Marco Argenti: If you look at the software developer lifecycle, a lot of that is changing. Developers are or are developing software by developing specs today. And so if you are too much in the weeds down there in that mechanics, and you don't adapt for AI or agents doing that work in software development lifecycle deployments, rollbacks, monitoring, observability and all that. I think that path will be very much disrupted. Or if you're adding a sort of a UX on things, that's another class. So you have a very simple process. I don't know, you're doing surveys or expense reports or whatever. And now people are going to start expecting, you know, their personal assistants or agents to kind of do all that mechanic for them. And so I think that path is probably something that has a bigger question mark on top. And so I always ask myself the process question first. The process transformation section, the question first, and then consequently what's the tool that is going to support that? Just to press on this point though. Have you replaced any third party software providers with something that's been developed internally through AI? We have we have terminated contracts already. Yes, absolutely.

Tracy Alloway: Hello, and welcome to another episode of the All Thoughts podcast. I'm Tracy Alloway.

Joe Weisenthal: And I'm Joe Weisenthal. Joe. The thing about AI, I feel like it's accelerated all of our timelines, right? Like, it's phenomenal to me to think back that, like back to the days of ChatGPT. And when did that come out? 20 2020 2220. 2222? That's just crazy to think. And then really unbelievable the gap. And I've been thinking about this like just a the explosion of capabilities. Yeah. And the thing I've been thinking about is that, you know, after the first year or so and it came out, you know, we talked to executives and we were like, how are you using AI and your workflow? And everyone's experimenting. It's great. Everyone is using changing because I was very vague. And now in 2026, the so story is that AI is so powerful that it's going to destroy all these legacy software companies. So what I would say is we must be past the age of experimentation. I think that to. Really use cases. Using it better have some example of like, here is a workflow where we're using. Well, exactly. And to this point, now that we're past the age of experimentation, I'm very curious how executives and managers are actually evaluating the return on investment in AI and what they actually want to see from it at this point. So, you know, are you going to replace all your third party SAS contractors with internal coders, and what does that look like from an actual headcount perspective, from a cost savings perspective? We can actually get some concrete details on this now. So I'm very excited to say we do in fact have the perfect guest, someone who we had on before to talk generally about AI and someone at a company that has been doing, you know, they got into it pretty fast. The last time we spoke to this person was in 2024, and even since then, I guess it just feels like light years in AI time.

Tracy Alloway: So just one last thing on the last few years of AI, which is that when ChatGPT used it, when it when it came out, I played around with a lot, you know, and I have read poems and all this stuff. And then I bet if you actually looked at my AI usage, went for through a trough, whereas like, I wasn't really getting any productivity, there was nothing it really could do that I needed and still sort of seemed like a toy. So I had this like intense burst of use for the first several months and then this trough. And now these days, with the expansion of capabilities, particularly cloud code, I'm finding all kinds of new things. So there is like we're coming out of the trough. I think a lot of people are actually finding things, at least if I can generalize from my own experience. Yeah, absolutely. So we do, in fact, have the perfect guest. We've brought back Marco Argenti, who is, of course, the chief information officer over at Goldman Sachs, someone we had on the podcast back in August of 2024. So, Marco, thank you so much for coming back on today.

Marco Argenti: Thank you for having me.

Joe Weisenthal: How much have things changed for you? Does 2024 seem like 20 years ago now in AI time?

Marco Argenti: Yeah, I barely remember even what, what happened back there? That's a nice way of saying you forgot what we talked about on the podcast. Maybe that. But literally, like, things are really changing on a on a weekly basis almost right now. And, and if I look at the evolution, not only, since, a year ago, but events in six months ago, I think, it has been nothing short than, than revolutionary. A year ago, we barely talked about agents or the word almost didn't exist. We were using AI is like a chat companion. Yeah. There was, Yeah, it was telling you. Oh, I'm sorry, I don't know who's the president of the United States, because, you know, my cutoff date is like a year and a half before or things of that nature. Now you can say, hey, you know, as a as a person, you can say, hey, my plane, my plane just got canceled and is going to redo all your plans, is going to check for, you know, like available flights is going to do all these, capabilities of personal assistant, and, and that translates in corporations also in a lot of utility that you can see in everyday tasks. So, I would say, you know, what I like to say, to, you know, to my people, but also in general, let's say this is not the drill. This is real. You know, it's not the age of experimentation anymore. This is a tool that now can do a lot for you. And, and so we put it to work and we put it to working, you know, starting from developers. But on expanding in many, many other areas. So, I would say, actually, if I look at the increase of capabilities of these models that, what well, we've seen in the last six months or so with really the evolution of, this advanced reasoning capabilities, they came out, I think, that finally got us the confidence that, you can use, AI for every day's work with the right supervision. And also, you know, in many cases for mission critical applications, is not a toy anymore is something that, you know, you can expect results from. And I think that's the biggest change. So today I would say that, there is nobody that is not touched by, you know, we gave our, GSI or GSA assistant to 47,000 people. Most of them use it every day. More. Most of them use multiple times a day. And what's interesting is, it's like, you know, like the first time you see a tool like, know Microsoft Excel, you can almost not predict what people are going to do with that. Okay. Maybe it's born for, you know, doing some form of accounting and then people write entire applications on top of that or use it for project managers, what management or things of that nature. And AI is kind of turning, turning that way. If I look at what people do with that, it's, it's, it's really things that surprises every day because,

Tracy Alloway: But why don't you give us some examples? So in production right now, what are some workflows or novel things that were not workflows before the you see within Goldman that I was doing for people today.

Marco Argenti: So let's start from, the GSA assistant. Great. That can answer really complex questions based on external and internal data that generally before used to take sometimes hours or even days, sometimes weeks to answer. Okay. It can do very complex research for you in topics. There are, you know, for example, we can ask questions, that come from clients such as a hey, how does the recent geopolitical events on the Hormuz Strait actually, impact this portfolio? What could be a potential rebalancing strategy? Or, you could ask, intersection of, you know, like, I don't know, how does, a certain fed decision on interest rate actually impact, you know, the volatility of certain assets. So you ask this multi-dimensional questions and, what GSA system does, it calls out the model, retrieves, the relevant information and creates a plan to answer that question. And that's kind of the key, because this AI's, they really plan before responding rather than just giving you the first thing that comes to mind. And so that's what kind of at the very surface thinks that one of the most, you know, common use cases, which is, we really enhance the client experience by being able to answer questions internally and externally in a much, much faster way. But really complex question, not simple questions. We had to wire up hundreds of data sources and also most importantly, which is something that I tell everybody that asked me, hey, give me a, you know, some, some, some advice on how to implement the AI in a corporation. Data quality is really the determinant between good AI and not so good AI. And so we do a lot of work to not only take a bunch of data, but also making it, understandable to, to the AI. So for example, to better be deeper, we have a tool called legend, the AI, which is our lakehouse, which allows you to go from, query to MC server, connect to GSA assistant, i.e. from data to answers. You can wiretap literally in 2 or 3 clicks. And it does all of that for you. And so the quality of the data, the quantity of the data, not only that, because not just the, you know, the bitter lesson here, but is also the lesson of, you need to curate your data. You get better answer disproportionately. That's something that is driven there. So that is kind of the knowledge, aspect of AI, which is, I would say the most widespread because every single one in the firm has that. And it's, you know, the highest users. We are like way above a million prompts per month. And it's growing really, really, really fast. And then, of course, you know, you're asking me like really impact in in production, every developer in Goldman is enabled with a gigantic AI. Okay. So we were probably one of the first, if not the first to launch Darwin almost like a year ago, which is the fully gigantic, developer assistant. We have cloud code. We have, you know, many other tools, GitHub, copilot, agent, etc. but, on that, you really see the step change. There is no question that, that is changing the way developers work. And by the way, it's not just about doing the exact same things more efficiently. It's changing the way developers actually do their work. And that is very, very easy to see. How, you know, that kind of changes the paradigm of what a developer does. You're much more of a product manager. You're much more of a planner. You're much more of an idea. Generally, the most important thing for a developer today is to be able to explain things rather than jumping and coding things.

Joe Weisenthal: And that resonates because I really like vibe coding, but I can't explain how any of it works. So if someone is like, I would, you know, I like build little like toy apps and stuff, but I get really anxious. I couldn't explain, that's why I'm not a software.

Tracy Alloway: Yeah, yeah. Well, just on this note, I mean, people tend to talk in generalities when it comes to AI boosting productivity. Or maybe AI changes the way we work, or it leads to some new ideas from your seat at Goldman. You know, you're a manager. You're looking at the bottom line of like all these businesses. What exactly is the outcome, the specific outcome that you would like to see from your developers using something like cloud code?

Marco Argenti: It's really about increasing the output. So I want to see I was actually having this discussion this morning. I was looking at some of the reports on some of the deliverables for our cloud migration, which is a very important thing for us. And I was looking at this really big, a project that was saying it was not only green, it was like two months ahead of schedule. And I was saying, this is how we know when things are going to work. You're going to consistently start seeing projects that are actually finishing ahead of schedule, which means that then people are ambitious, they want to do more, and therefore you end up with output that is much higher than what you had before. And listen with developers, obviously, the biggest question that everybody ask is, okay, what are you going to do? Are you going to cut developers this and that? So first of all, with all the innovation that I've seen in the last 30 years or so, I kind of never seen a moment where really people were reducing the number of developers, because if I look at, the things they were not doing in a certain year because of budget reasons, because of complexity reason, because of prioritization, the stuff that is below the cut of the backlog, it's a lot and not a lot of that is really driving the growth of the business. So it's good to have the optionality to do it. You know, you have the optionality of, say, I now I have 120% of my capacity. I have 130% of my capacity to do. I want to do 130% more. Great. If I don't, I have the option to reduce, so but so that's really what, how we measure it. It's really the impact on the timelines of delivery. It's output. It's basically quality and timeline becoming you know, quality gets actually better and the timelines get short. So that that's what we measure.

Joe Weisenthal: Obviously one of the big questions for the market this year is what is the impact of AI on legacy software provider. And there's various theories about how they could be disrupted. And there are reports, I think about, anthropic having quote, forward deployed engineers inside Goldman Sachs. So anthropic employees building out AI systems internally, maybe that could, you know, replace some legacy software right now. Can you say, like there is a change in the balance of power when there is a given piece of software up for negotiation?

Marco Argenti: I think generally there is. Okay. If first of all, that is kind of always been that tension because, imagine, for example, you know, like imagine when they were software, they didn't run on the cloud. And then all of a sudden a bunch of new vendors are coming to you and say, hey, wait, this like, why are you running that on your mainframe or your or your on prem? Why don't you run it in the cloud? Or remember when software needed to be installed? Then everything became browser based and so? So there is always been a little bit of a cycle and renewal. What I say is that today that cycle of renewal is is much faster. That's really what it is. And I would say I generally resist, making like really broad categorizations. I is, by the way, largest possible assuming it's like saying computers. Okay. Yeah. But even software is very broad. And so within the software, category, I think there are winners or losers and there are winners and long term and losers in the long term. But it's really like, people tend to make it the category and then then maybe throw the baby with the bathwater. So here's an example to me. The question that I asked myself, with regards to what is which vendors am I going to, which software am I going to have like a few years down the road? Is software generally is attached to a process or a certain ways of working? Okay. It does something for you. And and they put it in the form of, of an application that you use. The real question is, is that process and or ways of working going to be the same, or is it going to change in five years? Then you can determine what the is, basically the likelihood that the software is going to be robust to that or not. For example, is accounting or closing the books going to be very different from five years from now? I don't think so. Really, he hasn't really changed. I mean, everything changes, but he hasn't really changed much. And so if you are operating in the general ledger type of category, I don't think, there is a, you know, all of the sudden you take, you know, a GPU or a cloud that is going to close your books magically, you know, you should. Have to do the accounting, and. You still need to do a lot of that. And it's very regulated importantly. Right. It's extremely regulated, you know, jurisdiction by jurisdiction, country by country, you know, product by product, industry by industry. So that part is scandal to me, you know, in, in kind of the safe mode way. And then you go to the, you know, other end of the spectrum and you have, sometimes, you know, software that, you know, kind of is aligned to the way people do things today, like software being one of them. You know, like if you look at the software developer lifecycle, a lot of that is changing. Developers are, are, are, are developing software by developing specs today. And so if you are too much in the weeds down there in that mechanics and you don't adapt, for eyes or agents doing that work, software development, lifecycle deployments, rollbacks, monitoring, observability and all that, I think that path will be very much disrupted. Or if you're adding a sort of a UX on things, that's another class. So, you know, a very simple process. I don't know, you're doing surveys or expense reports or whatever. And now people are going to start expecting, you know, their personal assistants or agents to kind of do all that mechanic for them. And so I think that path is probably something that has a bigger question mark, on top. And so I always ask myself for the process question first, the process transformation section, a question first, and then a consequently, what's the tool that is going to support that? And just to press on this point, though, have you replaced any third party software providers with something that's been developed internally through AI? We have we have terminated contracts already. Yes, absolutely.

Joe Weisenthal: Okay. Now I'm not going to ask you your follow up questions again, because I know that your name's. On these stocks. Like I would I will. But overall, yes, absolutely. Absolutely. You know, the thing is like the whole buy versus build okay. Yeah. The equation has changed quite a bit because if you know before it well buy versus build was always like okay guys how long does it take to do the to build this. And you get that answer, which is where we can do it in like, you know, X amount of years and X amount of millions of dollars. Now I'm starting to see people coming to me and say, by the way, I had some time, you know, this week and then here is a perfectly working application. So the cost, or at least, you know, for simple applications, the cost of kind of, you know, but builder from a time perspective and from a, you know, extra cost perspective has gone down quite dramatically. So right now, you know, like the little things are most likely going to be build the very big large, the software that is to be, you know, deploy that scale across thousands of people and etc., so that that big complexity, as you know, from, you know, we all do toy stuff with our cloud coats at home and whatever. You know, there are still some rough edges like, you know, and so it's hard to think that all all of the sudden the big applications are going to disappear. And so that's really what, what I'm saying that if I look at the applications that I buy today, there is a lot of small applications. So I'm saying and so the build is kind of the pendulum is starting to swing back towards the build, at least for that category for sure.

Tracy Alloway: What is a forward deployed engineer? I know that's like one of the hot buzzwords of 2026. And I saw headlines that there were anthropic forward deployed engineers at Goldman. I have no idea what that means.

Joe Weisenthal: Yeah. You know, like what is that term? I think they do when they got there.

Marco Argenti: Okay. So I think, Lisa, that name has changed quite a bit. Out of date. No no no no no no I mean that is the latest. Oh so you are fully you are in the V in the V v latest of of that term. But remember, I mean there was a time where you used to call them solution architects. Right. And so, so the point is right now, I think one trend that I see, which is also kind of true for us, is when things change so much and so rapidly, you kind of want to go to the origin of who produces this new thing. Okay. So the least intermediaries you have and probably the faster you can go. And so going and working directly with the model providers is generally a good idea, because if you're putting someone in the middle, this company is going to have to be trained. It's going to have to be, you know, there is a cycle which, at this point, a very rapid change is going to slow you down. And so the first thing that that term means is, those are people that are actually normally building the product. They're normally building the cloud or GPT or X or so that's the first differentiation there, straight to the source of of of of of the, of the AI production in a way. And second is that, they are generally product people. So people that have actually built those tools, rather than people that are more like support and deployment people. And so this characterization is you take the classic, you know, sales support, team or solution support team, which was mostly doing integration. And when things are so rapid, you know, you know, it's like, you know, imagine if there is like something like, I don't know if, if the cloud style works changes so fast. Instead of a fashion assistant, you want to talk to the tailor because they can actually make it. Yeah. You know, things are changing so fast. So those are the tailors.

Tracy Alloway: So on this note, one of the things we heard in support of Sass was this idea that while integration is still going to be really important, and that's really going to be like the major hurdle for a lot of the stuff. Have you found that AI is making integration even faster at this point? Has that, you know, basically become irrelevant nowadays?

Marco Argenti: No, I think integration is is extremely important, especially, you know, for, you know, what the industry calls like systems of record. So when you do something like, you know, when you do a process, so then you have a source of data, like, you know, your CRM systems could be a system or record or, you know, you have your client system or record your accounting system or record and those when they become the authoritative source of an answer, they need to integrate with the rest of the FA and the rest of the data, the rest of the applications. So I can see that those vendors that sit on top of those, we can argue that they will implement, there's nobody that is better position than them to implement the AI that will kind of reach outwards and actually do that kind of integration. So I think those who will evolve, so that, you still get the same level of automatism and you still get the same, the same benefit of speed, but it kind of comes from within. I think that part is probably something that, that would be remain very valuable. And so in general, I don't have anything against the again, I don't have anything. It has the SAS category at all. But I overall but I have you know, as I said, of different opinions on who actually is going to adapt to the future and adapt to the future. And those who don't know.

Joe Weisenthal: You know, you mentioned this idea of people have a few extra hours over the weekend and they come in in the morning and they're like, well, you know, I had some extra time and I decided to to do this. What's the coolest or most novel example of something that people basically vibe coded in a limited amount of time that wouldn't have happened, say, two years ago?

Marco Argenti: So, I've seen people doing, like cloud migrations of, of, of legacy applications that were on premise, you know, once they have been enabled with, with those tools, a little tiny matter of hours, I've seen someone, build, a complete like, travel assistant, for corporate travel assistant to that, that looks at your calendar, it looks at the flight delays and look at the rebooking stuff, literally, like in a doing a meeting where they were not paying attention. So.

Joe Weisenthal: So those are some of the things that's. What I'm doing right now. And while we're doing this podcast.

Tracy Alloway: Actually, that brings me to exactly where I wanted to go next, which is I'm curious, like, do a large corporations have a token budget the way they would have a dollar budget in the past? So, like, I would love to have unlimited access to coding models and whatever and actually just play around and try to work on it. It's one of my favorite questions. But I'm curious, like how you think about token allocation within the firm and whether there is intra firm competition for compute?

Joe Weisenthal: For compute? Yeah, it token allocation could be like included in your performance. Right? If you do well you get more. Different teams and stuff like that. Whether that's part of what you think about for planning.

Marco Argenti: Absolutely. So, you know, a few months ago I did, and I, I spoke about predictions for 26. And one thing that I said was, you know, there's going to be the birth of the personal assistant. And that kind of happened with Open Claw and all that stuff kind of early on. And then the one was, there's going to be a token sticker shock for CFOs, right? And all of the sudden they're going to start seeing bills that they absolutely did not expect. Jensen Wong is it an interview today or. Sorry, not today. In recent weeks, something about like, oh, if I'm paying an engineer $500,000, I hope that you're spending at least $250,000 on tokens. Now, again, as many people pointed out, that's like the barber saying, oh, you really need to get haircuts every week. Nonetheless, we're talking about some pretty big numbers. A lot more than just like a Claude Max plan for 200. Right. Now. So talk about that right now. Okay, so first of all. Lesson number one is, You need to centralize the access to models, okay? So that you can monitor monitor it and then optimize it. Okay. So the wild West of everybody goes and calls an API and starts consuming tokens. And then you find out later on is a big problem. And so that's why we built, this GSI platform, which has, you know, what's called the model gateway. And the model gateway intelligently routes requests, to, you know, the combination, the Pareto frontier of, quality and, and cost. Okay. So you got to centralize that. It's not the one size fits all because many cases, if you're asking, you know, what's the weather, you don't need to call on a cloud or 4.6, you can ask it to leave any local model that you ran very cheaply on premise. So there is there are ways to optimize that way. Before you start even having the conversation. You're consuming too much, too many tools. We just this is very interesting to me is a big part of the problem that you're trying to solve is and we know, like, you know, ChatGPT they intelligently root, you know, they do some on there. You go to SI.com and they'll try to route it to the best model, and there might even be some conflict of interest because they probably want to rooted to the cheapest model. The user wants the most performant model. But how much of the work of your senior engineers is essentially solving this problem of the right query? Going to the Pareto optimal. Model, it is, a big a big part of the time of the spent by the AI central group. Okay. Platform group, the platform group work worry a lot about, where do I where do I get the right data, for example, for this question, and which model do I write around it? So that's a big, you know, because again, I spoke about potato frontier, meaning the optimization between a quality which we don't want to compromise and the and the actual cost, you know, and you can be ISO quality, very different, you know, price points because not not all questions are require the most expensive doc. Right. So that's point number one. So what I'm trying to say is my philosophy is to try to isolate, the developer or the user from the token. Anxiety is a little bit like with electric cars. Okay. At one point, if you have 80 miles of range, you're always optimizing routes. And maybe I'm not going to go there. I don't need this, ice cream today or later. You're self-limiting in ways that are kind of, you know, unused. Really no useful optimal micro optimizations. We don't want people to go there yet, at least right now. It's a time where people need to really find the the best way to kind of, you know, do more and do the best possible work with AI. And let us meaning internally in the sort of a central team to optimize it in a way that, you know, we're going to make it economical. And I think, reducing the talking anxiety is a big challenge. But I think it really frees up, you know, creativity and what you can do with AI. It's also like, a little bit like, you know, there are certain problems that you don't want to optimize too early. Okay. Okay. So for example, yeah. How much time do you want to optimize now for? I remember we used to kind of optimize the way to web pages because they were too slow to load. And then at one point, the editors or whatever say, why can't I put yet another image? And then people are starting to say, okay, why don't you do it? And then on the back, I'm going to work in optimizing your images rather than asking you at the most, you can put three images on the on the homepage. Right. So that's the approach I want. People right now, I would rather, have them, on the side of usage and let me worry about optimization. And the other point is really at the end, human hours always tend to be the most expensive cost. Okay. And so as long as, your token cost is, your token cost per hour is less than your wage per hour, that is a kind of a positive ROI. And so at that point, it's fine.

Joe Weisenthal: Well, just on this. No. What's your what's your feeling about future costs of tokens and whether they're going up or down because you hear different things on this, one of the things you hear is that, again, going back to the beginning of this conversation, AI has improved so quickly in the course of, you know, months, if not weeks that those costs are destined to come down. But on the other hand, we know that the hyperscalers are still losing money hand over fist for, you know, power users such as yourself at Goldman. So where do you think those are going over time?

Marco Argenti: I my my personal view is, token cost is going to go down quite a bit, but token numbers are going to go up. Yeah, probably even more. And so total token cost is going to actually we're going to have to accept that is going to be a major item of cost in any organization. And it's to be compared to the cost of people and not to be compared to the cost of, you know, it or tcpip packets or computer or any of that. If you look at just the number of tokens being used, for the same, use case, if you go the reasoning route or you don't go the reasoning route if you go the agent route or not, the agent route if you go the open law route, where, you know, it checks every, every, you know, starts having these tasks that are firing one after the other, and then you have to start to have verifiers, etc., etc.. So I think the trend will continue with regards to more and more of those. But the cost the the the per unit cost of token, I pretty I'm pretty sure that is going to go down also because as you know, GPUs are becoming more powerful. They you know, the cost per watt hopefully is going to go down. And then also like, to be fair, I mean, you know, these hyperscalers are doing a lot of optimizations to try to ram those stacks on their own hard right, which will potentially kind of also generate some economies of scale.

Tracy Alloway: Can Goldman employees like, run open claw on their work computers? And I'm curious like about the degree to which you have people who like want to I want to install this or this seems really cool. And then think about the security imperative and how you handle that aspect. Not the token anxiety, but the sort of I want to install this, this is awesome.

Marco Argenti: I hope it is. You know, like as a bank, we're pretty locked down in terms of what you can install. You cannot install stuff that is not, in the, you know, in the corporate, in the corporate App Store in a way. And so there's no way.

Joe Weisenthal: But do you feel like you should? Definitely. No way, though, because can't you just ask Claude how to install itself?

Marco Argenti: But he's not gonna be able to execute. It's not going to be able to create the actual executable. He can actually, even GSA system today can spit out a lot of code, but it spits out source code. Okay, now our source code doesn't run executable, so it needs to be built and it needs to be turned into an executable. It needs to be signed. Otherwise the operating system is going to refuse to run it. And so it just doesn't run unless you have that.

Joe Weisenthal: But do you feel an anxiety where you know, startups there? Probably. And you know, there's not a ton of startup investment banks, but there are various fintechs and other things I want to chip away at parts of your business, and they can run perhaps faster, and they can be a little bit more liberal about what their employees are allowed to do, etc. do you feel like you have to keep a certain cadence of expanding the list of those executables that are, able to be run?

Marco Argenti: So, I think, so I'll give you two answers. So first, I want to make sure that I answer your first question, which is, we're not using open claw. Okay? Okay. But some of the properties of open core has actually having informed the the way we are building our agent platform. Okay, agents today because of Open Claw have actually changed. There are three very important new if you if you break down warp enclosures that are it is my own interpretation. I never I actually never even spoke about this. There are three characteristics that make open claw what it is. One is it's a constant loop. So it's basically what, you know, in information theory you can call an observer pattern is something that continues to run and observe. So there is that it's a constant observer. So it runs constantly. The other one is it can schedule events every 7 a.m. do this or I, you know, like in personal life we all have some something like that that sends me the news in the morning and all that. So there is a schedule, ability of tasks. The third one is you can instructed to kind of change its own behavior because it has this files dot MD5 soul.md, the way you so you can say things like hey, I would like you to never use this term or please change that or change the way you filter news because and so it kind of writes its own software to do things for you without you even, you know, seeing what's behind the scenes. And so instead of letting people, install open claw on their computers, what we do is we incorporate some of those characteristics into our genetic platform so that it does things that are more similar to open cloud. So that's that's the approach. You're kind of made that the your second question was interesting because basically if I read behind the questions, are you asking whether there is a sort of a velocity disadvantage with regards to, you know, us versus others? Okay. And I think at, you know, I often say, you know, that there is a difference between speed and velocity. Speed is almost like, you have a certain sprint. Okay. But then at some point you're going to hit the wall a security wall, a scalability wall. There's going to be a bug. You don't know what you're doing and there's going to first, sooner or later you're going to be hit by that. You will be like, you know, most airplanes are in auto. I have autopilot that can do everything. So theoretically, you and I could go in the cockpit and for a long time during the flight. We will feel pretty good about this. Right? We will drink, you know, some some soft drinks or your tea. We might maybe watch some videos. We will be very happy. It sounds great. The passion is one point. At one point. In time in the flight, there's going to be some storm, and there's going to be an autopilot disconnect. And you and I are gonna look at each other, are going to say, oh, right, it was the pilot.

Tracy Alloway: Can I just say recently I was on a flight, I tell you this, and I was going to Newark. Oh, yeah. We circled three times. We tried to land. It was during a storm and they kept not landing. Everyone was like, starting to get pretty annoyed because we were up there for a while. And then the flight attendant comes on and says, unprompted, by the way, we have plenty of gas and everyone's stuck. We got no, it's like, this is a question. This is an answer that no one had. No. What do you need? I'm sorry. I just don't like to do that. Then everyone got really nervous. Okay. By the way, we have plenty of gas. But you know what? I was almost fearing that you would say. Is that a pilot?

Joe Weisenthal: No, no, I wasn't. And then we did land in Washington, DC. Okay. Oh you did? Yeah. From Newark. Yeah. No, no, that's not good. But that makes sense.

Marco Argenti: Yeah, yeah. So so that's why I mean by velocity is really like, is like the marathon is, you know, sustain the speed for a long time in a certain direction. Yeah. I don't think by randomizing that you actually gain velocity, you gain an instant speed of some sort. And so I'm kind of optimizing for velocity.

Tracy Alloway: But related to Joe's point, though, you are a regulated bank. Right. And so there are restrictions on what you can do in terms of technology. I am very, very curious what your discussions with regulators are right now, because a lot of regulators, this is still pretty new to them. A lot of the models basically are black boxes. How do you convince them that, like they're running as they should, that they're spitting out the correct output, that you understand how they're actually functioning?

Marco Argenti: So this is not the first time that banks use neural networks, okay. These are just much larger neural networks. But we've been using neural networks for like a decade plus. And so every bank has already gone through, the motions of explaining that the neural networks don't have perfect explainability. Therefore, you need to change the control system around them. You need to look at what actions can they actually do. And then you limit the actions, okay. And then you there is a function called the like model risk management, which is a very standardized, you know, function within every bank that for each of those neural networks, you need to have, an inventory, you need to have a risk tiering, and you need to put controls around that. So it is, not really that much of a new thing is more of an evolution, where now you have things that are much faster and much more powerful. But the basic pattern and the basic discussion with the regulators is kind of the same, which is are you classifying the risk tiering of the application? Right. And then which controls are you putting and are you putting human supervision and human in the loop. So for example, for code. We don't allow AI to auto approve the wrong code. Okay. They all they can do is publish what's called a pull request or merge request the same way as, a developer would do. And we kind of have a sort of a zero trust model there because we don't assume that, maybe a junior developer is going to be less, more bug free that than an AI, right? And so we have several controls in place. For example, there isn't to be there needs to be a human that's more senior than you that actually looks at the code. And then certifies and approves, and then after that, before it goes to production, it goes into something that is called Cicd or, you know, continuous integration, continuous deployment pipeline, where when it goes through the build phase, etc., there is a lot of checks that are injected into that. There are security checks, there are tech risk checks. So I don't think at the end of the day you really lose too much velocity or at all. You just need to invest more in those kind of things. And that's really what. And the regulators, I think if you bring them back into sort of a familiar territory and you're also honest on things that, you know and things that you don't know, and for the things that you don't know, you kind of put to higher protections. I think the conversation is generally very positive.

Joe Weisenthal: You know, we have an episode that we recorded several weeks ago that we still haven't released. I don't know exactly the timing of that one or this one. We interviewed, Scott Borg, the former CEO of Greenhill, the boutique investment bank. And part of the reason we had that conversation was because we want to know, like, if I is going to someday disrupt banking as we know, like what was banking as we know it. So we talked about the history of investment banking. But one of the things that he talked about was that a big advantage that the banks had was this sort of information asymmetry that they would know a lot more about their industries and so forth than their clients. And this was profitable. Now, going back to your answer, the very first question, you're like, okay, a client might call, Goldman. And they said, what is the Strait of Hormuz closure mean for this portfolio shock, etc.. I was going to ask this exact question. Yeah, I. Kind of think I could do that. I think I could you I mean, no offense, I'm sure your platform is a little bit better than like, but I think I could like get 90% of the way there. And I bet I could, like with a little bit of data, build a basket that says, I want, you know, helium shortage basket, which, which companies would I short if I think the helium shortage is going to get worse, I could rebuild a basket the way a trading desk would. Do you think long term, like the air erodes a certain structural source of profitability for banks? And you worry about that?

Marco Argenti: I think you can get to the 90%. But I think clients are really paying us for that extra 10%. Okay. So I think that's the answer.

Tracy Alloway: So what is the extra 10% in that context? Is it you model for slightly better or is it also the do we. Have access to you know,

Marco Argenti: We buy a lot of data that, you know, it's very expensive and it's a massive quantities and it's very up to date and very real time. So we have a little bit of a data advantage. We operate across multiple asset classes. So we see the trading side. We see the asset management side. So we have a sort of a correlation between assets advantage that we see those because generally rates move you know interest rates can move yields that can move. You know there is a correlation between all those indicators. So there is another advantage. We have a global advantage. We have people on the ground in, you know, 100 plus countries. And these people have relationships and information travels through those channels. And also, you know, like we generally deal with very complex portfolios. So this is not you and I may be having three stocks or 4 or 5. This is a like very complex, multi-year assets up with complex products like swaps or swap channels or exotic products, etc., etc.. And so that's really the 10%, that, you know, the clients that we have, you know, really value and where we really need to get, it's like at the end of the day, Lisa, if you look at, look at formula one, okay, the difference per time, a particular pair of per lap, between, you know, the, the, Mercedes, and, take, you know, your favorite last team. You know, it's sometimes one second out of the two me out of two minutes. And that is the difference between, you know, getting a $100 million a year sponsorship or a $10,000 sponsorship. So for for sophisticated clients, that 10% is really what the money is. And that's really what people are paying us for.

Tracy Alloway: So actually you mentioned all the different businesses at Goldman. And there are a bunch of them like asset management there's banking, there's trading. A lot of those businesses aren't supposed to talk to each other in various ways. And so when it comes to the data, is there like a data leakage issue where you might have a model that's, you know, in-house, like GSI, that's pulling data from different sides of the company in ways that, you know, maybe it shouldn't be. Maybe it's really hard to tell given the complexity of the model. Is that something you have to pay attention to?

Marco Argenti: Absolutely. So, we have the concept of info barriers, okay. And the info barriers are, are enforced, throughout the entire system. Okay. And, they're linked, to your ID or your account. Okay. So if I'm on the private side, I can only see certain information. If I am on a public side, I can only see certain information. And I cannot even know about the information on the other side, I can I don't have access to the files, to the C, to the folders. Nothing. Each I or each agent or each application. That's the beauty of decentralized platform. Needs to get an ID or a badge, and that badge is attached to the exact same info barrier says any application or any computers. And so these are enforced basically at the source. So even if it is the same, type of model, but that particular use of the model, that particular session or the model that needs to get a ticket or a badge and that badge or those keys, just take them to a certain place. And so this is one of this been, you know, it took us almost two years to build the, GSI platform. These are this back to the reason why you can't be casual about these things. This thing is not being built by some random vibe coders because you need to worry about cyber. You need to worry about the info barrier. So you need to worry about all that. And so when I talk about, you know, there are places where you can leverage and do correlations, but there are others where you absolutely cannot. And this scandal that is foundational to the fact that, you know, you need to have, you need to be ready for a, you can't be casual about that.

Tracy Alloway: So I take your point that there's never been a technology that you've seen in your career that has actually reduce the need for software engineers, and that the nature of the job of software engineers is changing, maybe gets more high level. And whatever. Setting aside the pure head level of headcount question is AI changing right now across anything software, technology or otherwise, the types of person you're looking for or changing something about the nature of the type of talent, your person?

Marco Argenti: Yeah, absolutely. Great question. So, I think, in this day and age, almost nobody is an individual contributor really. Because when you're working with agents, you need to have, at least three fundamental characteristics. One is you need to be able to explain what you want to get done. Okay. The second one is, you need to be able to delegate work. Guess what? Because you're going to have multiple agents. One is specialized, for example, in doing DCF calculations. And one is specialized in doing research. So you need to be able to break down the work into chunks that can be, executed in parallel in some way. And then three, you need to have the ability to supervise. You need to actually look at the output and say, okay, I'm good with this or go back. It turns out that those three things I explain, delegate and supervise are kind of the 1 or 1 of managers. Managers need to have those three, otherwise they can't manage a team. And so I is kind of turning everybody a little bit into a manager. And those are kind of the skills that we are actually looking for, for people that, they know that they're going to have agency on, tools that at some point are going to be even more proficient and specialized than they are. And so the most important thing is really the ability to ideate, to explain, to delegate, and then to really know what good looks like. And I think that is a big change. And I don't think everybody is going to actually rapidly go through that. And I think we're doing a combination of, training. There is a combination of, exposing them to, you know, other people. Like one of the advantages of having forward deployment engineers is also that there is a little bit of clash of culture that is happening around the table. And so people think really, really differently. And that pushes people outside their comfort zone. That's why I'm saying that, you know, there is a little bit of a metamorphosis happening. There is not just about efficiency is really thinking about is my job going to stay the same now is actually changing quite a bit.

Joe Weisenthal: So I'm thinking how to frame this question. But what's work life balance like now for a developer at Goldman? Because you have this existential existential angst about jobs potentially changing. At the same time, you have AI tools that enable more productivity, and you also have this thing happening where I feel like, Joe, maybe you know more about this than I do, but I feel like a lot of vibe coders, like like it's addictive. Yeah, right. It's like you're you're pressing the button of a slot machine, you're interacting with Claude and you're seeing what it spits back out over and over and over again until you get that big win. And so I've heard people talk about burnout among developers who are just doing so much with this right now that they're just hitting that button over and over. And it was a good discussion in the odd lot discord recently about exactly this. Some engineers and semiconductors feeling that the job has become less satisfying because it just all and I think is sort of what you're getting at the sort of slot machine. Yeah. Where it's like, oh, you're going to like hit the prompt. Okay. This is the great output. And then it's like they field of work as like less satisfying and stuff like that than actually like writing code. So yeah.

Marco Argenti: I mean listen again I this work, I love the fact that I'm a little bit, older than most to hear an engineer. Second, I've seen that the first time people had the Excel and I've sat for the, you know, the first time people had, Python. Oh, my God, I don't need to know Java. And then the kids start to code and there is this whole coding movement. And then you get the you start creating your applications. I've seen the first time people have mobile stuff and, you know, mobile apps. And so I think a little bit of that is because it's new, to be perfectly honest, you know, and I think there's there is a little bit of that, but there is a little a lot of novelty to that. And then I've seen that people have been using those tools, for a couple of years. They're taking them a little bit more like, okay, it's professionals tool. And I'm going to use it for what I actually need rather than just trying to discover. One thing that I've seen is that because maybe of that, but also because of, what you can get there is a sort of a in a way, reward cycle that is pretty quick. Yeah. People are very excited, actually. There's, there's some sort of a joy of the profession that is actually coming out as if engineers were feeling like this job is new again, because a lot of engineers have seen the same pattern, sometimes for 2 or 3 decades. Yeah. So that has been, something that I observed. There is also a lot of peer pressure. There is a lot of fear of missing out. And so people are rather than, you know, is no longer like, me trying to push the car uphill is more people are actually looking at their peers and they're looking at, oh my God, how could you do that? And so it's kind of spreading horizontally quite a bit, which is which is really nice to, to see. And so, so far I have to say that has been positive, positive change. And also one other thing that, you know, you're talking about burnout. I see that a lot of people get fatigue. I wouldn't want to talk about burnout, but they get fatigue when, you know, there are a lot of repetitive tasks, especially for a developer. Here's an example. Let's say you go from, you know, a version of, a Java library or spring boot. To another version. And then all of a sudden you compile and you get to rebuild and you get all these errors that says you need to upgrade. Honestly, upgrading libraries is not the most fun job. And if you need to do it 100 times, or is like someone says, by the way, guys, we have this new design, design, a new logo, new colors, implement it on like 200 websites. It might be fun. The first ten and then it becomes a drag. And so I think taking that away kind of in making them, they focus more and more on the plan, for example. And so right now let's do a migration plan to the cloud of a complex application. They spend maybe 70% of their time going back and forth with a very powerful set of eyes to really get the plan right. They feel a little bit more elevated. And the mechanical part, it's kind of, you know, left to, to, to to the machine the same way. I mean, listen, I started developing when I was literally flipping switches, okay. And then pressing a button, which would move the register up one and then came, you know, some languages, like, see, oh, my God, now, I don't have to flip switches anymore, but the guess what? I need to do? Memory management. I need to do pointers. I mean, there's a lot of heavy lifting. Oh, I have a memory leak. I'm going to spend a week before I actually finally identify that. And then, you know, it comes to Java or garbage collection. I don't have to worry about memory leaks anymore. Fantastic. And then comes Python, which is all all that rigidity. So much easier to you know, be type, free and so forth. And so every time you kind of keep raising the bar and a lot of the candle mechanics kind of goes away. I think this has been like a ten year jump, in, you know, in, in a matter of two years. But I think overall, nobody really likes to, to have that toil and that mechanical work. And I'm actually quite happy that people are going to spend that maybe initially more time because they're excited. But I don't think that I'm enjoying rather than things that they just, they dread.

Tracy Alloway: All right. Well, market will have to have you back on the podcast in another year and a half, I guess in discussions. Yeah. Three months. That's right. To go on the reduced air timeline. Thank you so much. Thank you so much for coming back.

Marco Argenti: Thank you both of you. Thank you.

Joe Weisenthal: Thanks for having me. Thank you so much.

Tracy Alloway: Thank thank so Joe, that was great to catch up. Yeah. One thing I thought was really interesting was his point about the discussions with the regulators and framing it like very similar to previous technological advances, where you're not necessarily explaining exactly how the models are coming to certain conclusions. Yeah, but you're more focused on actually limiting the risks and making sure that they're in the right bucket for risk assessment. Now, I thought that was really interesting, just that some of these technologies, the, you know, the black box. Yeah, the limbs are not the first black box. I mean, we've actually been talking about black box trading for years and finance before. So the idea of like, okay, there are these things that are happening, we can articulate them and whatever. Like it's not the first rodeo for finance is really interesting.

Joe Weisenthal: I'm also, you know, I thought the whole conversation about, token budgets and allocations are interesting. The idea of like, okay, part of the job here is you have a bunch of different models. Everyone, in theory wants the most performant model. But how do you find that optimization where you get the best performance relative to price? It sounds like a pretty interesting like engineering problem. Yeah, I would actually love to do more on that question.

Tracy Alloway: I would too, because it's such an interesting like a question of incentives. Right? And like how does how how do you actually like prioritize how each. Project, like what constitutes a good output and how when do you sacrifice a little bit of quality for like ten x less, token budget or whatever, like you said this would be, it would be very interesting to talk about how that problem specifically gets solved inside of.

Joe Weisenthal: An order token economics. Yeah. Efficiency optimization. Yeah. Well, we'll have Marco back on very soon to talk about all the new things that AI is doing. But for now, shall we leave it there?

Tracy Alloway: Let's leave it there. This has been another episode of the Odd Lots podcast. I'm Tracy Alloway. You can follow me @tracyalloway

Joe Weisenthal: And I’m Joe Weisenthal You can follow me @thestalwart Follow our producers Carmen Rodriguez @carmenarmen, Dashiel Bennett @dashbot and Cale Brooks @calebrooks And if you want more Odd Lots content, you should definitely check out our daily newsletter. You can find that@bloomberg.com/oddlots And you can chat about all of these topics 24-7 in our discord, discord.gg/oddlots And if you enjoyed this conversation, then please leave a comment or like the video. Or better yet, subscribe! Thanks for watching.