poor before, during, and after the stock market crash. It didn't affect me at all. And I'm suddenly getting this offer from Goldman Sachs. I was only dimly aware of it. I showed it to some Silicon Valley friends who were in the know about Wall Street. And they said, look, if you're going to go to Wall Street, M &A is where all the action is.

26:21This is the commodity trading business. That's a dusty, dark corner of Wall Street. Doesn't sound promising at all. By the way, Goldman just bought that trading business and the people in that trading business aren't even allowed to use the same elevators as the real Goldman bankers. And so I went out and I arrive at the J. Aaron trading floor. And my first observation was this place is a dump. The carpet was dirty. The furniture was ancient. I guess what I should have thought is these people aren't spending any money on office decor. because it's a partnership and that would be money coming out of their own pocket.

27:05And why would you do that? And I walk in and they say, what do you want, the math quiz or the computer science quiz? And I remember thinking, I have a PhD from Stanford in computer science. So, but I didn't say that. I said, I'll take both your quizzes. And for them, this was a dream. This is what they were looking for. I didn't realize that that was the right answer to the question of which quiz. And I took both quizzes. And the next thing I knew, there's suddenly a lot of attention and energy. And then after that, I'm in a room and someone's telling me, we're not letting you go back to San Francisco until you say yes.

27:50Then I thought, have I been kidnapped or what's going on? and they put an offer in front of me, which I didn't realize at the time was an unusual setup. Supposedly at Goldman, the average number of interviews is 42 before you get a job offer, but it all just got accelerated because we had arrived at a moment in time for which I was totally prepared. But the universe didn't tell me that I was preparing for a career at Goldman Sachs, But we arrived at a moment in time where there was a business, the Wall Street business, that had a huge need for the exact skill set that I had of modeling a complicated problem domain using math and software, creating a digital twin.

28:44And so when they found somebody who knew how to do that, they weren't going to let go. And they put an offer in front of me, and it was, I mean, it was 10 times as much money as I was making at the time. And I thought, wow, these people are crazy and serious.

29:05and I've also heard that Wall Street is an incredibly homophobic place and I had come out a few years before and in Silicon Valley it was a non-issue to be gay and I wasn't going to go back in the closet so I told them I think you should know that I'm gay and apparently this had never happened in a Goldman interview up until me in 1993. And their answer was, do you have a boyfriend? Which I thought was a very strange answer, but lovely in a way. Not an answer you'd be allowed to give today, by the way. And I said, yes, he's a securities attorney. Well, we'll get him a job at Sullivan Cromwell, does our legal work for us, and you can come out together.

29:56And I thought, oh, this place is gay friendly. Wrong conclusion. It was, it was gay indifferent. And that was enough for me. In other words, my being gay was not going to get in the way of business, and they weren't going to let it get in the way of business, which is something I always appreciated about, about Goldman. I did join, and I discovered very quickly that I was the only out gay person in the firm, and one of a tiny handful of out gay people in all of wall street like 10 in the street if if that right so but i was good at math and software and uh in in difficult environments in the past i'd always leaned into math and software successfully i did it then still doing it now and it's uh That's a formula that works for me.

30:54And what was your time like at Goldman? How did you progress? So I started off as what Goldman called a strategist or strat for short. Yeah. I don't love the term strat, actually, even though I was at some point co-head of strats. I still hated the term. at some point it got shortened I like strategist because it was suitably vague like what is that well it's just a made-up term for what the rest of the world is now standardized as data scientist we didn't have that term in 1993 and I was originally hired to work on a core piece of software that was given the unglamorous name secdb securities database and i remember on my first day i was told hey here's your new office there's three guys in an office off the corner of the trading floor their pencils stuck in the ceiling because whenever they got into a difficult mathematical or software problem their meditation was to throw pencils at the ceiling and they'd stick there.

32:11And they played Pink Floyd pretty loudly. And I thought, okay, this is like, this kind of like Silicon Valley, I can deal with this. And then they said, your first job is to write an object oriented transactionally protected database in C. And I remember thinking, that's idiotic. Why don't you just use Sybase or Oracle? And oops, I forgot to take the database class at Harvard. I don't really know anything about databases. Well, I don't recommend this, but, and I wish I'd taken that database class, but given that I didn't, I didn't know how hard and silly this was. So I just jumped in and worked on it and did it.

32:57And we had a philosophy back then. The only thing crazier than writing all your own software is not writing all your own software. And that strategy worked very well at Goldman. And this software enabled us to create a digital twin initially of the foreign exchange trading business so that we could lose money in software simulation without losing it in reality. And this gave us an edge on the other trading businesses. And one day the boss comes into my office and he says, the oil strat has resigned. Congratulations, you're the new oil strat. Go out to the trading desk and introduce yourself. There was no way he was going to take me out there and introduce me because it was an extremely unfriendly environment.