Finding a Single Source of AI Truth With Marty Chavez From Sixth Street
35:51 Share

I cannot believe he said this to me in one thousand nine hundred eighty one, but he said, the future of the life sciences is computational. The through art is my entire career. I've been building digital twins of some financial or scientific or industrial reality.

We looked at that and thought, wow, we Better do something about this very large unhedged position. That was the history of dod, Frank. Like, we don't really know what went wrong in the financial crisis.

So let's just go regulate everything. And I think ninety nine percent of IT was red tape that did not make the world a Better place. This was one of the many early nuclear winters here I I walked right into.

Hello everyone, welcome back to the asic T. C. podcast. This is your host, steffen. Now today we have a very special episode from a new series called in the volt.

This series features some of the most influential voices across the finance ecosystem, including, of course, our guest today, marty chavez. Marty is now the partner and vice chairman of six, three partners. However, his long had a ng for spotting how a healthy serving of technology can disrupt other industries.

From his P H D of applied artificial intelligence to medicine, to being one of the founding engineers of the team that created sec p that's the software that perhaps couldn't predict the global financial crisis, but famously help goldmine survive IT. So today, Martin sits down with asic sense e general partner, David abr, and they talk about a lot more, including where the puck is moving in this new wave of technology and the role of regulators and lawmakers within that. And of course, if you like this episode, don't forget to check out our new series in the volt. You can find that on our asia S E live view, which were also include in the showers. There you can find other episodes with global payment C E O jeff sloan and marco or gene, the cio of golden sex are you do IT take IT away.

Hello, and welcome to in the volt asic scenes fin tech podcast series, where we sit down with the most influential leaders in financial services in these conversations, we ever behind the scenes view of how these leaders guide and manage some of the country's most consequential companies. We also dive into the key trends impacting the industry and of course, discuss how A I will shape the future today.

We're excited to have marty chavez in the show. Mart is currently a partner and vice chairman of six street partners, a global investment firm with more than seventy five billion in assets under management. Prior to six treat, marty spent over her two decades of golden sex, where he held a variety of senior roles, including chief information officer, chief financial officer, head of global markets, answered as a senior partner on the firms management committee.

He was also one of the founding engineers behind the legende software system, sec db, which many believe helped go women avoid the worst of the global financial Prices. In our conversation, marty talks through the evolution of technology in financial services and the potential impact of artificial intelligence. Let's get started.

As a reminder, the content here is for informational purposes only, should not be taken as legal, business, tax or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a extensive fund. For more details, please see a extensive 点 com lash disclosure also。

Marty, thank you so much for being here. We really appreciate .

David day's pleasure. Been looking forward to this.

Martin. We had a fascinating career, obviously played a really pivotal role in turning the wall street trading business into a software business, especially turning your timing coben sacks and also now at six street. But you also serve on the boards of the broad institute on stand for medicine and a bunch of amazing companies. Maybe walk us through your career. R and what is sort of a three line in those experiences?

Well, let me talk about a few of things I did. And then the ark will become apparent. So I grew up an outcry king, new mexico.

I had a moment, really like the movie, to graduate when I was about ten, and my father put his arms around my shoulder and said, medicine, computers are the future, and you will be really good at computers. And this is nineteen seventy four. And IT was maybe not obvious to everybody, was obvious to my father.

He was a technical illustrator, one of the national abortion orties, and there was a huge computer that they just bought, that his team used to draw these beautiful blueprints for the weapons in the nuclear arsenal. And I really had the latest and greatest equipment when I was very conking, very expensive, and my dad, newer, was going so in new mexico, you don't have time and choices, especially that time is basically tourism and the military industrial complex. And so I went for the military industrial complex.

And my very first summer job, when I was sixteen, was at the air force weapons lab in alberni. The government had decided that blowing up bombs in nevada desert was really problematic in a lot of ways. And some scientists had this idea crazy at the time that we could simula the explosion of bombs rather than actually detonating them.

And they had one of the early crayon supercomputers, and so are a little computer deep kid. This is an amazing opportunity in my very first job was working on these big four train programs that would use many cars. simulation.

I got an early baptism in that technique, and you would simulate individual company electron's being scattered out of a neutron bomb explosion, and then calculate the ox magic post that arose from all that scattering. And my job was to convert this program from M, K, S units to electron rest mass units. That certainly seemed more interesting to me than jobs in the tourism business.

And so I did that. And then the next big moment was I went to harvard, was a kid, and I took a soft more standing. And did you buy any chance? Did you do software standing?

I didn't do soft standing. I also went to her. I think you, I think we also think about you study about chemistry and said so.

so you have to declare a major a concentration right away. You takes off more standing. And and I didn't know that, and I didn't know what majors going to declare is going to be some kinds science for sure.

And I went to the science center, and the science professors were recruiting for their departments. And I remember Stevenson sitting opposite table saying, what are you? And but a little bit like a hog guard's question, I suppose.

And I said, i'm a computer scientist and I cannot believe he said this to me in one thousand nine hundred eighty one, but he said, the future of the life sciences is computational and and that is amazing, right? And so provide and so prescient and I thought, wow, this must be true and he said, that will construct by ocampo jor just for you, and will emphasize simulation, emphasize building digital twins of living systems. And so I walked right into his lab, which was doing some of the early work on X A Christopher phy of protein capes and working to set up the protein data bank.

And who knew that? Will even back then, he wanted to solve the protein folding problem. And I remember you said, I might take fifty years, and I take a hundred years, and we might never figure IT out.

And that's obviously really important because that protein data bank was the road data for for, although old, which later came in and solve the problem. And so the through art is my entire career, i've been building digital twins of some financial or scientific or industrial reality. And the amazing thing about a digital twin is you can do all kinds of experiments, and you can ask all kinds of questions that would be dangerous or impossible to ask your perform in reality.

And then you can change your actions based on the answers to those questions. And so for wall street, if you got a high finality model of your trading business, which was something that I, with many other people, worked on as part of a huge team that made C, D, B happen, then you could take that model, and you could ask all kinds of counterfactual, what if questions. And as the CEO of golden sex, loy black fine, who really commissioned and sponsor this work for decades, would say, we are not predicting the future. We are excEllent predicted of the present. And i've been doing some variation of that ever sense, if fascinating.

I I don't want to spend more time kind of digging into sector b because I was also A A pression decision, obviously, ly during the financial crisis. But maybe just go going back and then you ended up doing some graduate work in healthcare in N N ai. How did you go from that into wall street? Maybe walk out of that transition because it's not poly obvious, maybe for most, and then would have to dig til your time. I got a bounder IT set.

I got so excited about these problems of building digital twins of biology that IT seemed obvious sto me, that continuing that in grad school was the right thing to do. I actually wanted to go and start making money. And I really owed to my mom, who convinced me that if I didn't get A P.

H. D, then I wasn't onna do IT. I'm sure he was right about that. And so I lied to stanford does my dream school. And so what happened is I was working on this program and artificial intelligence in medicine that had originated at stanford under ted short live, who was extremely well known, even back then, for building one of the first expert systems to diagnose blood bacteria infections.

And so I joined this program, and we and a bunch of my college, the program, took his work and thought, can we put this work, this expert system inference, in a formal asian, probably alister, framework? And the answers, you can. But the downside is it's computationally tractable.

So my PHD was finding fast randomized approximations to get provably nearly correct answers in a shorter period to type. This was amazing as a project to work on, but we realize pretty early on that the computers were way too slow to get anywhere close to the kinds of problems we wanted to solve. The actual problem of a diagnosis in general eternal medicine is you've got about a thousand disease categories and about ten thousand various clinical findings.

Laboratory findings are manifestations, are symptoms. And the joint probability distribution they have to calculate is therefore on the order of thousand to the ten thousand. And there is a big problem.

And we made some inroads, but it's clear the computers were just not fast enough, and we were all responded. And this was one of the many early nuclear winters of A, I. I walked right into a, and I stop saying artificial intelligence as embarrassed, right? Like this is not anything like artificial intelligence.

And a bunch of us were casting around, looking for for other things to do. And I didn't feel too special as I got a letter in my box at the department. And the letter was from a headhunter.

The golden sex had engaged. And I remember the letter. I prove IT somewhere.

I said, i've been asked to make a list of entrepreneurs and silicon valley with P. H. D. And computer science from stanford. And you are on my west.

And in nineteen ninety three, before linked in, you had to go do some digging to construct that's less. I thought I broke and I isn't going anywhere, anytime soon. And I have no idea what to do.

And I have a bunch of college friends in new york, and i'll scan this bank for free trip. And that's how I ended up in tax. And I didn't seem officious.

I just like the idea they were doing a project that seemed insane. The project was we're going to build a distributed transaction protected object, dorian database, that's going to contain our foreign change trading business, which is inherently a global business. So we can't trade out of exile, and we need somebody to write a database from scratch in sea.

And fortunately, I had not taking the database causes IT harder. Because if I had, I would have said, that's crazy. Why would you write a database from scratch? And I don't know anything about databases.

And so I just had the fortune to join as the fourth engineer and three person core C, D, B design team. And in a very lucky move, one day the boss comes into my office and said, the desk strategist for the commoditize business has resigned. Congratulations, you are the new commodity strategist and go out onto the trading desk and introduce yourself.

He was never going to introduce me to them. We were kind of scared of them, to be honest. And so there was in the middle of the oil trading desk, kind of an odd place for gay espana computer geek to be in nineteen ninety four. Was me.

It's such an amazing story. And you know, one of my favorite loans, which I believe, and I repeat often, is that opportunities live between fields of expertise. And I personally love expLoring those intersections. I feel like your career has heard been at these intersections may be fast forward coming into the financial crisis. Famously minor standing is that sec D B really helped the firm navigate that period and and really same what what was IT about sec T B, that was different than other wall street firms to, you know, lost billions of millions of dollars in that moment. And and how did you guys .

have to navigate them? Yes, this is where we're going to start to get into the pop culture heads because, of course, you have to mention the big short when you start talking about these things, right? And so C, D, B showed the regeneration CFO of golden sex during the financial crisis.

David venia that we and everybody else had a very large position in catalyzed dead obligation CS. They were rated trip away. So in sec, D, B, it's another thing.

And IT has a Price, and that Price can go up and down in their simulations where IT gets shocked according to probability distribution and then there's non parametric or scenario based shocks. And we looked at that and thought, wow, we Better do something about this very large unhedged position, namely so down or head IT. We didn't know that the financial crisis was coming.

Of course, we got in the press and elsewhere accused of all kinds of crazy things, like they were the only ones who hedge. So they must know that was coming. We were just predictions of the present and thought Better hedge this position, hence the big short.

And the question was, if lemon fails, what happens then? And we talk about lemon as if IT is a single thing. We had risk on the books to forty seven distinct lemon entities with complex subsidiary guarantee, non guarantee, clatter zed, non clatter ze relationships.

And so am a super complicated. But insect D. B. IT was all in there, and you could just slip IT around.

You could just as easily run the report from the counterpart side. Now I make IT sound like IT was perfect. IT was IT was a little less than perfect.

We had to write a lot of software that weekend. But the point is we had everything in one virtual place. And IT was a matter bringing IT together. So it's also part of the legend, but it's also factual.

We had our career show up at lemon s headquarters within an hour of its spilling bankrupcy protection for the forty seven entities, and we had forty seven sheet of paper with our clothes claim against each of those entities rolled up firm wide across all the businesses. And he took many of the major institutions on wall street months to do this. And so that was the power of C.

T. B. Of course, IT was wildly imperfect, but IT was something that nobody else had.

just like piggy. Back on that last point, what impact has regulation had historical on technologies, impact on financial services? And I think about, you know, the different asset classes, for example, in global markets that shifted to be traded electronically, right? Like was IT often historically driven by regulatory change, emerging technologies. Both are curious about that and also how IT informs the future.

yes. Well, so regulations, a powerful driver of change, and so is technological change. And some things are just inevitable. And a strong believer in capitalism with constraints and rules.

And we can, and we will have a vigorous debate about the nature of the rules and the depth of the rules and who writes the rules and how how they're implemented. All that matters hugely. But to say, oh, we don't need any rules or trust us, look after ourselves.

I just haven't seen that work very well. And so in some cases, the regulators will say something. For instance, in the dog Frank legislation, there's a very short paragraph that says that the federal reserve shell supervise a simulation was called the d fast simulation, the dot, Frank.

And I even remember what the rest stands for, right? And that will be part of the job of the federal reserve, a simulation of how banks will perform and a severely adverse scenario. And that was a powerful concept, right? You have to simulate the cash flow, the baLance sheet, the income statement several quarters forward in the future.

None of this was specified in detail in the statute, but then the regulators came in and really ran with IT and said, you will simulate nine quarters in the future, nine quarters in the future, right? The whole bank, all of IT the end to end. And then in a very important move, the acting supervisor for regular at the time that rule, the reserve governor said we're gonna a link that simulation to capital actions, whether you get to pay a dividend or whether you get to buy your shares back or whether you get to pay your people right.

Because he knew that that would get every boy's attention. If it's just a simulation, that's one thing. But if you need to do IT right before you can pay anybody, including your shareholders in your people, then you're going to put an awful lot of effort into its.

So that caused a massive change and made the system massively safer and sounder. We saw that in the pandemic. There's actually a powerful lesson for us in the early days of electronic trading, for the early days of artificial intelligence, right?

There was A A huge effort by the regulators to say we got to to understand what these algos are thinking because they could manipulate the market, they could smooth the market, they could crash the market. And we would always argue you're never gonna able to figure out or understand what they are thinking. That's a version of the halting problem.

But at the boundary between a computer doing some thinking and the real world, there is some A P, I, there's some boundary. And at the boundary, just like in the old days of railroad control at those junctions, you Better make sure the two trains can get on a collision track, right? And it's the junction where it's gna happen.

But then when the trains are just running on the track, just leave them running on the track, just make sure they're on the right track. That's going to be an important principle for l ms. And a is generally as they start agenting and causing change in the world, we have to care a lot about those boundaries.

And may is good transition to present day. You are a huge force in the zia of falling sax and Molly in general. And and kind of the rise, the developer, his decision maker, maybe talk a little bit about general AI specifically today.

How is this technology different from, you know, the AI of your PHD in one hundred and ninety one? And what are the impacts that you see not just in financial services but perhaps in other industries as well? Well.

for full disk, sure. I remember late eighties, early nineties and this program, stanford, we were, we were the asians, right? And then we would look at these connections ist to neural network people.

And I had to say, but IT is true with not sorry for them. We thought, like that all work, simulate neon. You got to be kidding.

And, well, so they just kept stimulating those neurons. And liquid happened. Now, in some ways, there's nothing new under the sun. I, at a fantastic talk not so long ago with Joshua, a benji, who's really one of the four, five women aries in this, in this, in this renison of A I that's that's delivering these incredible results.

And he was talking about how his work is based on taking those old asian decision networks and coupling them with with neural networks, where where the neural networks designed the beige networks, and vice versa. And so some of these ideas are coming back. But IT is safe to say that the threat of research, although the river of research that took this connection's nearly network approach, is the one that's sparing all the fruit right now.

And David way would describe all of those algorithms because they are just software, right? Everything is turning equivalent, right? But they are very interesting software. They, they started off with images, images of cats on the internet. People are putting up pictures of cats.

Well, now you ve got billions of images that people have labelled as saying this image contains a cat, and you can assume all the other images don't contain a cat, and you can train the network to see whether a cat or not. And then all the versions of that, how old is this cat? Is this cat ill? What ills does IT have? Like all of these things, over the last, maybe starting ten years ago, you started to see amazing results.

And then after the transformer paper, now we've got another version of IT, which is fill in the blank, or predict what comes next or predict what came before and is IT the transformers and all the chat box that we have right now. It's amazing. I wish we all understood in more detail how they do the things that they do.

And we're starting to understand IT IT all depends on the training set and also depends crucially on on a stationary distribution, right? So the reason all this works on is that a cat or not a cat, is the cats change very slowly and evolutionary time. They don't change from day to day, but things that change from day to day, such as markets, it's a lot less clear how these techniques are going to be powerful.

But here they are, are doing amazing things. We're using this in my firm and we're using IT in production, and we are deeply aware of all the risks. We have a lot of policies around IT.

IT reminds me a lot of the early wild west days of of electronic trading where we're authorizing a few of us to do some R N D. But very careful about what we put into production. And we're starting with the easy thing.

if feels like a unique moment or maybe there's unique to me, a lot of momentum happening. Go bottoms up and top down, bottoms up because no, I don't know something like forty percent of fortune one hundred is using maybe get up copilot union organization or microsoft dye product.

And in commercially, every CEO or every board member right can plug a prompt into one of these models and kind of understand totally the magic and imagine the impact that he could have on their business. And so IT seems like the employees of many these companies want the productivity gains that you're describing. Borders are like, you know, how is this an impact the human capital efficiency of our company? Like where can we deploy technology? I guess, when other ceos of large companies you come to you for your advice, like how are you advising them? I had to deploy A I in their organizations where within those companies like what's the opportunity to see maybe in the near term, in the middle of a long term.

really first order of business. And this is something that that that we work on a golden. For a long time and then happy that we left golden in a place where it's going to be able to capitalize on gena. I really, really quickly, which is having a single source of truth for all the data across the enterprise. Uh time traveling, source of truth.

So what is true today and what did we know to be true at this at the close of business on some day three years ago, right? And we have all of that and it's cleaned and it's created and it's named and we know that we can rely on IT because all this training of the eyes is still garbage, garbage out. And so if you don't have ground truth, then all you're going to do is spread about halcyons and you're just going to be caught in halcyons and imagining that are incorrect and not actionable.

And so getting your single source of truth right. That date engineering problem. I think a lot of companies have ve done a terrible job of IT. I'm really excited about the new gemini at one point five context window, a million tokens like that one. I just want to shot that from the mountain tops, like if you've been in this game and have been using rag retrieval augmented generation, which is powerful, but you run to this problem.

If I got to take a dog, a complicated dog that references pieces of itself and chunk, or you're gone to lose all of that unless you have a really big context window breaking that quadratic time complexity of the link context window is just monumental. And I think over the next few months, you're going to see a lot of those changes. Problems that were really hard are going to become really easy. I know. What do you think?

Look, I think every company needs a kind of using golden maybe as the analogy is so much of the organization, but particular even many parts, the federation, I think, can and should be leveraging software. And a lot of those workflows can be augmented with A I right from legal to compliance to end on boarding to you know risk management we're talking about. But I think it's going have a profound impact on the enterprise.

Obviously quite bias. I guess one topic that you know people debate quite off and is the impact of regulation on the adoption this technology. And just here is your view on the government's role in this in general. I and and what advice you have, income accelerating this verses? Yes, what responsibility they have.

Well, one of the things that I learned during the financial crisis was a huge amount of respect for the regulators and the lawmakers. They have a really tough job and and really important to to collaborate with them to become a trusted source of knowledge about how a business works.

And if I just meant the number of people who just go into a regulator and they're just talking their own book and and hoping that the regular or while maker won't understand that, I think that a terrible way to approach IT and has the very likely risk of just making them angry, right, which is definitely not the right outcome. And so i've been spending a lot of time with with regulators and legislators and a bunch of different and a bunch of different jurisdictions. And you you already heard a bit of a bit of what I have to say, which is let's please not take the approach that we first took without election onic trading.

That approach was write a big document about how your electronic trading algo works. And then step two was hand that document over to a control group, could then read the document and assert the correctness of the algoa. right? This is the halting problem. Squared IT is it's not just a bad idea. It's an impossible idea.

And instead, let's put a lot of emphasis, a lot of standards and attestations at all the places where there's a real world interface, especially where there's a real world interface to another computer, right? So the analogy is an electronic trading there was IT was not a lot you could do to prevent a trade from shouting into a phone, an order that would take your bank down, right? You could.

how? How are you going to prevent that from happening? right? And but what you really worried about was computers that were putting in millions of those trades, right? Even if they were very small, they could do IT very fast, and you could, you could cause terrible things to happen.

And so another thing i'm always telling the regulations is, please, please, the concept of liability, right? They start with this idea, let's make the lm creators liable for every bad thing that happens with L M. To me, that is the exact equivalent of saying, let's make microsoft liable for every bad thing that someone does on a windows computer, right? There's fully general.

And so there these elms are a lot like Operating systems. And so I think the regulation has to happen at these boundaries, at these intersections, at these control points first and then see where we go. And I would like to see some of these regulations in place sooner rather than later.

Unfortunately, the pattern of human history is we usually wait for something really bad to happen, and then go put in the clean up, regular after the fact, and generally overdo IT. That was the history of dod, Frank. Like we don't really know what went wrong in the financial crisis.

So let's just go regulate everything. And I think ninety nine percent of IT was red tape that did not make the world a Better place. And some of IT, such as the c car regulations, was profound and did make the system safer and sounder. And I would want us to do those things first. And not just the red tape.

but I know you're also very passionate about life sciences. You started your graduate career there and believe you now set on the board of recursion .

a pharma goals.

yes. yeah. Maybe talk through the implications that you're seeing four general vi in the sciences and biotech in particular.

Well, it's epic, isn't IT. I had an amazing moment just a couple months ago. I I had the opportunity of being the fireside chat post for jenson of invidia at the J.

P. Morgan n. Health care event. And there was a night that recursion was sponsoring. And we really talked about everything he learned from chip design.

So Johnson, incredibly modest, will say, well, he was just the first in that generation of chip designers who were the first to use software to design chips from scratch. And I was really the only way he knew how to design. And he he likes to say that in videos is a software company, which IT is right.

But that seems counter activity is supposed to be a hardware company. And he talks about the layers and lares of simulations. They go into his business.

Those layers do not go all the way to strate gas equation. And we can even do a good job on small molecules, right, solving to use equation for small molecules. But IT does go very low and IT goes very high to what algorithm is this chip running, and that's all software simulation.

And he said in that chat that at some point, hidden has to press the button that says, take this chip and fabricated, and the pressing of that button costs five hundred million dollars. And so you really want to have a lot of confidence in your simulations. Well, drugs have that favor very much so, except they cost a lot more than five hundred million dollars by the time they get their face three.

And so IT seems obvious to all of us that you ought to be able to do these kinds of simulations and find the drugs. Now the first step is gonna be just slightly improved. The probability of success of a face to a face free trial that's gonna incredibly valuable is, right now, so many than vail in the multibillion dollar of failures.

Res, but eventually, will we be able to just find the drug, the needle, in the high stack nature? This problem is mind blowing. There are, depending on the size of the carbon chain.

But let's just pick ze, there's about ten thousand trillion possible organic compounds and are four thousand approved drugs globally. That's a lot of zeros. And if A I can help us navigate that space, that's gonna huge.

But i'm gonna bet that we will met biology in this way. It's just, biology is so many orders magnitude more complicated than the most complicated chip, and we don't even know how many orders of making tude and how many where's of abstraction are in there. But the questions, do we have enough data so that we can train the L O, ms.

To infer the rest of biology? Or do we need an awful lot more data? And I think everybody's clear, we need more data. I think we're less clear on is do we need ten orders of magazine de more data or one hundred more orders of my tude? We just don't know mazing .

time to be a life.

That's time ever. We say this at the alphabet board and it's what an incredible group of people. And when I hear sargon, Larry says the best time never to be a other scientists know. Of course, I agree with that. It's magical, marine.

Thank you so much for your time. Always a pleasure you've had such a fascinating career, and we really appreciate you at spending .

time with great talking.

With thanks, i'd like to thank our gas for joining in the wall. You can hear all of our episodes by going to a sixteen z dot com back flash podcasts. To learn more about the latest in fin tech news, be sure to visit a sixteen z dot com backlash fin tech and subscribed to our monthly fin tech newsletter. Thanks for tuning in.