Apr 30, 2024 · 29m · a16z
Marco Argenti (Goldman Sachs): Turning Developers into Clients
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of a16z's 'In The Vault,' Goldman Sachs CIO Marco Argenti discusses his career across major tech platform shifts, enterprise engineering strategies, and how generative AI and hybrid architectures are transforming financial services.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 2.7 Guest teaching 3.3 Guest disagreement 0.3 The host pushing back 0.2
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 22:55 Marco broadening host's scope from engineers to all knowledge workers
In a very mild reframe in an otherwise highly agreeable interview, Marco gently expands David's point about 10,000 engineers to emphasize that AI impacts all 45,000 Goldman Sachs employees.
Hardest push from the host ▶ 14:02 Host pressing on Goldman's historical build-it-all biasDavid challenges the traditional financial institution mindset by citing Goldman's historical tendency to build proprietary tools like internal email clients and asking how Marco justifies buying third-party software.
Biggest teaching moment ▶ 8:58 Control theory applied to regulatory complianceMarco educates the host on control theory from his academic background, explaining how institutional controls and regulation act as feedback loops that accelerate long-term innovation.
The host holds their own ▶ 22:33 Host bringing specific institutional metrics to ground AI ROIDavid demonstrates strong institutional context by citing Goldman's 10,000+ engineering headcount to illustrate the massive financial leverage of 10-40% productivity gains.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| Legal and Financial Disclaimer | 2 | 2 | 0 | 0 | The host opens the episode with a background overview of Marco's career across major tech shifts. Marco provides an extensive narrative of his 40-year trajectory in technology without any pushback or disagreement. | |
| Transitioning to Financial Services and Engineering Tenets | 2 | 4 | 0 | 0 | David asks about the transition from tech giants to regulated financial services. Marco educates the host by drawing on control theory to explain how regulation and engineering tenets actually increase long-term innovation velocity. | |
| Turning Developers into Clients | 3 | 3 | 1 | 1 | David brings up Goldman's history of building legacy internal software tools and asks how Marco decides between internal engineering and external vendors. Marco details the framework for looking outside first and eliminating undifferentiated heavy lifting. | |
| Generative AI as a Historical Technology Shift | 2 | 4 | 0 | 0 | The discussion turns to generative AI as a platform shift. Marco elaborates on the historical nature of AI, comparing it to the Gutenberg printing press removing knowledge barriers. | |
| AI Adoption at Goldman Sachs and Productivity Gains | 4 | 3 | 1 | 0 | Marco shares metrics on developer productivity gains from AI. David demonstrates firm expertise by highlighting Goldman's 10,000+ engineer headcount to contextualize the scale of those gains, which Marco broadens to all 45,000 knowledge workers. | |
| The Long-Term Impact of AI on Financial Services | 3 | 4 | 0 | 0 | David prompts Marco on open-source versus proprietary AI models. Marco outlines his strategic vision of a hybrid AI model architecture combining large general reasoning models with specialized domain models. |
Statements from this episode (14)
Argenti: 1990s startup launched online store before Amazon
“One of the very first online stores. In fact, we launched a little bit before Amazon”
Argenti: Nokia's Ovi Store reached 70 million daily downloads
“We were doing something like seventy million downloads a day or something like that.”
Argenti: Regulatory controls ultimately accelerate long-term innovation velocity
“Regulation and other controls are actually very important for sustained speed of innovation. So you might be slow at the beginning, but then eventually you'll gain velocity.”
Goldman Sachs operates roughly 10 core enterprise developer platforms
“Now again, fast forwarding, we have about 10 of them firm wide, like, there are heavily adopted like cloud platform, the data platform, or the identity and access management platforms the API platform, mobile platform, and so forth that developers can just bui…”
Argenti: All corporate leaders must understand developer mindsets
“And I think also today, given the importance of developers I think it's also a responsibility of a leader, not only a technologist to try to really understand what developers are thinking and what they want.”
Argenti: Pitching client technologists is as critical as pitching executives
“Like for example, some of the hedge funds or some of the quant funds, as you know, I mean, you know, it really, really well that, you know, we're talking to the technologist is just as important as talking to the business people.”
Argenti: Goldman Sachs looks outside for software before building internally
“So at the very beginning, we kind of introduced our digital strategy tenants if you remember, and one of them was look outside first before you build.”
Argenti: Goldman Sachs builds compliance scaffolding around open-source software
“And so many, in many cases, you know like we decide to use a vendor or we decide to use open source, and then we build a kind of, you know, like, let's say a scaffolding around it to make sure that A, it integrates with our systems and with our data and B, you…”
Argenti: AI removes knowledge barriers like printing press removed physical ones
“Compare the invention of AI to the invention of the printing press in a way that it removed a major barrier. First one was a physical barrier. This one is a knowledge barrier.”
Goldman Sachs selected 15 to 19 AI use cases for pilots
“So we have set up you know, a working group or a committee that kind of reviews all the ideas when we selected out of hundreds idea of ideas, we've selected about 15 or 19 actually at this point that we're kind of, you know, at different phases of piloting.”
Goldman Sachs sees 10% to 40% AI developer productivity gains
“And we're seeing the developer productivity go up, you know, can be anywhere from 10 to 40%, depending on the developer, depending on the use case, which is massive, even if it stays at 10%.”
Argenti: AI will accelerate real-time financial operations and complex modeling
“You know, you need to go through front to back extremely quickly that, and so all that, I think this shift towards more of a real time business is something that AI definitely will accelerate the ability to have very complex models of the world and of the econ…”
Argenti: Large proprietary AI models lead in reasoning capabilities
“Probably nobody beats those large models with regards to actually reasoning capabilities.”
Argenti: Enterprise AI will combine proprietary reasoning with specialized internal models
“And I think the same is going to be with AI, where you're going to have a reliance on, You know, large proprietary models, mostly for the reasoning capabilities, almost like, you know, I really want to understand what you're asking. And then kind of they will …”