#1082 Samir Vasavada The 22 Year Old Genius Who A Built $1 Billion Company

What's up everyone? This is Anthony Pompliano. Most of you know me as Pomp. You're listening to the Pomp Podcast, simply the best podcast out there. Now let's kick this thing off. Samir Vasavada is the co-founder and CEO of Vise. In this conversation, we talk about artificial intelligence, Vise, direct indexing, personalized portfolios, OKRs, how to run certain meetings, how to actually operate with a board of directors, why Kanye West may be the most brilliant philosopher of our generation, and what exactly the wealth 3.0 movement is really about. I really enjoyed this conversation with Samir, and I hope that you guys enjoy it as well. Once you get done actually listening to this episode, let me know on Twitter or in the comments what you think. All right, let's get into the episode. Hope you guys enjoy it.

Anthony Pompliano runs Pomp Investments.

All views of him and the guests on his podcast are solely their opinions and do not reflect the opinions of Pomp Investments.

You should not treat any opinion expressed by Pomp or his guests as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of his personal opinion. This podcast is for informational purposes only.

All right guys, bang, bang. I've got Samir here with me. I'm very excited to talk to you. You started a company when you were very young. You've scaled it at an incredible pace, but I thought a great place to start would actually be with a tweet that I dug up from the past, which was Kanye West is the most brilliant philosopher of our lifetime. I tend to agree, but I'd love to hear your thought process as to why Kanye is so philosophically superior to others maybe that are living right now.

Yeah, I mean, I haven't been tweeting a lot lately, so maybe that one stands out for sure. You know, I think I was listening to a Kanye song and I was just super inspired. I think there was this quote of his of like, you know, "Not having money is everything, but having money isn't everything." And I was like, hmm. Pretty insightful.

Yeah, well, that's part of, I think, most good songs is like, my brothers and I always joke like bars, right? If a good tweet is essentially just like a one-liner, whether that's a standup joke, whether that is some kind of lyric from a music song, like ultimately the ability to communicate complex ideas very simply and in short form is a pretty powerful concept today.

Yeah, and I think he's incredible at doing that. And I think like some of the greatest philosophers of our time are also crazy people. And I think Kanye is a crazy person, and crazy people have different ways different views on the world that, you know, can sometimes be reflected in—

How do you know if he's crazy or if he's right?

You know, I don't know, but I think I've heard some stories from some friends that know him that seem absolutely crazy that, you know, even the most right person would not do. So I don't know.

Well, it begs the question, like, if you start especially thinking of like startup founders, like, can you be crazy and right? Right, like to some degree, I always think about like, it's crazy until it works, right? So when people sit down and they're like, I'm gonna do X, everyone's like, yeah, yeah, yeah, okay, crazy person. And then all of a sudden you accomplish it, then it's like, oh, you were right.

Yeah, exactly.

But at the same time, what's his line? I think like, show me a genius that's not crazy. Yeah. And so it begs the question of like, how do you label when somebody's doing something different? Ultimately it's just, are they right or not?

Look, if you're gonna be a startup founder, you have to be crazy. And I think the startup founders that fail, right, sometimes fail in really meaningful ways. Like, if you look at, like, Elizabeth Holmes, and yes, like, she committed fraud, but, like, if somehow Theranos worked, everyone would make her look out like a genius, right, even if there was the fraud behind it, right? Or Adam Neumann, or any of these other people. You know, they're all crazy. Every founder is crazy. It's just the ones that are successful are the ones that write the history books, not the ones that fail.

Yeah, and when you think about that, where do you see that showing up the most? Like when you say crazy, is it the speed at which they wanna build the company? Is it their vision of doing something so disruptive to a market? Like how do you measure whether a founder, if you were to sit down and talk to them, is crazy or, you know, maybe they don't have that it component to be a founder?

So I think it always starts with a vision, right? If you are a founder and you believe you can disrupt an industry that like has probably done the same thing the same way for the last, I don't know how many decades, you are objectively crazy to think that you can take that on because all of the odds, the whole deck is stacked against you. So you have to be crazy in that sense. I think the other side that it shows, where you can look at any of these founders like Steve Jobs or, you know, others that like, they are just crazy in how they operate, right? They have extremely high bar, like for talent, for expectations from others, and they expect everyone to operate at a certain level that, you know, people just generally don't tend to operate if you're not a non-founder. So they have those expectations for the people they work with. Like, they can be difficult people.

Yeah. How much does age play into it, right? Like, you're a young person who, I think you started Vise when you were 16. You're how old now?

I'm 22.

22 now. So for 6 years, right, you've been running this business. 22 is still very young compared to most people in the workforce in kind of the quote-unquote traditional market. Do people lose the craziness as they get older, or do the special people just keep it no matter the age?

So what happens is you get a little bit more jaded as you get older. So when you're young, you have this like certain naivety about you that like you believe because you don't know how hard something is going to be that you can take it on. So like when we first started, so my co-founder and I met at the age of 12 at summer camp. And—

Which summer camp?

Our parents are like, oh, you cannot do normal summer camp. You need to do college courses. So we did college courses at Northwestern.

Okay.

And his dorm was across from mine. And the two of us—

12?

Yeah, 12.

Okay.

And the two of us, you know, we started an app development firm. So we started building apps for small, medium-sized businesses. But, you know, when we came out of that, like, our idea was like, how can we use AI to build apps? Like, what if you could just, what if anyone could just type in an app idea and our system will build you an app using artificial intelligence? That is a really, really, really hard problem. And, you know, we didn't know that we couldn't solve it at the time and we couldn't end up solving it. But when you're really young, you believe you can do things that you didn't know were possible. And sometimes you can make the impossible happen. And just because you didn't know that that could happen is why it happens. Mm-hmm.

And when you think of kind of this idea of, I just want to dictate to a machine to create something, if we were talking 10 years ago, like, that sounds insane, right? Now we're starting to see with GPT-3, maybe with DALL·E, like, a couple of these services, you can't maybe go create the full application that's super complex that would serve all the needs of a user. I've seen some people do, like, super simple things simple things like, you know, it'll auto-click something, uh, or whatever. But also we're getting, uh, hey, write me an essay, or hey, create this piece of art out of just a couple of words that I dictate to the machine. And so does it feel like the path we're on, eventually we will be able to solve that problem where somebody could just describe an application and build it?

Yeah, I mean, I think the limiting reagent has always been like, it's a data problem. Okay, right. And I think the challenge was like chips weren't designed previously to be able to, you know, process nearly as much data, like the training datasets you need to build, let's say, an app or write an essay or something like that, just previously were not possible, right? And for now, like, especially as you look at the evolution, Moore's Law, right? The data we're able to process, right? The data GPT-3, for example, is able to process is truly incredible. And the things that that thing will be able to create over the next 5 to 10 years are gonna be pretty endless.

Talk, for those that don't know how this works, right? Like, so you get a dataset, where are they getting the datasets from? And then how are they actually training these models on those datasets so that when I go and I use it and I just type in, you know, hey, write me a paper on something, it just kicks it out and I think it's magic. But like, what is the work that went into actually creating the quote unquote artificial intelligence that I'm using?

So for like a Grammarly or something like that, you are collecting lots and lots of data from, like, you know, preexisting, like, content, right? Like, either, like, broken-down sentences or, like, what it looks like to be correct. The challenge is, like, data labeling, right? So my buddy at Scale AI created this really great company to kind of help solve this problem. But it's like, how do you train the model on what's actually right and what it should pattern match? Like, if you think about any of these AI models, they're just functionally pattern matching. So, like, how do you pattern match to what is right versus what is wrong that the algorithm can pattern match to? So you need to, the challenge in this problem is how do you effectively label these datasets? And the way people label the datasets is they have, you know, teams of people in like India or, you know, really low-cost places where they're manually labeling these datasets. And then those datasets labeled are gonna be fed to these algorithms. And when that problem is solved in some kind of automated way, which I think we're still kind of far from, but Scale has definitely made a lot of progress towards doing that, you know, you're gonna see a lot more progress on, you know, broad AI.

So a couple of things that jump to mind, Mechanical Turk, right? Obviously Amazon has this service where you basically can put in some data, it will get labeled essentially for you, and then you can get it back and you can do quality control or whatever. But for the most part, that is one thing that you could do. Another is a lot of the CAPTCHAs now will show you, you know, 6 or 7 photos and it's like, hey, click for, click the 3 that have the sidewalk, right? Or the chimney or a school bus or a bicycle or whatever. Uh, and in some way you're getting people on the internet to actually train the data for you, right?

Exactly. Or like in, you know, in the instance of Tesla, I think like what's brilliant about their self-driving car is like they, you know, you've got all these self-driving car companies that have to send cars out on the road and like, you know, have engineers go out and sit with them and, you know, drive them and collect data. Whereas Tesla has the data of every Tesla out there that's going out and collecting self-driving car data.

And is it the idea, like Tesla as an example, the compounding nature of every single day they're basically collecting almost as much data probably as Google's collected, you know, in its lifetime? And I don't know if that's exactly correct, but definitely directionally. And so if you do that for a decade, Tesla just has such a mountain of data and so many more insights than any of the other companies, and therefore that advantage is almost insurmountable?

Exactly, right? You can use your scale. Like, that's what the, genius of some of these big platforms are, is they're able to use their scale, leverage that to collect data, and then that data can kind of improve their products and kind of unlock things that they couldn't do before.

You mentioned Scale AI. What are they doing?

So Scale AI is like a data labeling company, right? So they're building, you know, technology basically augmented by artificial intelligence to help like label datasets.

Got it. And the datasets are just anything somebody can feed them, or do they use specific types of datasets?

So I think it started with self-driving cars. Now they're like working with the military. It's different types of datasets. So you could feed them a dataset and theoretically they could label it.

Got it. And so when we think about this labeling of data and then the ability to extract insights and then apply it for whatever the use case is, one thing I always hear my nerdiest friends say is like, "Artificial intelligence doesn't exist, it's machine learning." When you hear that, like talk through a little bit of the nuance between like machine learning and artificial intelligence.

Artificial intelligence is an umbrella term, right? It always has been. Like maybe you can conflate like AGI, artificial general intelligence, with like the broad term of AI, But like artificial intelligence is a general term that can kind of encapsulate lots of different things, machine learning being one of them. And like where we've seen the most progress, and I'm a little rusty on this because we've kind of strayed away from the AI focus, which I'll talk about in a little bit, is like machine learning, right? So narrow intelligence. So you've got AI applied to one problem set, one kind of narrow solution that you are applying kind of some kind of machine learning algorithm to.

So once you have a dataset and it's labeled, talk to me about the machine learning itself. Like, what is happening there in terms of the insights or the applications people are able to use with this machine learning on a labeled dataset?

So you have a recurrent neural net, or like some kind of neural network, and the neural network has nodes, and it's able to draw conclusions and basically draw a pattern. So given the set of data, this is what we expect to happen. Not necessarily an algorithm where it's, just a set of instructions, basically. So it is a black box in some ways, because we're trying to interpret a pattern given this broad set of data. And the better data you have, the better pattern recognition you're fundamentally gonna get.

[Speaker] So in the most basic sense, like, you could think of, hey, I've got a set of data, I've already kind of massaged it and helped synthesize it to some degree via the labeling and all this stuff, I throw it to a machine learning model which has these neural nets in it, and all I'm really saying is like, yo, find the patterns. And then the machine's like, cool, here's the patterns, and then I can go do whatever I want with those patterns.

Yeah, and your model's only as good as the data you feed into it, right? And the challenge is oftentimes explainability, right? So when you think about financial models, right, like in our instance, we sell to wealth managers, right? Wealth managers want to understand the why behind their trades, right? Why was this decision made, right? And when we started, we built a pure AI model. And I think the challenge was, you know, we couldn't, you know, the model performed very, very well, right? You couldn't explain, you know, here's how this algorithm jumped to the conclusion of why we made this set of trades. We were looking at a set of data, we were trying to find the patterns of that data, and then we, you know, had an output, and we didn't know how the model basically came to that output.

And was the problem in that scenario, like, people, they're like, I don't trust the box, like, I don't trust, I don't trust the computer, I need to understand what the computer's doing in order to put my trust in the conclusion, or what was the— the kind of limiting factor there?

So generally speaking, I think people just trust the box. Like, in most problems, with a finance problem, people, when they're giving you all your money, um, all their money, they want to understand how did the algorithm make decisions. But it was like split, right? Some people are like, look, this is the future, like, we're going to trust this algorithm. And others were saying, you know, I want to understand how the decisions are made. So I think the reality is, is you have to get somewhere in the middle.

Yeah, I always think of like Google Maps, right? Which again is just looking at streets and trying to figure out how do you move from point A to point B. Would people follow the directions if they only showed you the next step? Like if there wasn't actually the line on the road that showed you, okay, you're here, this is how you're gonna get there. If they just said go right and you didn't know what was coming after that, like there is an element of humans being like, ah, should I go right or what, you know, where are you taking me? By being able to see the full route, before you kind of hit, you know, start, I think people are like, "Okay, cool, like, yes, that does get to the end." Or like, "Is this the best path to getting there?

Like, what if there's a better path to getting there, right?

How does Google Maps know?" Mm-hmm, and do you think that consumer behavior has to change, like consumer psychology? Obviously, Google search results, right? There's some machine learning that's going on there and kind of learning what your preferences are, what's the best information out there, helping to kind of surface the best search results. If I think of music recommendations, and if I'm using Spotify or iTunes or something like that, and they go ahead and they tell me what's the next song that they think I'll like. We've talked a little bit about investing and the ability for analysis of a dataset and then kick out kind of a conclusion. Like, there's plenty of people who are trusting this stuff today. What's holding back the rest? Is it just time and the changing of their consumer psychology?

So I think about it in a couple of ways. You need, like, for mainstream crypto adoption, you need consumer behavior to change, right? They are doing something one way. And now they have to do something a completely different way, right? So that's gonna take a long time and it requires like a major shift in psychology. However, for AI or for machine learning or whatever you wanna call it, it's happening in the background and it's progressively getting better and better every day. And it's just happening as we kind of run through our day-to-day lives. So there's either people that are using technology or they're not using technology. And if you are using technology, it is going to start to have an incremental advantage over your life every single day.

How important are optimizing for specified outcomes? So in the example that you gave earlier where you're like, oh, we thought we could type something in and eventually we would just like create an application. Obviously, it's pretty important if you want to build something that looks more like Instagram versus maybe an enterprise software tool, right? You'd have to tell the machine, I want X, and then it would figure out how to go do that. When it comes to machine learning, do you need to know what insight you're looking for, or is there the ability to just serve it the dataset and say like, hey, you just come back to me with whatever patterns you find. I actually don't know what I'm looking for yet.

It's an interesting question. It depends. So I think the way we think about it is that problem, for example, is a very broad problem. So the algorithm wouldn't necessarily be able to get you, like, the perfect app, right? The algorithm would not be able to tell you, like, you know, based on what you're thinking in your head, this is what the app you want look like. That's why it was a really hard problem to solve. But for specific instances, right? So for example, I'm trying to, like, I'm trying to do stock price performance and I'm looking at all of the historic price-to-earnings on a particular stock, right? And I want to understand, like, given all of the previous historicals of the stock, what do I think is a likely kind of next outcome? It's very precise. What do I think the, like, forward-looking price-to-earnings might be? That is a clear scoped problem that you can task the algorithm on. But, like, having something that's very creative, very broad, is a very difficult problem to solve.

And when you think about it in the investing framework? Is it rather than like, hey, pick a stock, instead you can say like, which is the one that is likely to go up the most by the end of the year? Or pick the thing that is likely to go down the least in, you know, a certain type of environment? Like, how do you think about this machine learning, artificial intelligence approach to investing?

I think it's less broad even than that.

Okay.

Like, rather like on the stock level, it's more on the specific indicators, right? So like sentiment is a great example of this. And we don't necessarily know if sentiment is a leading or a lagging indicator. So, like, is a stock with good sentiment going to go up? Is a stock with bad sentiment going to go down? But the way we think about it as part of our models is, you know, stocks that have, you know, call it high or low sentiment, could just be a good indicator not to buy. So you take an example like a Valiant Pharmaceuticals, right? This company took a huge hit because, you know, fraud, all kinds of horrible stuff going on. And, you know, they still did meaningful revenues, right? Any kind of value stock screener would say, OK, Valiant Pharmaceuticals makes sense as a purchase. But any investor, any institutional investor that knew the stock would say, I wouldn't touch this company with a 10-foot pole. So what this can tell us is, like, look, you know, the fundamentals of the company might be good, but the sentiment of that company is pretty bad. We should just stay away from that company rather than using it as an indicator to buy or sell. The way we think about it, though, is that sentiment is like sentiment is a clearly scoped problem where you can use artificial intelligence or machine learning to understand what is the stock sentiment, right? So we can look at tweets, we can look at news articles, we can look at analyst commentary, we can aggregate those various datasets, we can scrub that data with relative ease, and then we can, you know, create a sentiment score.

And when you do that, how accurate are these types of models? So let's just stick on sentiment, for example. If I gave you two stocks, could you come, you know, pretty close? Is it something that like, oh, we're always improving, but like, I still don't have a ton of confidence in actual sentiment like measurement. How do you think about accuracy?

I would say it's very accurate when you think about is something high or low sentiment, right? So a company that is getting, you know, bashed in the press, for example, our model is going to say that is a low sentiment stock. A company that is getting lots of great positive commentary is probably going to be a high sentiment stock. The challenge is what do you do with that data, right? How do you know whether or not there's something that you can trade on that stock or not. There's a lot of people that have tried. They've created hedge funds around this. How do we try trading on sentiment? How do we trade on stock tweets? But it's a really hard problem, because there's no real answer as to, does sentiment move a stock up or down?

[Speaker:SAMIR_VASAVADA] So sentiment's a thing where if I go and I read the news, I probably can come to the same conclusion as the computer, right? Or if I'm on Twitter, or if I'm in forums, or kind of whatever your— whatever the means of content ends up being, like, a human likely will come to the same conclusion on sentiment. Are there things where humans have a really hard time coming to the same conclusion as the computer that you guys look at?

The challenge when you think about holistic portfolio construction is you're looking at so many different types of assets, right? So a human is going to be able to do a similar to just as good of a job looking at one particular stock, right? We might be able to look at way more data points, and do it faster. But if you're thinking about building a holistic portfolio, you need to be looking at tons of data points for tons of different stocks, you know, hundreds of stocks potentially. And a human just, you know, can't do that nearly as well as a computer can with the speed and the precision that a computer can do it.

Makes sense. And also, the computer doesn't sleep, the computer doesn't get sick, right? All these things that I think kind of older folks usually joke about the computer, but like they are true to a degree. And so when you think about kind of the market that you're going after today, how much of the finance world is already using some of these technologies and kind of up to speed on what I think you and I would look at as like no-brainers in the future, like these technologies will be used, versus it's still being done the same way it was, you know, 10, 20, 30 years ago?

Very far behind.

Okay.

And everyone knows it needs to change, But there's a couple problems. First is switching costs. So when you think about some of these large institutions that have built software, there's so much tech debt that's continuously stacked up on software that is like 20+ years old, you know, trading systems, portfolio management systems, software that it will cost these institutions billions of dollars, like hundreds of millions, billions of dollars to move away from onto something new. And a lot of those institutions just don't want to go through that pain. They don't want to go through that hassle. It's part of the reason why IBM is such a big business, right? Because someone has to maintain all of this. But that's going to slowly start to erode over time. Because what's going to happen is technology right now might not hold one of these institutions back, but in 5 years, in 7 years, in 10 years, as people start to do more things on their own through technology, these institutions are going to have to start to change. But the time it's going to take them to change, the time it's going to take them to evolve is going to be too far gone.

We interrupt this podcast to tell you about the most anticipated crypto event of the year. Messari's Mainnet Conference 2022 is happening this year, September 21st to the 23rd in New York City. It's 3 days of can't-be-missed keynotes, fireside chats, demos, networking, and more. They'll have folks like Ryan Selkis of Messari interviewing Ripple's Brad Garlinghouse, and that one is going to be absolute fireworks. They also have keynotes from Balaji Srinivasan and OpenSea's Devin Finzer. You can get $300 off if you go to mainnet.events. Again, just type in www.mainnet.events, use promo code POMP, and you'll get $300 off. I'll see you this fall at Mainnet 2022. This episode is brought to you by Alto IRA. They can help you invest in Bitcoin and crypto in a tax-advantaged way that helps you preserve your hard-earned money. Alto's Crypto IRA lets you invest in Bitcoin and over 200 other different coins and tokens, and it has all the same tax advantages of your traditional IRA. There's no setup or account fees, and it's all you need to do— invest in crypto tax-free. Let me repeat that again: you can invest in Bitcoin and cryptocurrencies tax-free. So, are you ready to take your investments to the next level, diversify like the pros, and trade without tax headaches? Open an Alto Crypto IRA to invest in Bitcoin and crypto tax-free. Go to altoira.com/pomp. That's altoira.com/pomp. Start investing today. This episode is brought to you by FTX US. They're the safe, regulated way to buy and sell Bitcoin and other digital assets. Trade crypto with up to 85% lower fees than the top competitors. There are no fixed minimums, no ACH transaction fees, and no withdrawal fees. Download the FTX app today and use referral code POMP to earn free crypto on every trade over $10. The more you trade, the more you earn. Go download the FTX app today and use referral code POMP. Taking this as kind of like a viewpoint that you have today, 6 years after starting the company, take me back to when you started the company. So you meet your co-founder when you're 12 at— I'm going to call it intelligent camp because unfortunately my parents were not forward-thinking enough to send me to a, to a camp like that. Between 12 and 16, you guys were still working on things together or just like staying in touch, or what was your relationship between when you met and then you actually started the company?

Yeah, so like the specific story is, it's pretty funny. So I'm from Cleveland, Ohio. My co-founder's from Detroit, Michigan.

Okay.

And, you know, very traditional families, very traditional backgrounds. And when we got back from camp, we're like, well, we stayed friends and we're like, well, we wanna make some money, right? Like, school's easy, like we should make some money. So we were like—

That's why we're different, by the way, is 'cause you said school was easy.

Like, we should make some money. And the way we think, you know, we should make money is by building apps. You know, it started as an iPhone game we built, then we started building apps for small businesses. So I created a little webpage and posted a press release, actually, to get my first couple of apps. And small business owners reached out. We had, like, the Canadian Craigslist was one that reached out. We had, like, a gas station. We had a couple conferences, conferences like Zaps. And we were able to start building apps for all these people. Thousands of dollars.

How much were you charging?

We were charging different prices. It was like, honestly, I wish I had some kind of pricing formula, but we didn't. We just, whatever we thought we could get the person to pay is what we would charge them. So anywhere between kind of $5,000 an app all the way up to like $20,000, $30,000 an app.

That is usually how most young people start with pricing is like, do I think they'll say yes to $1,000, right? Like, oh, they said yes, damn, I should ask for $2,000, right? You just kind of work your way up until somebody starts telling you no.

And you know what's funny? I was like, you know, the amount of value we were able to provide for the little cost compared to like, you know, some of the bills I see at Vise for specific things, I'm like, wow, these people got a lot of value out of us for very little cost. But we started building apps and, you know, Roonik, my co-founder, is this math prodigy, and he was one of the brightest math minds in the country. And a professor reaches out from a very notable institution, you know, asking, hey, you know, I run the artificial intelligence lab here. Can you kind of join? You know, you're a math researcher, and math research translates very well over to AI research. Can you join, and can you start, you know, helping me solve some of these problems? And he's working on restricted Boltzmann machines, and it's a type of artificial intelligence. And we were wondering, like, you know, all these small businesses want apps. It's 2013, 2014. They all want apps. This is becoming the hottest thing. Like, every business was expected to have an app. Like, what if we built the software? We called it Syscat. That you could type in your app idea and our system would build you the app. And we spent all of the money we made on building apps, you know, bootstrapping it to build this, you know, company. We were like, we're not gonna be able to raise any venture financing. We're gonna find some people on the internet and they're gonna be our first engineers. And we just went on all kinds of different web forums to try and meet people. We ended up meeting this engineer who actually unfortunately had to— hacked into Sony Media as part of the hacks and went to jail, so that was unfortunate, but a number of different engineers, and they helped us bootstrap the first initial version of the product. But we were running out of money. Like, we didn't think we could solve the technical problem. It was just too complicated of a tool to solve. At this point in time, my grades in school aren't that great because I'm spending all of my time building the startup. But we had met this guy who was a former investment banker, and he's like, look, you know, even though this company might not be working out, You guys know so much about AI and machine learning. There are all these consulting opportunities through these expert networks and working with these financial institutions specifically. You can consult on how AI works and teach financial institutions how AI works. So we're like, that sounds great. So we started doing that and we're charging $700 an hour to meet with financial institutions and educate them on how artificial intelligence works. And these are like big bulge bracket banks. So, you know, MassMutual, RBC Royal, Deutsche Asset Management. Big firms, and we would get on the phone with managing directors, analysts, whomever, and talk through artificial intelligence, talk through the latest and greatest technologies. But what started to happen is that we were talking to different people in these organizations, and the wealth management divisions kept on coming up. The sense that we have these massive wealth management groups, 50% of Morgan Stanley's revenue, 30% of Goldman Sachs' revenue, is like massive scale to these wealth management orgs. But our wealth managers, you know, we believed at the time were great money managers, but in reality, they were good relationship managers. And it was why they weren't going away. It was why robo-advisors weren't going to replace them because they had this relationship they could keep with the client. But all these institutions were trying to understand how can we leverage AI, machine learning to be able to make our wealth managers smarter, to be able to help them make better investment decisions, make better portfolio decisions. So we start to think about this problem. And we do a couple projects with, you know, large banks, and we realize, what if we just build this on our own? What if we build our own software, portfolio optimization software using machine learning, and we sell it to large institutions?

And what were they trying to get at? They were trying to get at getting a better return for their clients, at like tax loss harvesting, something else? What were they optimizing for?

So at the time, what I think what they were optimizing for and what we were optimizing for was slightly different. They were optimizing for how can we help tell the story to our wealth managers? How can we help, you know, how do we have one kind of platform that can deliver insights on clients' portfolios that can help them recommend investments all powered by AI? What we were thinking about is like all of these institutions have all these portfolio managers. They don't need to have portfolio managers. An algorithm can do their job just as well. And the advisor themselves can work directly with the algorithm, can work directly with the platform, and you can cut the portfolio manager out.

So basically, there's a bank, let's call it, right? You know, ABC Bank. That bank has the financial advisor. Now, the advisor in some capacity is really doing customer relationship, right? It's managing the relationship. People do not want to talk to a computer. They want to talk to a person. They're able to sell, but they're also able to do client retention and those types of services. But a lot of times the financial advisor is not the one actually making the investments. They're simply just saying, okay, you know, uh, client, uh, uh, Joe, why don't you take, uh, 10% of your assets and put it into our, uh, public market, you know, whatever vertical-focused, uh, fund. Oh, take another 10% of your assets and put it in this other fund or whatever. And then there's portfolio managers for those funds that are actually managing the capital.

Exactly.

And that's where you guys thought that you could basically rip those folks out and replace them with technology.

Exactly. Sorry if there's any portfolio managers that listen to the podcast.

There's a lot. Don't worry.

But the idea was like, you know, there's all of this, you know, all of these people in these institutions that are delivering relatively generic investment solutions, and we can replace it with one kind of piece of software that can deliver highly personalized investment solutions to each individual client. But at the time, we were thinking more at the fund level. So we were trying to think like, okay, given a strategy that one of these places might have, they might sell through their advisors, how can we just automate that strategy? We called that tool FSAI, Financial Services Artificial Intelligence. The two of us built it ourselves.

And how long did it take?

It took us 6 months. And the funny thing was, there was a state stock market challenge going on, and I used the algorithm to allocate my portfolio in the State Stock Market Challenge. Did you win? And I came in second place, so.

Interesting. Who was the winner?

I got $1,000 out of some kid from some other school.

And he was just licking his finger and sticking it in the air and picking stocks?

Must have been. I don't know.

That or inside trading. Yeah. Right? Who knows what he was doing?

But either way, the idea back then was returns. It was performance. But there was a school trip to Detroit Startup Week. And Jamie Dimon had this big initiative in Detroit, and he was speaking on stage, and he gets off stage. We rush him on stage, and we're like, Jamie, this is what we're working on, FSAI. Do you think that JPMorgan would buy this? And he's like, look, JPMorgan has a lot of tools. I don't think JPMorgan's gonna buy this. You know, it was a whole discussion. And, uh—

It was a long discussion?

No, for 5 minutes.

5 minutes, yeah.

But we were like, well, shit. Like, you know, if JPMorgan's not gonna use this, and, you know, we were young, and They're not going to use this. Like, what do we do? So we ended up meeting this financial advisor in Philadelphia. And he had this $300 million independent RIA practice. And he said to us, advisors are leaving big institutions at such a rapid rate. They're all going independent. Advisors are leaving the big bulge bracket institutions, and they are taking their books of business with them. They're starting up their own practices. And they have very little technology that can help them build and manage portfolios. They don't really have an operating system around their investments. So they would eat this up, they would dig it. So we were like, okay, we're going to build the software for independent advisors. We're almost going to build the Shopify-like operating system for independent RIAs to help them build and manage portfolios. So every single day after school, we would talk to like 5 to 8 financial advisors. We would post these job ads and just have lots and lots of advisors apply to the job ads. And it was our growth-hacked way of meeting advisors and doing customer research. And the way we thought about it was, how do we understand the psychology of these advisors? Right? Most of these people in the space are just building investment products. They're not actually trying to enable advisors to scale their business and, like, make them better at their jobs. How do we understand the psychology of the advisor and then sell into that psychology? So we spent all of this time talking to advisors. We had hundreds of conversations with advisors, and we realized there was something there. And one of those people we hired was a former advisor, and he loved the product, he loved the vision that we were going after. He's like, okay, I'll be your kind of first employee, and I will build this company with you guys. And he had sold his RIA practice, he could afford to work on equity. We were using the money from consulting to kind of bootstrap the business early days. And we met these two PhD machine learning AI quants, and one worked at Morgan Stanley, the other one worked at Google. And they were the early team that basically helped us build—

Where do you meet those people?

The first version of Eyes. AngelList, actually.

Interesting.

So some, Roonick actually worked with in the past. Others, we were— there were resources that were available that just weren't utilized very, very well by startups back in, you know, 2015, like AngelList and other job sites that now are probably inundated, where there was a lot of really high-quality talent that wanted to just work on startups with startup problems. That was a great place to recruit from.

And was the team all together when you first started, or was it remote?

Everyone, it was actually, we all started remote. I was running it out of my parents' house in Cleveland, and so was my co-founder in Detroit. We were in high school at the time. And then the team was actually all, for the most part, based on the West Coast in San Francisco, which is an interesting part of the story, because what happens is I realize that this is becoming such a core part of my life, and this is what, Some people have a passion, others have a calling. Like, this was my calling. This is the thing I wanted to spend the rest of my life on. And I decided, you know what, screw school. I'm gonna drop out and I'm gonna go move to San Francisco. So I moved to San Francisco. I lived in the Tenderloin, which was not too fun of an experience, but it was all I could afford at the time. And we had a WeWork and we couldn't even afford the WeWork membership, so we just snuck— we had one membership and we snuck all people in every day and no one seemed to notice. And my co-founder at the time was still in high school. And his parents said, you know, you have to go to college. And, you know, at this point, like, my parents don't like me because I left school.

So your parents were pissed that you left?

They were pissed. They ended up coming around to it, but they were upset.

And your parents are Indian?

Yes, they are Indian.

And are they traditional Indian parents who wanted you to be a doctor or a lawyer?

Yes, doctor or engineer.

Yeah, I didn't know this whole thing. And I've got a couple of friends who are Indian and like, The joke is always like, oh, Indian parents, whenever something happens in the news, they like text their kids and like, why can't you be like, you know, whatever they see in the news?

Oh my gosh. Is it real? There's so many people that we were compared to. You know, my co-founder has this huge chip on his shoulder because his cousin is like this brilliant kid who like went to Harvard and got a job at Palantir. And his parents are always like, oh, you know, Harsh did this big thing very well, you know, live up to that. So every Indian kid has this chip on their shoulder. And, like, I definitely had one. I think, like, a lot— like, they're very, very comparative.

Yeah. How old were you when you moved to San Francisco?

I was 16.

16.

So I just about turned 17.

And did you know anyone in San Francisco other than the people who were working at the company?

I knew the people that were working at the company. And I cold emailed and cold LinkedIn messaged a couple of people to, like, meet my first group of people.

And did they think you were crazy?

Oh, everyone thought I was crazy. But I'm living in the Tenderloin, and my co-founder gets an invite to this barbecue at this place called the Crypto Castle. I'm not sure if you've heard of it.

Jeremy Gardner.

Jeremy Gardner. So Jeremy Gardner says, hey, look, I've got this bunk bedroom. You can live here. So I lived in the Crypto Castle for about 2 years, and I saw the Bitcoin run-up, which was a pretty cool thing. And it was kind of my initial launchpad. In San Francisco, and I was around a lot of young—

Lots of people coming in and out of there, right?

Exactly.

And a lot of people building companies, all of that. When you actually quit high school, you didn't graduate.

I didn't graduate.

OK.

But I did some community college credits. OK. And I did these summer college classes, so I was able to get enough credits at the end of the day to get a diploma, like, a couple of years later.

Got it. And so when you do that— Is this like a burn the boats moment? Like, if you're leaving, there's no like coming back home and be like, "Mom and Dad, you were right. Like, I should have gone to college and like, I apologize." Or is this like, "No, like I gotta make this work 'cause I'm definitely not going back." Oh, I had no choice.

Like, I think that entrepreneurs, when they have a set of, like, when there's desperation, right? When they have like no other choice, when you have no other optionality, you are forced to make it work. And that's the situation I was in. I was forced to make it work. No going back. And it was like more, like, it was my passion, it was the thing I wanted to spend my life on, but like, I had no alternatives, right? Like, there was nothing waiting for me on the other side, and I had to do it with very little money.

So at 16, you start to run the company, right? What are some of the things that, looking back now, like, what are some of the mistakes that you made where you're like, "I was 16, like, I just didn't know"?

I mean, I would say every mistake is probably what ended up happening. I mean, I would say I probably have better learnings given the period after we were funded and after we started to have scale about what I would've done differently. But when I was 16, I was trying to build a company, I was doing whatever I could, and I was trying to be as scrappy as possible. So I think the lesson, you know, going forward that I think I did well was I stuck it out, right? I had like a semblance of grit and like I didn't give up. And I think in order to be an entrepreneur, and like you can see it in this market environment, you can see it, just broadly speaking, if you quit, right, because something is hard, because there's not something proven, whatever it might be, because it's the easy way out, like, you're never gonna find, like, a meaningful amount of success. And I think that if you continue to work at something and you continue to iterate, success is gonna start to compound over time. It's gonna continuously get easier, and you're gonna learn more, and that learning is gonna continue to compound over time.

What was fundraising like initially?

So— What happened with fundraising was, so my co-founder decides to go to college, and I'm like, shit, my co-founder's going to college, what am I gonna do? So I decided to move onto his dorm room floor. He has this little mattress topper, it's 2 inches thick. We put it on the floor, he's single, and I just sleep on that. And I'm like, Roonick, if you drop out of school, what needs to be true? And he's like, look, if we raise $1 million, I can convince my parents to drop out of school. So I'm like, sure, like, let's go out and raise some money. And originally, we actually tried to raise money, and I had just made cold calls and cold emails. I cold emailed 100 people, and I would just continuously email them and no one would ever reply. So I was like, okay, well, the only way that I can get fundraising to work is if we have some kind of social proof, if someone like introduces us to someone. But like, at this point, we still don't really know anyone. So we're at UPenn. We start using the UPenn resources. There are some investors, there's this guy named Josh Kopelman who went to UPenn, who's an incredible VC from First Round Capital, and he had introduced us to these two guys, Nat Turner and Zach Weinberg, who founded Flatiron Health. And they said, look, we don't know if this whole RIA wealth management technology thing is gonna work, but we like you guys, so we'll give you your first $100,000. And it came in perfect timing because the employees, they were working for had, you know, basically no salary for 2 years, and they were finally starting to get pressure from their spouses to, like, you know, take some cash.

Why are you working for a 16-year-old for no money?

You know, it's funny, because we didn't even tell them our age at the time. Yep. But they asked for cash, and, you know, we were, like, about to call it quits. We were like, look, if this doesn't work by the end of the year, if we aren't able to raise capital— and it was probably February, it was the lowest point they'd ever been. I was very depressed. I was like, if we can't make this work, if we can't raise capital, like, we'll call it quits. Like, we'll stop the company. And we ended up getting that first check. I was the most excited I'd ever been. And I knew I could, like, finally pay to, you know, cover operations and pay for salaries. And Nat and Zach helped us raise the next couple hundred grand, and then—

They introduced you to, like, angel investors, to VCs?

Angel investors, VCs. Keith Rabois and Ben Ling co-led our seed round. So I think we were Keith's first investment at Founders Fund, first or second investor. And then Ben Ling, with his new fund, they had co-led our seed round. We raised $2 million. So then Roonick dropped out of school, moved back to the Bay Area. We were back in the crypto castle because we still didn't pay ourselves any salary. We realized that wasn't sustainable. So we paid ourselves something little, lived in an apartment, and just started scaling the business. We got registered with the SEC. We were actually the youngest ever people to be registered with the Securities and Exchange Commission, which is very fun for us. And we realized that we needed to be a money manager. We needed to actually manage the money. With the general premise of Vise being, we're going to sell the software to independent RIAs. We're going to help them build their clients' highly personalized investment portfolios. We're going to automate the management of those portfolios. We're going to provide the technology to explain those portfolios. So advisors, instead of building portfolios, could now be focused on their client relationships, can now be focused on growing their practice. Everything else would be handled by our system. And that was, like, kind of what we were working towards at that point in time. We— I had seen the show Silicon Valley, and always, like, wanted to do this, like, startup battlefield thing. I thought it would be interesting, and I would recommend any entrepreneur do this, and the reason why is because there's a lot of great recruiting benefits to having that kind of pitch or video on the TechCrunch website. But TechCrunch was having this happy hour with Sequoia. And we went to this happy hour. We didn't really feel like we needed Sequoia or any kind of investors. We had Founders Fund at the time, we had some money. We're like, we don't need to raise any more capital. And we're sitting at the pretzel station, because we just love free food. We didn't wanna pay for food. And chowing down on pretzels, and this guy comes up to us, his name was Sean, and he's like, what are you idiots working on? And—

It's basically how he tweets, so.

We love Sean. Sean's great. Brilliant. But, you know, he's like, what are you guys working on? And we're like, oh, we're building the software for financial advisors to like help them build portfolios. The whole vision is to kind of automate asset management. And he's like, that's really interesting. Like, I was at this retreat with John and Patrick Collison, and someone mentioned that, you know, financial advisors are the one part of the market that for some reason technology hasn't replaced them, and they continue to grow year after year after year, but no one's built any software for for them. And, you know, shortly after that, Sequoia preempts us on a seed round. And then Sean actually helped individually recruit lots of talent for us, which was helpful. And then we decided we need to move the company from New York, or from San Francisco to New York, because New York was a better hub for financial talent, because all of the financial DNA was actually in New York, wasn't in San Francisco. Sequoia then preempts us on a Series A round, and the pandemic hits. And shortly thereafter, Ravi at Sequoia preemphasizes on a Series B round. We've now raised like $50 million. And we start to scale assets on the platform. We're still only managing a couple million bucks in assets across a handful of financial advisory firms. But it's proving that the product works. And it's working pretty well at this point in time. And then we start to build out a small sales team. I'm starting to spend a lot of time on sales. And we scale basically in the period of 6 months from zero, close, call it $4 or $5 million in assets, basically nothing, to close to half a billion dollars in assets.

Got it. And why was Sequoia preempting the rounds?

I think they liked us. They saw the market opportunity. They saw the sizes of the market. And they saw that, you know, standard VC, like, if they didn't preempt us, someone else would. And I think that was pretty clear, because lots and lots of VCs at the time were kind of swarming around us. And VCs love to pattern match. So I think when they see young founder, big market, like, oh, this is the next Stripe and we're going to invest. I think the challenge is history, you know, seems like, you know, some of these companies were built much quicker with much more ease than they actually were. And when you think about tackling a space like the asset management space, where Vyse's vision is almost to build the next BlackRock, right? We want to build one platform that all investors, regardless of age or net worth, can get a personalized portfolio in an automated way across all asset classes, which is a very challenging thing to do given the size of the market and how, you know, taken over the market is by kind of legacy incumbents. Although it is the direction the market's heading in, it's going to be the direction the market heads in over a, you know, 10, 20-year period, maybe 30 years. It's not going to be something where, you know, you see immediate venture-scale success. In the period of, you know, 6 to 8 months, whereas, like, you can apply, you know, some capital, you have a standard kind of SaaS operating model, and you're able to continuously start to prove revenue. It's very much a space where you can compound growth more and more over time.

So there's a lot of people who are investors who will listen to this conversation, they hear you say personalized portfolio, and they're asking themselves, what the hell does that mean? What is a personalized portfolio compared to a normal portfolio?

Yeah, that's a great question. So if you think about what financial advisors deliver to clients today, they deliver ETFs, mutual funds, managed investment strategies, very generic investments. They don't take into account your values, your goals, your risk tolerance. They're just kind of generic investment strategies. And there's a lot of drawbacks to this, right? They're expensive. They don't, you know, optimize for your taxes. They aren't very explainable. And they're sometimes like pretty hard to manage. So what Vise does is working with financial advisor, we can take in all these different inputs about a client, how much money they have to invest, their goals, their net worth, needs they might have, their environmental preferences. Like, do they wanna invest in environmentally friendly companies? Do they not? Strategies they might wanna take. Maybe they care about value investing or growth investing. Maybe they've got embedded positions because they work at a company or they, have stock that they inherited at a low cost basis, and then career risks or risks they might have, right? Maybe they own real estate, they have angel investments, they work in a highly regulated industry, whatever it might be. Vise can take all of those inputs in one simple questionnaire, 30 seconds later, build a highly personalized portfolio of individual stocks. So stocks that match up, a portfolio that matches up to the client's individual needs. So we will give them a risk-adjusted portfolio over a certain time horizon. That is adjusted and continuously readjusts to the client's various goals minus all of their preferences. Let's say the client works at Facebook, they won't have exposure to technology or nearly as much exposure to technology, they will have no exposure to Facebook, they will be de-risked from their Facebook holding because Facebook makes such a large percentage of the S&P 500. Even if you're just investing in index funds, you're overexposed to Facebook. Or those environmental restrictions, you want companies that are environmentally friendly in your portfolio, you'll take out the companies that are not environmentally.

Is there like a core philosophy that you all follow at like the highest level, like modern portfolio theory, 60/40 stocks and bonds, and then you kind of operate off of that or move off of that based on the preferences? Or is it you start from scratch, look at the preferences, and then build a portfolio? And if somebody's preferences lead to 90% fixed income and, you know, 10% REITs, sure, that violates everything in a 60/40, you know, portfolio philosophy, but like, starting from scratch gets you there, and like, that's the goal? Like, how do you think about which direction?

We do both.

Okay.

So we have a philosophy, like, kind of grounded in empirical evidence, like a value-based investing strategy, kind of like Dimensional fund advisors, that advisors can customize based on various clients' needs. And then we offer all kinds of other different types of strategies as well. But the idea is the advisor has probably their own strategies, and we want to be able to operationalize their strategy. So the way our optimizer is built, it's very modular, it's very open-ended. So given a number of different kind of restrictions and inputs, client inputs or inputs from the advisor, and strategies advisor might have, we can build this personalized portfolio for them. But let's say something doesn't make sense. Let's say they have a goal that's unrealistic. Let's say they have a risk tolerance that's too high given the client's specific needs. Our idea is that We will tell the advisor and the investor through the whole process the risk and return tradeoffs and the implications of the decisions they're making on their portfolio. So the advisor will understand, okay, given these set of inputs, this is what my portfolio is expected to return over this period of time, kind of like a forward-looking Monte Carlo, and then the back-tested returns, you know, fitted for that particular portfolio given a set of individual stocks. But I think what's also interesting about Vise is because the customization is basically endless, it can service any different type of client, and it can do it in a highly efficient way. So you think, OK, what are the other drawbacks as to using Vies over index funds? Like, what about tax loss harvesting or something like that, right? And the reality is, so most advisors today don't actually do tax loss harvesting. Most investors don't do tax loss harvesting. But Vies, not only do we do tax loss harvesting, but let's say you're using a robo-advisor or you're using kind of a— generic model portfolio of ETFs, you're tax-loss harvesting on the fund level. So if the fund goes down, you can tax-loss harvest the incremental, you know, down. But if you're investing in the individual stocks, you can tax-loss harvest the individual stocks that go down.

So if inside of a portfolio or a fund, one stock goes down 20%, but the overall fund's only down 2%, most people are tax-loss harvesting the 2%, not the 20% on that specific position.

Exactly. So whereas Vize, you can— you're investing in the individual stocks. And this is important because you can understand what you're investing in, you can understand why you're investing in it, but you can tax loss harvest the individual positions.

Who gets screwed if you guys are successful?

I think the traditional mutual fund companies, the traditional asset managers. Like, I have this general belief, and you can already see it, mutual funds are seeing record outflows, and it's going to continue to compound over time. I think there's always going to be a place for ETFs. But I think mutual funds are going to see record outflows. I think the challenge is the retirement business is so big, and it's built on top of the mutual fund business, which is why it's not happening faster. But fees are very—

What's drawing the withdrawals?

ETFs were the big initial driver, and then step 2 is what you can call direct indexing or custom indexing, which is part of, a large part of what Vise does. And there's only a handful of companies that are doing it. But I think like 4 years ago there were less than $3 billion in direct indexing assets. Now there are hundreds of billions and we're on track for trillions of dollars in assets over the next 5 years.

Do you think that VISE or similar types of direct indexing products will eat ETFs and mutual funds completely? Or do you think that it's something where, no, they may just be products within the platform? So if somebody comes in and gives you your preference, they may go and buy individual stocks, but like, could you still allocate to an ETF or to a mutual fund?

For sure, for ETFs and mutual funds for other assets. But for core equities, I think over some period of time, assuming the on-ramps are gotten right, so the way people start to initially invest and making it easy, low account minimums, things like that, it will be very, very much direct indexing or custom indexing for those core assets.

So financial advisor goes ahead, adopts the technology. Is this something where they're going to their clients and they're like, ha, we've got, you know, we've got the black box, let's fill out this questionnaire, and then in 30 seconds it's going to tell us, you know, exactly what you're going to be invested in? Or is this something where they're still having the same conversation and kind of relationship management with their client, but then they're going back to the office and they're using this tech? And what I'm trying to get at is like, how much awareness does the client have that it's actually this software product that's determining what the portfolio is? Versus they think that it's still the RIA or something like that?

It's both. I mean, it's disclosed, so the client has to fill out paperwork, but sometimes advisors kind of take credit for it. They use it as a sales tool. Advisors are like, I have spent this time building you this personalized portfolio of individual stocks that perfectly matches your needs, compared to I've invested you into, you know, this generic model strategy or set of mutual funds. And advisors, when stacked up head to head, You know, the strategy of personalized stocks will just always do better. Yeah. Ideally.

How do you guys make money?

We charge an asset management fee.

OK.

So we charge a BIPS fee.

Got it. And when you go to the RAs, what is the biggest objection that they have?

There's probably 3 or 4 different objections, one being technology. So advisors haven't adopted technology historically until relatively recently. So technology adoption in the space has been slow. And it's been really slow around investment management. So as it starts to kind of increase, it's going to continuously go up. I think that's one key one. I think two, which we talked a little bit about earlier, is that some advisors just see themselves as stock pickers. I think it's 10%, like 80 to 90% of advisors now outsource some or all of their investment management. So it's becoming a broad part of what they do. So advisors realize they need to outsource. Someone gave one of our advisors a really good piece of advice that, you know, the advisor told me, which kind of resonates, is that in this business you can do one of two things. You can either manage the money or you can manage the people. And managing the people is a hell of a lot easier than managing the money. So most advisors realize that their job is to be their client's Sherpa. It's to be their client's therapist, their marriage counselor, their closest advisor, their financial planner, the relationship manager, so to speak, between them and their money. And they're staying away from the money management. So they see themselves as the manager of managers. And it's Vice's job to enable those advisors, not the advisors that are saying, OK, I'm a hedge fund manager. I'm going to give you my 10 stocks, and we're going to outperform the market.

Yeah. How much impact are the RIAs seeing right now through things like the rise of Robinhood and SoFi of the world, meme stocks, kind of this, like, day trader mentality that seems to have taken part in some part of the market? Like, is that affecting inflows to RIAs from younger people, or are they not seeing that?

This is really interesting. I spent a lot of time diving deep on this. So the average account size on like a Robinhood is like, you know, $5,000, $6,000. The average account size on Betterment or Wealthfront or traditional robo-advisors, probably in the low $20,000s. And what starts to happen is it is a great on-ramp for investors at the earliest stages to start investing. And to initially get into the markets. But what people realize is when there's any meaningful amount of wealth, like anything, you know, $50,000 plus, they are so far away from understanding how to manage that, even with technology in the loop, that they need to go talk to an advisor. Someone recommends them talking to an advisor, someone recommends them to talk to a human, and that human is a critical part of that process. So I would say that those traditional fintechs do a great job at getting, you know, people to invest earlier on. I think that, like, people meme stock trading is just kind of replacing the gambling budget. I don't think that's actually, like, investing. I think that's gambling. And I think a lot of those people know that it's gambling. And the money that people use to gamble is very different than the money they use to invest. You're not day trading— at least you're really not a smart investor if you're day trading your life savings, your retirement.

Somebody out there is for sure doing it, but we should not—

the vast majority of people probably should not.

I would not encourage those people, but 100%, somebody's out there doing it.

But advisors, the thing people always wonder about advisors is that advisors for younger clients, right? Like these older, you know, baby boomer advisors aren't going to be the same types of advisors that are going to be working with younger clients. And people wonder like, well, my, you know, like I'm a young client. Like why would I want to work with a financial advisor that doesn't understand me, that doesn't understand who I am? And the answer to that is pretty simple, which is more people that look like you that are young are becoming financial advisors. And there is a new generation of financial advisors, independent financial advisors that understand all kinds of different types of clients. So we've got clients that specialize in musicians. We've got clients that specialize with tech people. We've got clients that specialize with doctors, with lawyers, right? Like there's even LGBTQ-specific financial advisors. Advisory firm that just works with and understands that client base. So advisors are starting to change, become, you know, take different forms than the traditional kind of golf and country club, you know, advisor in their mid-50s that, you know, your parents probably use.

This episode is brought to you by Core, the free non-custodial browser extension built by Ava Labs, which is more than just a wallet. Did you know that you can also bridge Bitcoin natively across the Avalanche bridge and take advantage of the thriving DeFi ecosystem in that community? With Core, any crypto user can easily swap assets, display NFTs in a beautiful interface, and store your assets in a Ledger-enabled wallet. Plus, you can put real dollars in your Core wallet in just a few clicks. Go to core.app to access the full power of Web3 today. This episode is brought to you by Unstoppable Domains. They've partnered with blockchain.com to create NFT domain names ending in.blockchain. It's the perfect ending to show that you're a believer in a decentralized future. The Blockchain.com community can join a short waitlist to get one for free at blockchain.com/waitlist/blockchaindomain. Free NFT domains provide all the benefits of premium Unstoppable Domains, including fee-free lifelong ownership. If you don't have a Blockchain.com wallet, no worries. There's new free domains available to everyone. Either join the waitlist for for a free blockchain.com domain or visit unstoppabledomains.com to buy your domain today starting as low as $5. Unstoppabledomains.com. This episode is brought to you by Bullish. They've reinvented the digital asset exchange. They give you access to DeFi features like automated market making and liquidity pools in a regulated environment. It's a whole new way to generate alpha. Bullish's total trading volumes have exceeded $25 billion just in the 7 months since it launched. And their industry-leading order depth means you can trade confidently when you want at scale with better pricing and lower risk, all within a regulated market environment. Good reason to be bullish. Learn more at bullish.com/pomp and follow @bullish on Twitter today. What about Bitcoin cryptocurrencies? Like, it seems on the internet to be this really dominant asset class and everyone's super excited about it or they really, really hate it. And you kind of get the extreme reactions. But $2 trillion-ish for the entire industry is like pretty small compared to equities or any of the other asset classes. Like, what are you seeing there?

See, there's $85 trillion in assets sitting with financial advisors, and it's growing at this staggering rate. And I think what's interesting is like, Anthony, you and I live in a bubble, right? We live in a bubble in, you know, the tech world where it seems like crypto is the most important thing. Or one of the most important things. And the reality is, for most financial advisors, if you're a financial advisor in Iowa, or your client's a person, a teacher in Iowa, right, you probably have heard of Bitcoin a couple of times. You don't really know what it is or how to invest in it. And if you want to see mainstream adoption of crypto in people's long-term investment accounts, you need both institutional adoption on like retirement accounts and kind of on a mainstream scale to get to investors that aren't sophisticated. And have it as a default portfolio allocation, or you need to educate the advisor really well. And advisors are starting to seek out getting educated. And through a platform like Avise, or like a tool that services advisors, having crypto on that platform— and it's part of our vision, right? We want to have crypto as an asset class. We're probably going to roll it out relatively soon to having crypto as an asset class that any investor can invest in as part of a standard portfolio.

And what about regulators? How do they look at some of these technologies, right? It's very clear, like, how do you regulate a portfolio manager who's making decisions, they either fulfill their fiduciary duty or they don't, and there's certain rules that they have to follow, and for the most part, 99.9% of them do. When you start to introduce machines and algorithms and artificial intelligence and machine learning, like, do any of the regulatory environment change, or do they look at this differently, or is it the exact same thing?

Yeah, I mean, the regulatory environment has gotten very difficult for fintechs. Okay. The SEC is going after fintechs really hard, but more about are these fintechs doing what's in the best interest of their client, or are they doing what's in the best interest of themselves? Are they treating the client the way they should be treated? The way that they think about these algorithms, it's interesting because we've gone through some of these processes with regulators, is they want to understand how they work. They want to understand how does the algorithm work, how does it make decisions in the same way that an advisor would want to understand. And understand it.

And the same way the consumer, as we talked about earlier, they don't want to just know like, "Hey, just kicked out this random output." It sounds like the regulators wanna know as well, "Well, if you gave it this dataset, it came to this conclusion, why?" Exactly. Mm-hmm. And are they enforcing rules there, or is this more of like, "We're on a fact-finding mission. We wanna learn and understand how this stuff works, and then we'll decide if we need new rules or not"?

I think it's fact-finding right now.

Okay.

There is going— because it's still very new. The SEC is still— most of the regulatory agencies are still— they're very smart, which is, I think, contrary to what a lot of people believe, is that these regulators are very, very smart, but they're very slow. It takes them a long time to enact new rules and to understand real problems. I think that we're still in fact-finding, data-gathering until we can create new policy around some of these software. However, regulators are taking note. They are starting to act in a meaningful way. You can see it in a very public way against companies that are fintech that are not doing what's in— that's clearly not doing what's in their clients' best interest, and they're being deceptive or otherwise. And there's almost like a zero tolerance policy for it by the SEC, which I think has been pretty interesting.

Yeah. You're a young person who's now run a company for 6 years, or this company for 6 years, you obviously had a company before, and what I find fascinating is I think that you've learned a lot, you've changed your mind about a lot of ways of actually operating the business itself. So I thought that what we could do is I'll throw out a couple of topics about just running a business and you mind dump on me in terms of how you do it and things that you've learned over the last 6 years. The first one, which I know you have changed your mind on, it's OKRs. What is your current thought process and why did you change your mind on OKRs?

All right. So I thought OKRs were the most brilliant thing when someone introduced them to me. I think it was an ex-Google executive.

Of course.

A really big Google executive who's friends with us. And I thought they were brilliant. Our whole exec team was like, hey, we should do OKRs. We'll do it on a quarterly cycle. We'll have our main objective. We'll have all of our different key results and our feeding inputs. And then we'll have this big deck. I created this big deck presentation that outlines each OKR, each input, and the owner of the input. And here's the problem. Maybe for a big company they work. For startups, they're terrible, because there's usually one thing in the startup that is holding you back. There's one limiting reagent problem over all of the rest that the entire company needs to be focused on. One metric, one objective, one issue. And the problem with OKRs is you're distracting yourself from the key focus. If you've got 5 different objectives, or 3 different objectives, and 8 different key results, or 9 different key results, you are looking at so much stuff. And to keep a small team focused on a broad set of priorities and objectives, you are naturally going to create yourself into a situation where you kind of half-ass each of them. You kind of make some progress on each of them, maybe you make a lot of progress on one of them, but it's not the right one. Whereas to work in a— to be successful in a startup, you need to have one thing that is holding you back between now and hyperscale that maybe it's one objective. So in our case, it was like assets under management at the time, and one kind of leading input into that assets under management, which is like onboarded AUM from net new customers that everyone focuses on, the whole organization obsesses over. The whole organization spends all of their time and energy thinking about that. Everything else, you can just let run, you can let it operate. But if you're thinking about, oh, here's how I kind of slightly improve this, and slightly improve that, and slightly improve this, you're not really gonna be able to make step function order change in a, like, major way across all of them. And you need to see success in one in a meaningful way.

Let's walk through that decision. So, when you decided Aside what the one thing is, how did you guys go through that process to understand it was the AUM?

I mean, what is the difference between more customer adoption and less customer adoption? Like, what is— usually everyone knows, right?

It's something that's so obvious.

It's so obvious, right? Like, if you're having a problem with retention, right? You have all these customers that are joining your new app, and the initial growth is incredible, but retention is staying really low. The whole organization is going to know that, like retention is bad. When you ask everyone in the company, what is the biggest problem you think the company has, everyone's going to have an answer. And hopefully, like 50% to 70%, if not more, are going to have the same answer. That is usually your answer of what is the key objective.

And then how do you communicate that to the whole team? So like once you and the executive team decide this is the thing, all-hands meeting, is it like an email, a Slack message? How do you actually communicate it?

Just the mind of a CEO that, you know, to you, it might be the most important thing. To everyone else, it might not also be the most important thing. So when you communicate— this is the other problem with OKRs, because there's so many of them. When you communicate one of them on an all-hands or send an email about it, like, it might be important that day. People might not show up. People might not remember. Like, if you've got one key objective, if you've got one key focus of your business, one metric, that you're tracking, it needs to be plastered frickin' everywhere, right? You need to talk about this in every conversation. You need to be asking every single person every single second of the day, what is going on with this one particular thing? And then when you nail that, you can focus on everything else. But until you nail the main thing, until you nail the main focus of the business, nothing else matters. And I think the challenge is, as founders, we all think, okay, there's so much stuff on our mind, there's so many balls we're trying to juggle. And, like, we're going to talk about it once in an all-hands. We're going to send one company-wide email, and everyone's going to get it. And that's just not how people work. You need to talk about it everywhere.

In one of our companies, every single meeting starts off with a reminder of how many days since the company was started, and then what the mission of the business is, just over and over and over again, to the point where sometimes we call on people and say, hey, what is this? And if somebody doesn't know, I don't care if you're brand new or if you've been around for a while, you better know the answer to that question, right?

I agree.

And I think that your point about the repetition ends up being very, very important until it's literally ingrained in people's heads.

It's all repetition. And it's got to be simple, right? It's got to be a very simple thing that everyone can think of. So we had a new product initiative that we wanted everyone focused on. The whole company had to be focused on it. We just had a very simple name for it. When everyone thought about that product initiative, are we working on PCE2? And everyone focused on working on PCE2. It's a simple, easy thing. When you have these long, drawn-out objectives, it's really hard to kind of keep that in your mind. But if you have a simple acronym or simple target that everyone can think about, then everyone's going to resonate to it.

How do you do recruiting today? Like, what are some of the things you've learned over the last 6 years on how to be good at recruiting high-quality talent?

So I was really good at this early on. I made a lot of mistakes later on.

OK, interesting.

There's a couple interesting things. So the first is having a Sequoia. I love Sequoia, but having Sequoia is a blessing and a curse. And the reason why it's a curse is because having such a high-quality venture fund helps you, and like big backers, and all this funding, and all this press, and all this excitement helps you recruit people that are actually the wrong people to recruit at your organization, because they are there because you're a rocket ship that they think they're going to make lots of money on. But the reality is, you need people that are obsessed with your mission, that are willing to work their asses off, and they're only there because they love what they're building. They're willing to take a pay cut, they're willing to work extra hard, harder than they would have before, and they're willing to just take shit that they wouldn't have taken before. And I think the, like, pattern— like, if people look at, you know, the success and the draw that you have behind your company, like, they're gonna join for that reason, whereas I think early on, And when we didn't really have all the backing, we only had people that joined because they love the mission, and they wanted to join an early-stage startup, and they wanted to work really hard, and they were willing to take a cut, and they were willing to be extra loyal. Versus what happens after a certain scale is you start to get people that are there because you are this big flashy company, and you need to avoid that like the plague. Like, I used to think, like, part of the reason why we were able to raise so much capital, and we wanted to raise so much capital, high valuations and with big investors is because we thought it was going to be a magnet for talent. And it is a magnet for talent. It just oftentimes is the magnet for the wrong talent.

Did you run into problems with raising a lot of money or with high valuations?

It depends.

OK.

So there are problems in the sense that there's this general expectation with high valuations. And last year, I think everyone had a high valuation. You needed a high valuation, because what would happen is, Engineers would come and they'd say, hey, this company I'm interviewing at has a high valuation. Why don't you have a high valuation too?

Which is actually the opposite of what they should be asking, right? Because it means that there's less growth.

That's how people think, right? Because the way that these companies say is, we're giving you $2 million of equity. And we might be giving them, you know, $300K of equity, but it's at a way lower valuation that could be, you know, $10 million of equity in the near future, whereas that $3 million of equity at that other company You know, could be—

$1 million.

$1 million of equity, right? But that's how people are trained to think, and that's how they were trained to think.

They assign value of the company as a proxy for quality and likelihood of success.

Exactly. And we, like, you have high expectations. Like, you have to grow really fast. And I think the reality is, like, it was a good learning lesson for us, but the reality is, like, this is a business that is going to take a long time to build, and you do need a lot of capital. So it was a good idea to raise a lot of capital. Do you need a lot of people? I'm not sure. I don't believe so. I think that the narrative in Silicon Valley was always like, you have to hire a lot of people, you have to hire executives, you have to scale your team in a really big, meaningful way, and you're going to get more progress and you're going to grow faster as a result of it. I don't actually think that's true. I think if you are crafty, you build a really great product, you have a team that is incredibly loyal and mission-driven, even though they might be a much smaller team, but they're obsessed with what you're building and they're in it with you every single day in the trenches, then you're probably gonna have a higher likelihood outcome of success than the company that has 150 employees, big executives, you know, a lot of pedigree that, you know, probably moving a lot slower. They're probably moving like a big tech company, not like a nimble startup.

Yeah, when you think about working with the board of directors, what are some things you've learned there?

A lot. Okay. The challenge with boards, generally, even speaking outside of Eyes, is you've got lots of different perspectives. So everyone has lots of different perspectives, and they're operating with very little information. So they're only operating with the information from what you've told them, or from what they've seen historically in financials, or from other metrics or memos you've given them. I think the challenge is they're giving you advice, based on what little information they have, and they typically expect you to take that advice. And sometimes you need to know when to listen to them and when not to listen to them. And it's a hard line to draw, right? You don't know when the board might be right and, you know, their experience might be great. You don't know when, you know, you actually are right on this particular decision. And, you know, you should take that decision the way that you want to have it solved. And I think that The other issue is that incentives are very misaligned, right, for early-stage businesses, where it's like you have a financial investor, and the financial investor has their LPs, they have their partners, they have their reputation. Investors care a lot about their reputation. And, you know, quite frankly, that doesn't really matter for your business. What matters for your business is, you know, what is right for your business, what is right for the growth of your business, what is right for the customers, your employees. And, you know, you're oftentimes the best person to make that decision, not a board member. And I think it gets to this even broader point, which is a problem that not a lot of founders talk about, but it's actually a really big problem, is advice overload. There are too many people giving you advice too much of the time. There's this founder that you know, there's this— there's your employees, there's your customers, there's your board members, there's other investors. And you don't know which direction to go. And I think that it's important just to kind of listen to yourself and know, you know, take advice from a small amount of people that are consistent in their advice and that can deliver one way or another.

How have you balanced doing the work yourself versus delegating? Are you good at that or not good at that?

I'm not great at this.

Okay.

And the reason why I'm not great at this is because when you start to scale an organization, you're kind of playing this game of telephone. So you'll give work to a leader, and then that work will give their— will give that work to their team, and they'll give you some version of what you want back. But oftentimes it is very far from what you actually wanted originally. And I think that the larger the organization gets, the harder it is to communicate, the more you're playing that kind of game of telephone. And I think when you think about remote work, it becomes a catastrophe. Like, I just don't believe that early-stage startups can be done in a remote way. Maybe if you've been working with the team for a long time and like you have a lot of embedded trust and, you know, spend a lot of time with them in person, somehow it can get done. But I don't believe remote companies work because of this communication, this overlap.

Mm-hmm. When you see that you're not good at something, right, or you think, hey, I could get better at this, like, what are the steps you go through to improve on something like that? Are there specific things? Like, do you go seek out an advisor? Do you go and try to find someone who is good at it that's like another founder? Like, what have you done to try to get better at something like delegation or whatever?

So historically, I will call the person I think is best. And the nice thing is we've had lots of resources, the best founders, the best executives, the best investors, that we can call and understand how do you delegate. The challenge and why I'm not good at this is everyone has a different perspective. Everyone has different advice. And you don't know what advice is always right for you. You don't know what the right direction. You can take some people's advice on some things, other people's advice on other things, and they're not always going to fit in sync with each other when those two things are related. So that's why I think I'm not that great at it, because like, I don't— always know what direction, like what consistent path to take from an advice standpoint.

Talk to me about the touchpoints with your executive team, right? So as you're running the business, do you meet with them once a day, once a week, once a month? Like, what does that kind of cadence look like? And I think this is more a style, right? Is each business leader thinks about it differently. Some people want to, hey, I want to talk to you every single day. Other people say, look, we've got one We have one big meeting on Wednesday for 3 hours. All the executives come together, and we kind of review the business. How do you work?

So we've gone through multiple different iterations of this. Like, we've had executives, we've lost executives, we've worked without executives. I don't actually believe that, again, like, early-stage companies, like, slightly post-product-market fit, work with, like, seasoned executive teams. So if you're pulling seasoned leaders out of, you know, big companies expecting them to work well in startups, you know, it's a lost cause. It's not gonna work. And the structure that those people typically like to try and work in don't work too well for startups. So what we did was we did weekly leadership meetings on Wednesday. It was 2 hours. We would have everyone put together a memo the night before. We'd have a general memo, you know, and what happens is everyone gives their opinion on every different thing. And it's not clear, is this a decision-making meeting? Are there— how many different decision makers are there? Everyone wants to feel like their input is heard on a particular decision. Everyone wants to feel like an owner of the decision. And you as the founder, the person that's running the meeting, is like going crazy because everyone wants to have their input on the thing. And certain people have different levels of impact on a particular decision. But if they're all in the meeting, their voices are all relatively equal, right? So if you're thinking about a product decision, and your chief people officer is weighing in a ton on the product decision, and she's the loudest voice in the room, and your head of product isn't really talking too much, which is something that commonly happens in a leadership meeting, you know, you're diluting that meeting, right? So you kind of need to either tell people, like, hey, like, you know, stop talking, and like, you know, this person's allowed to talk.

How do you handle it? You literally just, right in the middle of the meeting, say, hey, you know, be quiet?

So I started to do that. In the beginning, I was very polite. What I actually think works in the right structure doing this is leadership teams don't work as big teams. Like, you have your team of, you have your group of functional leaders. So this could be 5 to 7 people. And then you have your S-team. And your S-team, maybe it's your, like, CTO, maybe it's your finance leader, maybe it's your co-founder. It can't be any more than, like, 3 or 4 people. It's gotta be a small group, and you have to have a lot of, like, respect and trust in that group, and you were talking to this team every single day. Like, they are helping you run the whole business, and they are filled in, they are keyed in on everything. This team will help you make the decisions. This team is the decision-making body. Then you have the information body. In the information body is all the different functional executive leaders. They might run the people function, they might run the legal function, whatever it might be, and they will help— you keep them informed. You do your 2-hour weekly leadership meeting, and you will keep them informed informed as to what is going on. You will get their perspective, but the decisions are not made in that meeting. The decisions are not made in that meeting.

And you make the decisions before or after the meeting?

That— it's kind of— you're making them all the time.

Okay.

But you're clearly logging those decisions. You're logging who's the DRI, who's responsible for this particular decision. You have to be very clear. You have a leadership operating system. So in our case, we had a leadership operating system. This is how we make decisions. This is who— how we appoint DRIs. This is what a DRI is responsible for.

What's a DRI?

A directly responsible individual. We have AORs, which was really helpful. So AORs are areas of responsibility. We basically had a big spreadsheet. In early-stage companies, people don't know who's responsible for what. You literally write out to the specific thing, what is the area of responsibility? Who is the owner of the AOR? What are they specifically responsible for? What is the backup? So who will, in the case they're not there, who's the person you go to? And then what is the process write-up? How do you go through this actual process? So the way it works is when someone quits, when someone leaves, when someone's out, it's really easy for people to swap in and take over transition AORs.

And when you do all of this, how much are you having to train employees on it when they first show up? So a brand new employee, and you're saying, hey, here's how we run meetings, here's how we do information sessions, here's who the leadership team is that's going to actually make the decisions, all that? Versus it's more of, hey, you are an engineer, you are a marketing person, you are on the customer service team, here's what you're gonna need to know on a day-to-day basis, and then they'll kind of over time through osmosis essentially just pick up, okay, this is kind of how the organization works.

So we did the latter, but moving forward, I think having a formal like week-long training program, even if it only works 1% of the time, that 1% will continue to compound. And hopefully it works much more than that. To how— here's how the company operates. Here's our culture. Here's how we view our values. Here's how we make decisions. All of those different things, and deeply embedding it, like almost having an onboarding class, and making sure people get it right, is really critical. And the other part to culture is like, how do you hold people accountable when they don't do things like that?

How do you do it?

We're still figuring out the specific, the right way to doing it, but it's usually rewarding people when they do. So we have Slack channels where we call people out for positively doing it, and then I always give people feedback of like, hey, I would have liked you to see this differently in a one-on-one setting. Like, never criticize people in public.

And how often do you have those feedback sessions or those one-on-ones? Are you doing that multiple times a day, all week long? Week long with various people in the company? Are you doing it just with the leadership team?

So previously, we did it just with the leadership team. And we did two things as far as feedback. So we did feedback at the end of every meeting. So people would write down their feedback. They would say, OK, here's how we thought the meeting went. We'd carve out 10 to 15 minutes at the end of the meeting. Here's how the meeting went. Here's some of the learnings. Here's what I would do differently next time. And everyone would write it down, and we would review it as a team. And then we would put a next action. Of like, next week we're gonna do this better. And then we'd always make sure to remember that as a team after every meeting. The second thing we did was on one-on-ones, I would do basically check-ins. So I would ask people, like, I'd ask people for specific, like, questions about how they would rate, like, working with us. So how do you feel our working relationship is this week? Is it a 2? Is it a 5? Is it an 8? Is it a 10? If it's less than 8, what would make it an 8 or higher? And you would score it, and you would do this like once a month or once every 2 weeks, and you would see the scores progressively change and get better over time.

And when you saw the score low, is that your problem or their problem?

I mean, it's sometimes both. You have to understand like what is the component part, like why. Sometimes it is your problem, sometimes it is their problem, and then you work out a plan, like what is the next action? What are we going to do differently?

The last question I have for you is, after running the business for 6 years, if you could go back— and there's plenty of young founders who are starting companies today— what's the 1 or 2 things that you wish that you knew and you would really hammer home?

Don't— so there's a lot of distractions that is really important to avoid. And I think there's more distractions now than there was in 2016, 2015, whenever I was starting. And I think avoiding those distractions is really important. There's only one thing that really matters, and it's your customers and your product as a result, so maybe two things. Other than that, product and customers, and maybe you can start to think about, okay, so once you nail product and customers and you've got culture, and, you know, some of these other things to running a company. Nothing, none of that other stuff matters, right? Founders get too obsessed, especially young founders, with status, right? I fell into this trap. I personally, like, it just seemed like what everyone else was talking about, and it seemed like the important thing. If you can't nail customers and product, you're gonna fail at everything else. Like, the types of VCs you raise money from, the pedigree of the employees you recruit, The time that you spend on going to founder networking events, all of it, it's a waste of time.

I 100% agree. It doesn't feel like that in the beginning, right? But over time, I think that you realize that.

I'm sure you had a ton of impact on making this together, but like, and I'm a huge fan of Keith Raboi, but like the Miami Tech Week thing, like half of our employees were like, oh, like we wanna go to this Miami Tech Week thing. And I'm like, look, like, The VCs that are, like, retired and, like, want to hang out, and, like, they can go to Miami Tech Week.

Keith's gonna kill you for saying retired.

They can— maybe not Keith, but— because Keith has a broader mission, which is he's trying to get people to Miami. But Keith aside, like, you need to stay focused on the core objective, which is our customers and our product. And anything else that isn't that, Miami Tech Week, is a distraction. If you're going for vacation, go on vacation. But like all of this other stuff, if it's not vacation, it is a distraction.

Mm-hmm. I think, look, it's just singular focus on what you feel is the most important thing, right? I think it's a pretty fair way to look at a business. And I actually think it's a very good way to look at a business. And for those that end up not liking that, like, they can go work at a different business.

Yeah, exactly.

Right? Where can we send people to find you on the internet or find out more about Vice? Vise.com, and—

V-I-S-E. vise.com. And I guess I'm on Twitter. I haven't been tweeting much, but maybe I'll start again.

Well, if I can find the Kanye tweet, that means you definitely got to start tweeting more.

Exactly. Awesome.

Well, listen, I really appreciate this. I think people will learn a ton, not only about artificial intelligence, machine learning, but also what you guys are building, and then how you actually run the business. And we'll definitely have to do it again in the future.

Cool. Thank you so much for having me on.

Thanks so much for listening to today's episode. I really hope you enjoyed this one. Make sure you're subscribed on Apple, Spotify, or your favorite podcast player. And if you're looking to transition into a brand new job in the Bitcoin or crypto industry, we've got you covered. Head over to thecryptoacademy.io. My team and I have been working with the top HR teams in the industry to develop an intensive 3-week training program with over 50 live events. We teach you exactly what you need to know to break into the industry, including live interview prep and resume review. Our students have been hired at over 75 of the world's best Bitcoin and crypto companies. Go to thecryptoacademy.io to learn more. Again, that's thecryptoacademy.io. If you enjoyed today's episode, make sure you share it with your friends, and I'll see you all for the next episode.