We Built an AI CFO to Help You Grow Your Net Worth | Anthony Pompliano

They are coming after us. Peter, do you know what it's like to wake up one day and you've been waiting for a trillion dollar company to compete with you? That happened to me. It was awesome. These companies have tens of thousands of employees. Our engineering team is less than 10 people, but it's not about how many resources you have now. It is all about how do you get the most leverage using the right approach with the models. That is ultimately the goal of the business, right? If you use this product, does your net worth grow faster than if you did not use the product?

Absolutely. Great to see you again.

It's one of my favorite things to talk about because I think everyone has a different definition. And you can learn a lot by asking someone what their definition is. Some people are like, 'My extraordinary life is I want to work as hard as I possibly can and solve problems for other people.' I think other people say, 'I actually want to work as little as possible and I want to travel around the world and experience things.' Some people judge their extraordinary life by how much time they spend with their kids or how many material possessions they have. And I think the conclusion I've come to is most people actually don't define what it is, and so they don't really know what they're chasing. For me personally, I think about two components: is I want to work hard and solve problems for other people, but I want to do it in a way where my kids can see me doing that on a day-to-day basis so that they implicitly learn things like hard work, effort, etc. And so, what that means is I want to blend my life. My professional and personal lives come together, and I want to build these businesses, solve these problems, but I want to do it in front of my children so that they're able to kind of soak that stuff in.

Correct.

Yeah, look, I think there's two types of people. Some people are kind of, let me put together a 10-year plan and figure out what I want to do and then work backwards. Those people usually they kind of know where they want to go to college when they're starting high school and then they work towards that path. There's folks who in college say, 'Hey, I want to be an MD at an investment bank.' Then they work backwards and say, 'Well, in order to do that I got to get a job at this organization. I got to stay there for three years. I got to do this, this, and this.' My career's been very different, right? I kind of took the second path, which was I'm very interested in business, technology, and finance. I want to stay on top of what the latest, greatest, newest things are and then I want to find opportunities and be as opportunistic as I can. And so what that allowed me to do was even if you think about Facebook, I had never thought about going and getting a job at any company. I right out of college started my first business and after those first two businesses, I really realized that, 'Wait a second, I don't know how to scale any of this stuff. I'm really good at going zero to one and then I kind of have a like a basically successful small business that isn't really venture backable.' And so the idea to go to Facebook was maybe I could learn that there. They're pretty good at scaling, they're pretty good at growth and analytics. And so the opportunity was I have a problem, let me go solve that. I'll go to Facebook. If you think about after Facebook, I wanted to learn investing and it was cuz I didn't have an idea to go build my own business. And so it was, 'Well, let me go find other people. I'll just put some money in their business and then I'll help them.' And maybe that'll help spur an idea for me to create another business. And if you fast forward to today, the Silva product is maybe the best example where the best engineer we had working at any of our companies, he was thinking about leaving, didn't really know what he wanted to do. And I said to him, I said, 'Look, I know the one thing about building a company is if you can solve a problem, you're able to build the business.' And so I just thought for a couple days, 'What is the biggest problem I have in my life that also the audience would likely have as well?' And I had a spreadsheet, and that spreadsheet was a big pain in the butt because I'd have to go update it every single week or month. And so I just said to myself, 'Well, let's solve that problem for me, and then we'll see if other people want it.' And so whether it's getting a job at Facebook or starting a company, it literally is how do you go and find the biggest problem and then opportunistically go and tackle that. I think it has for me been a much better path than 'What do I want to be in 20 years or 10 years?' and then working backwards.

I don't know what I want to do next year.

So, if you remember when we were at Facebook, for about 90 days I'd worked on the growth team that led Facebook pages growth. And at some point I got tapped on the shoulder by some of the growth executives and they said, 'Hey, you got to go and you got to spend these 90 days working with the VP of comms and the executive speechwriter, and you're going to help Mark Zuckerberg grow his audience on Facebook. And you guys are also going to help Sheryl.' And I remember kind of being like, 'What?' You know, is that like a real job? And long story short was the same things that we'd been doing to grow the Facebook page ecosystem and understanding the algorithm and the product at Facebook and helping small businesses and brands do that was very helpful to help Mark figure it out as well. And the conclusion was very simple things. Like, 'Hey, put your dog in the photo. If Mark's in the photo with his face versus not, certain words in the text description.' All of those things we kind of learned over time. And while I was there, I started to meet other brands who were big on Instagram but wanted to go to Facebook. And so I kind of cut my teeth, right? It was very much trial by fire trying to figure out how to help these people. And when I left, I realized I'd never done it for myself. And so I had a problem in that I wanted to invest and I knew people in Silicon Valley because of Facebook, but I didn't know like the next generation of founders. I didn't have a great network of finding these people or winning access to these deals. And so my thought process was, 'Hey, audience is going to be a currency in the future.' I kind of just got a front row seat to that if one of the most successful founders in the world and Zuck is saying, 'Hey, I want an audience.' Maybe the rest of us should go get one. And I knew that if I had an audience, I probably could get access to certain deals I wanted to invest in and give me kind of competitive advantage. And so that was in 2016. I said, 'Okay, I'm going to go do this.' And then from there kind of figured out, 'Okay, what are the steps in order to be able to build this thing?'

To be honest, I didn't know one if I was going to be good at it and two I definitely didn't think I was going to make a lot of money in it because the very first fund I ever had was like three and a half million dollar pre-seed fund where we didn't even take any compensation in terms of a management fee or anything like that. So this was love of the game if you will and it was really a test for me and my partner at the time Jason to say, 'Hey, do we like doing this? Do we think we're good at it?' Right? We don't want to go and try to raise some huge fund and then be bad at it or not like it. Let's start small and kind of walk before we run. Now, what I found very quickly was the reason why I enjoyed it was you kind of got to see the future. You got to meet really cool people, even the ones you didn't invest in. You got to hear about a lot of the technology trends, you got to learn a lot, but you also started to see trends by oh, all of a sudden there's a ton of companies that are using this technology. What if you could apply it to this other industry or use it in this other kind of example? And so to me it was almost like I was having the opportunity to potentially make money in the future, but being able to learn on a day-to-day basis and so the learning was actually a return on my time that I was just that alone was worth it and then the fact that maybe over the long run I could make money obviously just sweetened the deal.

Yeah, I mean look, I tell people all the time that right when you and I worked at Facebook, that was a pretty special time. I don't know if I will ever get the opportunity to be in an environment like that ever again. Every single person was super high quality, the company was scaling incredibly fast. I think it went from like 3,000 employees to 12,000 while I was there in a two year span and it felt like there was still this Facebook versus the world, you know, type feeling inside the building because we were constantly being attacked. There was all kinds of issues. There was people who were do I trust Facebook, do I not? Can they monetize? All these questions. And so when you have really talented people who have this kind of very tight-knit loyal kind of teamwork approach and you're executing very quickly, I mean that is special and so there's probably other companies over time that have had a similar thing, but I don't know if I personally will ever get to really experience that again in my career. So, I'm really grateful for that opportunity.

Yeah.

Yeah. Well, it's also if you can get a lot of the people who could be one-person companies, but you get them all on the same team, then they're basically their one-person divisions, right? Or they're like one-person products or whatever. And when you put them together, it's like 1 + 1 = 4 or 5. And I do feel like that is what Facebook That was part of the magic was the quality of the people and the ability to communicate quickly, kind of no-bullshit environment really allowed everyone to kind of, you know, do this amazing work. And then you add up thousands of people all doing that at the same time given the leadership and vision of Zuck, Sheryl, et cetera. The results speak for themselves as to what that company was able to do.

Man, I got We could spend hours on this. I've gone through maybe different seasons in my life of, you know, the types of folks that we're trying to hire. When I first started out, it was anyone we could convince to work with us, right? We had no name brand, we had no successful track record building businesses, and frankly, we didn't really understand what we were building or how to describe and get people excited. So, you start off everyone at the same place of good luck. And you just start trying to hire people, and you realize, 'Okay, we need to get a better story. The people are interested, but they're scared the business is going to go out of business.' So, how do we make sure that they're confident in the financial success or kind of stability of the business? You learn all this. At some point after a couple of years of figuring all that out, I then got the idea I was like, you know what I'm going to do? Hire the best people. I'm going to go to these big companies and I'm going to go hire people with great resumes and I'm going to bring them in. My conclusion over the last, you know, I don't know, four or five years of sporadically doing this is every single one of those people were amazing. They did every single thing that I asked them to do, but it was like a car wreck. You brought in somebody from one culture, right, who is used to lots of resources, frankly not a lot of internet exposure in terms of how they think about their job. Everything is risk management, everything is about kind of big bureaucratic process, and then you put them into an environment where people are like, we have an idea at 9:00, we're trying to launch it by 10:30. Those two things it's like oil and water or something, right? And so what we had to realize was those people are great at doing what they do in the environment in which they're doing it, but that's not our environment. That's not our culture. That's not the way we build businesses. And so we had to kind of go back to our roots and start to say, okay, look, what we need to do is we need to find people who frankly don't have a lot of experience. We need to find people who have a lot of agency, who want to work hard, who are very ambitious, but if they have been trained inside of one of these large companies, they're probably actually not going to succeed here because the environment, the culture, and the way we work is so different. And so it was almost like we had to go down that path and try it to learn what we don't want to do. And frankly the people who we had brought in, they also recognized it and they're way happier because they went back into the environments that they're like, hey, I tried that kind of start-upy type environment, that's not for me, I'm much more comfortable here. And so one of my conclusions over the last 15 years or so was like, there's no such thing as good or bad culture. Culture is all about employee-company fit because there's cultures as I've seen that are, I think, horrible. And there's people there that are like, we love this place. And vice versa, right? There's cultures I love and there's somebody sitting there, I hate this place. So the matching of employee to company is actually really important when you think about culture, not so much just do we have a good culture or not.

We have almost exclusively gone to if you want to work at one of our companies, you got to do a trial. We'll do two to four week trials and what I've learned over time is that was never a thing 10 years ago. I never I don't think I ever heard of someone being like a work trial. It was like there was projects during an interview process, but it was kind of you get it, you work on it for two or three days and you bring it back and then they judge your abilities based on that. What we do now is especially engineering, product, analytics, all that kind of stuff. We say, 'Look, we'll pay you the market rate for two to four weeks and you get so you're coming in the slack, you're getting access to all the systems, everything and you're going to start working. And we're going to judge you not only on your productivity, but also the quality of your work. We're going to be able to very quickly tell do you gel with the team, how do you communicate, how do you take feedback, how do you deal with any sort of issues or kind of corrections we give you, all that stuff. And what we find is not only one is it very obvious very quickly who's good and who's not, but two is it's a better thing for the candidate as well. Because what we don't have to do is we don't have to go and say, 'Okay, we want you to go through six, seven, eight, 10 interviews, then give us all these references, take a four week process, you know, blah blah blah, whatever.' We literally will meet people, I'll talk to them, co-founder will talk to them, and then all of a sudden we're just like, 'Hey, let's do the trial.' And we just jump right in. And so now two to four weeks later, we're like, 'They're good or they're not.' And in the same amount of time that it would have taken us to do all of that work to be like, okay, let's give them a job offer. We already know with certainty they're good or they're not, and it's just so much more efficient process that I think is better for us and for them.

Especially with AI, right? Like, if you okay, hey, go do this project, and you're like, by the way, I think AI helped in this. I kind of sort of don't care cuz I want them to use AI once they're inside the company, but also are they just good at prompting or can they actually critically think? Can they actually solve the problem, right? You don't know. And so, one of the aspects that tech engineering maybe is a good example. We had an engineer recently who we met, we liked, we thought that he could do the job. He'd worked at another company that we really respected, and we had a very specific feature that we wanted to build. It was basically a demo account for Silva. So, one of the problems is that people don't want to immediately just sign up and connect, you know, a million accounts, and they're like, hey, what about security, privacy, whatever, before I go and I connect my bank account or something, let me try it. And so, this demo account basically allows you to do that. Well, we just said to him, hey, you got two weeks, build it. And that's not a lot of direction. It's kind of all a figure it out, which is exactly what we would do to somebody on the team anyways, right? And so, when we did this, he had some questions and tried to get the lay of the land in terms of the infrastructure and in the code base, whatever, but he built it. Well, dude, I don't have to question whether you can work here or not. You were able to do the thing that we would have given to someone on the team anyways. And by the way, not only do we have confidence that you can work on the team, you have confidence you like us, and guess what? We didn't waste four weeks to get to that point. We just shipped a new feature, and so, we all feel great about this cuz we already have shipping velocity, and it's your introduction to the team. And so, I think that's where the world is starting to move a little bit more towards is it's kind of less credentialized, and it's much more just here's the problem, solve it. If you can solve it and you have agency, then, you know, you got a much better shot to get a job or work on something.

Do you still see a lot of people who are kind of leaning on that or do you think that a lot of the employees and candidates for jobs, they've started to kind of move away from it cuz they realize the companies are moving away from it?

No, no, no, we just tell them. We I don't think we've ever had someone in the first day where we're like, hey, it's not going to work. But definitely there's times where they end us maybe like, hey, this isn't going to work, right? And so, I think part of it is expectation setting where you're going to do a two or four-week trial and by the way, if at any point you're out or we're out, like we're just going to talk about it. And people I think just appreciate the like we're not going to you, you don't us, let's go. And it also allows for you to have kind of more direct conversations throughout the work trial as so, hey, we like this, this isn't good, this needs to improve, can you go pick up this skill to be able to finish this feature or whatever. And I think that just overall it better mimics what it's going to be like to work together, which is ultimately what you're trying to do anyways.

I think of it in the opposite direction. I always think about it in terms of if I hire this person, how much more money do we need to make in order for this to be a good hire. And so, what I try to figure out is less about hey, I make X dollars. Let's just say I have a business that makes $100 of revenue, my EBITDA is 30 bucks. If I hire this person, now my EBITDA is going to go from 30 to 20. I think about it from the perspective of, okay, they're going to cost me 10 bucks a year. Well, I make 100, can I make 130 in revenue if I add this person? And so, it's much more, how do they generate revenue versus how do they impact margins? And what that also does is there are very specific types of roles, they just don't generate revenue. They have no impact whatsoever on revenue. The number of people who work for us that are in those types of roles is very small. I don't know the exact number, but we do not have huge HR departments or even finance departments or any of that type of stuff. We look for really successful people at these jobs, and we basically say to them, you're going to kind of be like a one-person team or a two-person team in this specific function. You're going to use a lot of technology. We're going to have vendors that support you. All the things you would expect that are necessary. But, this is not a job where you're, you know, if you come in and you run finance, you're going to have 20 people. And as soon as you find the people who understand that and are excited about that, then what you find is you can use your resources and actually hire people who are revenue generating, and that is where we've seen a lot of success in terms of growing these businesses.

Yeah. But, that's the test.

Right? If they're dying, then they should be dying out of your process. Just be like, hey, this isn't the spot for you.

So, there's a couple of lessons that I think I've learned over time. The generation of the idea for Silvia was I wanted to solve my own personal problem. I explained earlier I had that spreadsheet. And to the credit of Shane Nour, who's CTO, he basically once we had this dynamic dashboard that basically I could see my net worth in real time as it updated based on prices, etc. He was like, "Well, what if we start to overlay some of the AI stuff?" And we started off where like a ChatGPT is today. It was a ChatGPT model with software around it that basically fed the context of my personal finances plus the query. And so that is the definition of a ChatGPT wrapper. That's where we started 18-ish months ago. And as we started to go and we eventually publicly launched it, we started to get users, we just obsessively talked to as many of them as we possibly could, we realized a couple of things. One, why are so many fintech products focused on budgeting and savings? Because usually the people who build them, that is the problem that they know and they understand. So they are building a product for themselves and it also happens to be the widest addressable market because everyone or most people have budgeting and savings problems, not a lot of people have lots of assets. So by us targeting a very small section of the, you know, US population or global population, basically wealthy people on the internet or self-directed or self, you know, independent investors, there's just not a lot of solutions for them. The second thing is I'm fortunate to be in that group. And so I intimately understand a lot of the problems. The example I always use here is if you think about somebody who has assets and works with let's say a private bank. The private bank wants a personal financial statement every 90 days. It's a pain in the butt to constantly keep it updated. So we simply made a one-click generation of a personal financial statement that you can then forward right to your banker. It takes 30 seconds. The regular engineer who does not understand this cohort of people, they would never even know that that's a problem let alone how to solve it. And so we think of Silvia as there's technical innovation, there is intimate understanding of this cohort of people, and then there's this third element, which is ultimately it is about memory and trust. And so, when you think of adding those three things together, we moved over 18 months from a ChatGPT wrapper to today we have basically rebuilt the entire infrastructure, and we have become one of the first public companies that is able to show owning your own intelligence is not only one possible, but two, it increases performance, it increases accuracy, and it significantly decreases cost. And so, our general thought process is like general intelligence is great for a lot of things, but in something like personal finance, the specialized workflows and owning your own intelligence is going to be the winning strategy.

Today, when you go in Silvia, there's a couple things. We have the harness that we've built. That harness is custom-built for our specific use case and our architecture. As part of that harness, we have everything from a custom-built model router all the way to the tool calling, the agents, etc. Then we have also trained what we call internally the Silvia analyst, which is a proprietary coding agent that is specifically trained on our tasks, our architecture, and our use case. We also have trained our own models, and what we can show there is that the Silvia model is more accurate than all of the frontier models when it comes to answering tax, mortgage, credit card-related topics, etc. And so, not only are we able to do things in terms of our proprietary data pipelines and kind of all the normal stuff people would try to do in terms of a use case-specific model, but also we probably have one of the largest data sets in the world where wealthy people are talking to an LLM about their personal finances. And so, we're able to create our own rubrics. We're able to do all of the RLing. We're able to actually improve this thing to the point where we don't want to be good at physics. We don't want to be good at interior design. We don't want to be good at crazy creative image generation. We just want to be the best in the world at personal finance.

Well, Anthropic engineers now are starting to enter into this group, you know, they're doing just fine for themselves. So, shout out to all of them. They're doing fantastic. In terms of use cases, so let's start with a very basic one. Everyone would like to pay less in taxes, right? You mentioned it earlier. If you go to Google, if you go to ChatGPT, etc., and you say, "Hey, how do I get my tax rate down?" It can't give you an answer because it doesn't have the context of your full finances. But it's not just, "Oh, connect it via Plaid to your bank, brokerage, credit cards, all that stuff." It's also things like, "Well, what real estate do you own? What are the trusts? How many dependents do you have? What city do you live in? What is the tax code?" All of this stuff goes into answering the question of how do I get my tax rate down? And I'll give you a specific example. There was a user on Silvia who went and said, "How do I get my tax rate down?" And what Silvia said back to them was, "You have high income that has been persistently high for many years. You have at least one entity where you are the sole owner and the sole employee. And so, for you specifically, what you should do is you should set up a defined benefit plan where you are the sole beneficiary of that plan." Saved that person hundreds of thousands of dollars a year in taxes if they implement it. But you need the context of not only income, legal entities, where they live, right? How many dependents, all of that stuff goes into answering that question. So, in that case, it's actually a pretty simple question. Everyone would like to get lower taxes, but it's the technical architecture of the context. Now, fast forward all the way to the other end. One of the things that young people in particular don't think a lot about is not only what happens when I die and my kids or my relatives need to know where my assets are, but something like an instruction manual. So for example, I have a friend, he built a custom home about a year and a half ago. And I was talking to him and he basically says, "I left my wife some instructions of what to do if anything ever happened to me." And I said to him, "Why'd you do that?" And he said, "Well, because my wife hates debt. And the first thing she'll do if something happens to me is she'll pay off our mortgage." But he had like a 2% mortgage. So he was like, "The last thing she should do is pay off the mortgage." But he goes, "I know my wife. I know she'll do that." So what we see people doing inside of Silvia is they will literally attach all of their assets and they will at times even use it as almost like a diary. Here's why I'm buying this asset. Here's what I think's going to happen in the future. Here's how you should think about this if you take over this portfolio. And because we've built an entire custom memory product, we literally had to go and rebuild a genetic memory and this whole file system, etc. What it allows us to do is empower the user now to go and actually start thinking in generation. And when you do that, you really start to change the way that people think about these fintech products. It's not a product for trading. This is a product that allows you to think super long-term around estate planning, taxes, those types of things. And that's what we think is very different about this group.