Invest Like the Best with Patrick O'Shaughnessy - Vlad Tenev - Navigating Robinhood's Evolution - [Invest Like the Best, EP.384]

Episode Date: August 13, 2024

My guest today is Vlad Tenev. Vlad is the CEO and co-founder of Robinhood. It was such a treat to sit down with him and discuss the behind-the-scenes of a revolutionary business we all know well. He details Robinhood’s journey to zero-cost trading and what it means to build a consumer-centric financial product. Vlad believes in finding the harmonies across mathematics and art and applies this lens to everything he builds. We discuss Robinhood’s new credit card and more products on the horizon, the company’s toughest moments, including the Gamestop episode, and the compelling future of AI in financial services. Please enjoy this conversation with Vlad Tenev. For the full show notes, transcript, and links to mentioned content, check out the episode page here. ----- This episode is brought to you by Ridgeline. Ridgeline has built a complete, real-time, modern operating system for investment managers. It handles trading, portfolio management, compliance, customer reporting, and much more through an all-in-one real-time cloud platform. I think this platform will become the standard for investment managers, and if you run an investing firm, I highly recommend you find time to speak with them. Head to ridgelineapps.com to learn more about the platform. This episode is brought to you by Tegus, where we're changing the game in investment research. Step away from outdated, inefficient methods and into the future with our platform, proudly hosting over 100,000 transcripts – with over 25,000 transcripts added just this year alone. Our platform grows eight times faster and adds twice as much monthly content as our competitors, putting us at the forefront of the industry. Plus, with 75% of private market transcripts available exclusively on Tegus, we offer insights you simply can't find elsewhere. See the difference a vast, quality-driven transcript library makes. Unlock your free trial at tegus.com/patrick. ----- Invest Like the Best is a property of Colossus, LLC. For more episodes of Invest Like the Best, visit joincolossus.com/episodes.  Past guests include Tobi Lutke, Kevin Systrom, Mike Krieger, John Collison, Kat Cole, Marc Andreessen, Matthew Ball, Bill Gurley, Anu Hariharan, Ben Thompson, and many more. Stay up to date on all our podcasts by signing up to Colossus Weekly, our quick dive every Sunday highlighting the top business and investing concepts from our podcasts and the best of what we read that week. Sign up here. Follow us on Twitter: @patrick_oshag | @JoinColossus Editing and post-production work for this episode was provided by The Podcast Consultant (https://thepodcastconsultant.com). Show Notes: (00:00:00) Welcome to Invest Like the Best (00:03:56) The Next Frontier in AI: Reasoning and Logical Deductions (00:06:19) Challenges and Approaches in AI Development (00:09:08) Formal Mathematics and AI Integration (00:11:23) Practical Applications of Mathematical Superintelligence (00:17:30) Robinhood's Journey to Zero-Cost Trading (00:24:38) Building a Consumer-Friendly Trading Platform (00:28:52) Robinhood Gold and the Future of Financial Services (00:35:51) Understanding Robinhood's Business Model (00:42:34) Navigating the GameStop Crisis (00:49:17) Improving Customer Satisfaction (00:52:43) Reputation Repair (00:54:52) The Future of Financial Services (00:59:06) Crypto and AI in Finance (01:08:09) Building a High-Performance Culture (01:11:42) The Kindest Thing Anyone Has Ever Done for Vlad

Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest like the best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at join colossus.com. Patrick O'Shaughnessy is the CEO of Positive Sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of positive sum. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of positive sum may maintain positions in the securities discussed in this podcast. To learn more, visit psum.vc.

My guest today is Vlad Tenet. Vlad is the CEO and co-founder of Robin Hood. It was such a treat to sit down with him and discuss the behind the scenes of a revolutionary business we all know well. He details Robin Hood's journey to zero-cost trading and what it means to build a consumer-centric financial product. Vlad believes in finding the harmonies across mathematics and art and applies this lens to everything he builds. We discussed Robin Hood's new credit card and products on the horizon, the company's toughest moments, including the GameStop episode, and the compelling future of AI in financial services. Please enjoy this conversation

with Vlad Teniff. So, Vlad, when we last talked, this is a strange place to start a conversation with you probably for most people listening, but it was sure intriguing to me when we last spoke. Can you explain what mathematical super intelligence is to us? Mathematical superintelligence is basically an artificial intelligence that can solve math problems at a superior capability to the sum total of all humans on Earth. So I think there's various degrees of mathematical intelligence. And one thing I personally believe, which I think is a little bit not controversial, but I think most people don't understand this. When you say mathematical superintelligence, most people

think you're building some fancy calculator that can multiply big numbers or come up with very, very large primes. And those are impressive things that computers have been better than humans at for a very long time. But really, I think the next frontier in artificial intelligence is building agents that can reason and can think like humans. And a lot of the big AI labs are talking about reasoning as sort of this next goal. And you can see the limitations of the current models with reasoning because they just make obvious mistakes and they very, very confidently say wrong things. It's particularly bad with math, but also language-based logical deductions. So my belief is that mathematics is reasoning. And so when you talk about building a mathematical superintelligence,

that's another way of saying something that can reason, take logical steps, make deductions, figure out itself when its own logical deductions are wrong and course correct, and really something that can think and solve problems at a level that's in excess of what humans are capable of solving. And I think the litmus test for this will be Millennium Prize, problems, which are basically some of the world's most difficult unsolved problems in mathematics. And these are things like the Riemont hypothesis, probably the most famous unsolved problem. And these are very rarely solved by humans. I was privileged to study under this professor Paul Cohen at Stanford, who solved one of them,

one of the most famous ones, and he won a Fields Medal for it. it was called the independence of the continuum hypothesis. And so like solving one of these problems is very, very rare. And my belief is that the next one will be solved by artificial intelligence. And once we get to that point, all future Millennium Prize problems and like difficult mathematical problems will likely be solved by artificial intelligence. And I think that's not too contrarian of a view. If you look at, for example, Metaculous, which which is a prediction market where people make predictions about advances in AI and when certain things will happen. Last I checked, it was a 43% chance that the next Millennium Prize

problem will be solved by AI. So I think the chances is higher than that, and there will at least be AI augmentation in the critical workflow. But I think once we get to that point, we'll basically have mathematical superintelligence at our disposal. What do you think is the right way to think about building this. Like, is this a whole adjacent pursuit to the text-based, chat-based foundation models from Open AI and Anthropic? Is it something that they should do and fold in? Is it something that should be built separately? I've got a million questions about this, because if it is the reasoning engine, then that seems really important. And like, we want to pursue that and build it. So, like, how does this get built? Is it way different than what we've

become used to with just scaling something against a ton of data to get a foundation model? And the traditional AI path, say a bit more about how we might get to this. And then the next question will be like, why is it so useful? Oh, yeah. We could talk for hours about this. Let me tell you my sort of contrarian take on this and why it's different. So I'd say the approach that is predominant among the large AI labs is that reasoning capabilities, which includes mathematical reasoning is going to be an emergent phenomenon, an emergent property of bigger and bigger models. So basically by feeding these models more and more English text, they'll sort of get better at reasoning automatically. And I think that might work to a certain extent. I'm not even necessarily

saying that that's incorrect, even though there's all sorts of problems with manufacturing data. And we've basically trained on nearly all the public data that's on the internet with the latest generation of models being GPT4 and Cloud 3.5. So now there's a bit of a data wall. And one of the big questions is how are you going to get more data so that these models can continue to progress on the capability curve? And people are looking at various ways to overcome that, including synthetic data, which basically boils down to using the last generation of models to create more content that then you use to train the next generation. And I think there's a big open question about whether that particular approach will lead to recursive self-improvement because the great thing about

the human data on the internet, it's pretty diverse. You get new things from time to time. It increases and you can get some real nuggets in there. But then can synthetic data generated by the previous generation of the model actually increased capability. My bet is not significantly. It'll probably be much less valuable than the human data because otherwise you would have some kind of perpetual motion machine. I don't think the capabilities of the models can actually extract and learn and self-improve that way. So that's a problem. One of the really interesting things that's happened in the past couple of years is you have these models and the advances in AI. You also have advances in the field of formal mathematics. Basically, there's a language

called Lean that just came out a couple months ago. It was fully released Lean 4. It was created by this guy Leo de Mora, who was a research scientist at Microsoft Research, but is now at Amazon. And Harmonic's a early adopter of Lean. What Lean has done is it's taken math out of the realm of chalkboard and chalk and put it into the realm of VS code and GitHub. So it's turned math into essentially a programming language where now you see prominent mathematicians like Terry Tao from UCLA or Ravi Vakil from Stanford. And these are kind of the young millennial Gen X mathematicians. So the guys that are a little bit more technologically savvy, they're adopting all the new

tools. The older mathematicians are kind of resisting this to a certain extent. But the new mathematicians are basically formalizing their papers in this language. They're using it to actually find errors, and they're doing these large-scale collaboration projects where they kind of distribute the work to mathematicians around the world, and you can actually break apart complicated math problems into small bits and have graduate students and others formalize them. This is all happening in lean without the use of AI. And so one of the things harmonic is doing is combining these two things, lean and formal mathematics that digitizes math and turns it into a programming language and all the advances in AI popularized by Alpha Zero and GPT4 that take a first step into reasoning.

And we kind of put these two things together. In service of what? I want to come back to an insight you had about the sorts of things that the best say human mathematicians are doing to solve a problem that's not just brute force data, but like inserting concepts into a problem that might help you solve it or something like that. So I want to come back to like the how, but let's just say we could like snap the thing you wish existed into existence and tomorrow we have the first mathematical superintelligence. What problems do you think that we would send that thing to solve most immediately? Like, why is the tool exciting to you? long, long term, mathematical superintelligence, if you believe that math is reasoning, which

I do, then mathematical superintelligence is a more specific way of saying like artificial general intelligence. So it can be used to solve pretty much any problem. And I think in particular, it'll be problems that have like a quantitative element to it and require actual quantitative reasoning and engineering. So you have a couple of specific domains that this is important. Engineering, computer science, if you boil down writing any sort of software and in particular mission critical software, the fundamental element in software is algorithms. And mathematical super intelligence, one of the direct things it'll do is write software really well and ensure that that that that software is provably correct so that there's no bugs and it actually works to its

specifications. It can make new discoveries in engineering and physics. It can run experiments. Over time, I think it can run organizations. Anything that a very, very smart human that requires reasoning to do, it will be able to facilitate. Now, there's some things that it might not have a huge advantage over current things in. For example, we have great calculators. GPT is great at writing history essays, so you probably won't need mathematical superintelligence to write a nice history essay or a poem. But I think new fundamental research, discoveries, anything that requires correctness in a chain of logical thought, I think could benefit from mathematical superintelligence. That's long-term. Verified software synthesis I talked a little bit about, I mean, you get to the point where if you imagine how much a company is paying a software engineer, if something could replicate the capability of a really amazing high-throughput software engineer, not only in writing code, but proving that that code is correct and bug-free, that's a huge value proposition, particularly in safety-critical domain.

like financial services, blockchain smart contracts, where there was a high-profile exploit due to faulty smart contract code just a couple of days ago. And you're seeing this, the wormhole exploit that bled out, I think, $400 million from this wormhole bridge a couple of years ago. Then you get into automotive and aerospace,

and if we're going to be building satellites and landers and things that go to other planets, those things are expensive right now. And so the standards for them operating correctly are really, really high. So I think you'll first start to see impact in those domains. But if you have something that can write bug-free software, that'll be generally useful. And I think it'll expand to other domains. And this is a thing right now.

People are writing software and verifying it, but they're doing it by hand. So it's kind of like an artisanal thing that's very expensive. And it's only basically done where there's, there's already a huge financial impact for doing things wrong. That's why Leo de Mora, who's the author of Lean, is at AWS, because AWS has high scale. And if there's something wrong with the encryption library or their cryptography, the results could be catastrophic and incinerate tens to hundreds of billions of value.

So it's those domains that they invest in manual verification now. I think that'll all be done by AI. Let's ground this in your day-to-day. How, if this snaps into existence, do you start thinking about applying this to Robin Hood, your products, your business? What are the sorts of things that experiments or ideas that it would start to unlock for you and the company? One very practical application is Robin Hood is fundamentally an engineering company in financial services.

is we produce a lot of software. The end product that you see as a user of Robin Hood is software, basically. And that software is high criticality. We've had scenarios in the past where a relatively small error or bug in the software has cost the firm tens of millions of dollars, and those are all very, very disappointing. And so let's say it was possible to create software that you could validate and provably validate is bug-free,

then a great place to deploy that would be our trading stack or our clearing stack, anything that has to do with money moving or trades being settled. And not to mention the crypto business where we're certainly deploying Web3 technology, engaging in smart contracts, moving large amounts of money. And so I think if it were possible that you could build software, validate software at very low cost, and prove that it holds certain properties, it can't double spend, or it uphold certain risk checks, I think that we'll get to a situation where the regulators will actually require that. The only reason it's not required right now is that it's just impractical to implement.

It's so costly that you would have no software. So if the cost of proving certain things and making software conform to standards of correctness was actually practical, I think you'd see every regulator requiring that to be implemented. I would love to go back to one of the episodes in your history and hear about the reasoning or the thinking behind it. Like, tell us the story. When I was in an adjacent business building software that was also trying to push some dimension of investing, in our case, custom indexing, like index quantitative strategies.

We often use the Robin Hood analogy where, like, if you plotted the cost per trade at the major brokerages and the discount brokerages over time, you saw this natural downline in the trend of the price. And then something incredibly innovative that Robin Hood did is, like, let's just take this to its natural conclusion, put it at zero. It strikes me that that was an incredibly important decision and trigger point for like the adoption of the platform. Can you tell us about that behind the scenes. What was going on? When did that idea come up? Was that like the fundamental idea? Was that something different? Like, I'm just always fascinated by how companies decide to do something like that, which on its face is like, wait a minute, that's the source of

revenue. So it can't be zero, but obviously it changed the business model. So I would love to just hear you riff on that story and how that came to happen. Yeah. Well, first, it was the understanding that that was possible even. Because I think if you talk to someone who's a lay person in that era and they're used to Charles Schwab charging you $10 to $15 every single trade, the first thing that comes to mind is there must be some kind of reason. Trading seems complicated. There's a lot of things that have to happen for that trade to get cleared and settled and executed and probably some people have to touch it and there's a cost to it. And I think we've also been trained as Maloney's a millennials, some of the great movies that we've grown up with are the trading movies, at least,

involve people throwing the paper tickets in the trading pit. You probably remember trading places with Dan Aykroyd and Eddie Murphy. That's what I thought trading was. And when I was growing up, to some extent, it still was that. So I applied for some jobs out of college. And one of the things I was considering going into was options trading. And this was really before options trading. was fully automated. And the options traders were doing all of the calculations and calculating the Greeks really quickly in their head. And some of them were in the pits. Others were clicking mouses and actually trading manually. And you don't have to squint too hard to imagine that there's humans involved somewhere in the process. And so it's expensive by its nature. But then my co-founder,

Bejew and I decided out of grad school to start a high frequency trading firm. And this was in part because he got a job, Bejou did, at a high frequency trading firm in 2008 on a whim and kind of learned the space. And it was a very, very new space. So I hadn't really heard much about it. And then we figured out that this is an interesting problem. It's a new market. It's growing. You're not really hearing much about it. But the approach makes so. much sense. And the approach is put servers close to the exchange matching engines, write really efficient code that takes advantage of arbitrage opportunities between different exchanges, build up the capability from there. Eventually, this high-frequency trading business that we

tried to create, which wasn't very successful, led to us creating a software business where we served other high-frequency traders and big banks that didn't really have the know-how to get into this new space, but saw that it was the future and didn't want to get disrupted. So they were buying our software. And what we realized pretty directly is you can kind of see what's going on here. The old mental model of trading involving humans didn't exist anymore. And in the equities markets, which were the broadest and deepest markets and the ones that most consumers wanted to trade, everything was fully automated. And not only that, but the lane season and the costs were dropping precipitously.

So at the time when we got into the business, which was late 2000s, 2010, the different exchanges were competing with each other on latency. So NASDAQ came out and said, we can offer you matching engine latency is of 40 microseconds. This became purely a computer science problem. And the cost of the trades basically was zero. I mean, we had customers trading tens of billions of dollars a day. And we had three people managing the software. So it was very, very scalable.

And we were charging like a monthly SaaS fee for our enterprise software. I was at a party in San Francisco one time. This was my previous company, Kronos, that was selling the software. We had a New York office and a San Francisco office. And I went to start the San Francisco office because Bejou and I were Stanford grads. So we wanted to get closer to Stanford because that's where we knew the engineering community and where we could bring engineers that could help us. I went to this party, and I met a couple of people there.

I was talking about what I was building, and I was like building this software. Our customers are trading millions of times a day, and we're charging them very little. And one person was looking at me and he says, can I use your software? I love trading. You're telling me that your customers can trade millions of times a day,

and it's very, very cheap, I'm paying $15 a trade at Schwab. I would love to have a modern, easy-to-use technology solution where I don't have to pay that $15. And at first I was like, no, that's ridiculous. You have to be a hedge fund. This is for sophisticated people. But it did kind of plant a seed. And we started digging into why retail didn't have access to this technology.

And we thought there must be some reason. Maybe there were different regulations governing retail, but there was no fundamental reason. There was nothing that said you could actually like replicate the stack of a sophisticated high frequency trading firm and combine it with a retail platform. And I think what we figured that out, we just got very excited because we knew what was possible then. We were like, well, we could deliver free trades to people. And it's very, very clear that that's going to be a huge market opportunity.

And when we got that idea, we were also thinking through what the product would be. And a lot of initial skepticism about Robin Hood was that they looked at us and they were like, okay, here are a couple of mathematicians and physics people. They're not going to know the first thing about building consumer products. They don't have that consumer product DNA. And so there was like skepticism that we would make something good. We knew our way around latency and back-end systems and performance and all this technical stuff, but we wanted to build a consumer product that was beautiful.

And we had this feeling that mobile was going to be really important. And this was at a time before Venmo was really popular and before financial companies were really taking mobile seriously. And I think there was a little bit of hesitancy from people to actually do serious financial stuff on their phones. And so the popular financial products that came out at that time kind of felt like toys. Bemmo was a little bit of a toy. It had this social thing. And those were the things that worked because people were skeptical of putting their serious money into a mobile financial app.

But we thought there was one major advantage to customers that are trading, which is that this device is with you at all times. And if you're an active trader and you really care about what's going on in the markets, that's just the capability that you can't get on desktop. So we really wanted to focus on that. And then we also understood that we didn't really have a fundamental advantage on web. Of course, the cost was an advantage, but these platforms were scaled and established and they had reasonable websites.

The last revolution where the internet came about led to a lot of new trading platforms. And so we thought if we could be the first in mobile and focus on that, territory that nobody was really paying much attention to. We sort of like did two hard things with that product. One has introduced free trading, but the other is we kind of redefined the whole market. The market is mobile predominantly now. That's the growth segment and all of the apps that came after us. And to some degree, the incumbents that had to rush in to make sure they didn't lose their existing market had to adapt their interfaces to look exactly like Robin Hood to because that's what customers were expecting. If they did something different, the feedback that you would get would be,

why can you just make it look exactly like Robin Hood? That's the interface that I want. I love the story because I think if you ask the average random person about Robin Hood's business trajectory, they would say, oh, it was this brilliant, simple, fast, light mobile UI plus this really clever, low-cost customer acquisition tool. And it's so interesting that the actual story is like a deeply technical story first and then the mobile UI story laid on top of that. And it makes me wonder, like having done both, what your product philosophy has become, if so much of your success was adopting the next platform, making it low friction, simple, beautiful, elegant, whatever, how would you distill down many years later now your overall product philosophy for all the things

that you do, whether that's design, performance, UI, U.I, U.S., however you want to approach it. I do think a lot of people just assume that if you really care about math and engineering, then you don't really care about design and art. But I actually think that there's a huge overlap between math and art. When you talk to mathematicians and they're talking about a proof to you or some result that they really like, they sound like art enthusiasts talking about a beautiful painting. They'll use words like elegant. Oh, that's really elegant.

That step is really beautiful. There is sort of a difference between different types of proof. Some are shorter and more elegant, and they have an element of creativity to it. Others maybe are a little bit more computational. And mathematicians care and understand about this stuff. And I do think when you're solving a consumer problem, there are differences. maybe it'll get solved either way, but there's like a certain artistry

in creating an interface that solves it in the best possible way for the customer. And so we obsess over that here, and we do really care about how things look and how things behave and making it intuitive for people. And I think Bejou really cared about this, and I really cared about it as well. So I don't think there's necessarily a conflict.

I know not a lot of people do it, but we are motivated by aesthetics and beauty and not just performance and accuracy. Maybe you use like a recent release like gold or something to bring that to life. What was the process like of conceptualizing, designing, rolling that out through that lens of elegance in product? Yeah. I mean, the gold card is probably the best recent example. We announced the gold card at an event in March of this year.

And the event itself, we spent a lot of time on. Not a lot of financial companies use events to launch consumer products. But we cared a lot about the vibe. We wanted to have a certain aesthetic, which was more of like this 1980s luxury vibe. So we looked at print magazine adverts as inspiration. If you look at the Robin Hood Gold Credit Card website, it should evoke the feelings of like an old horse ad or like a 1980s Rolex ad that you would find in a magazine with the serif fonts, spotlight, and camera features.

And people at Robin Hood, we have a lot of great artists and they want to build beautiful things and get excited when it's something more than just rolling out the features and meeting the requirements. So at every level of that product, there was like high quality care and craftsmanship. We spent so much time obsessing over the card itself. We got really excited about the physical card, the app experience. We acquired a company X1 that had an existing product, and it had all these great capabilities, and they really cared about design as well.

Some of them were, I think, buried in the product. We wanted to bring them out and surface them a little bit more. And then we asked ourselves, can we go beyond just having really good design and a great product? And can we disrupt the economics, much like free trading did? And that's where we got the 3% cashback on all categories value prop, which if you talk to customers is so shocking that they just assume we're losing a bunch of money and that this is going to be bad for the business. And it's sort of in the too good to be true category, much like free trading initially was

and then you kind of put this together. And I mean, you should look at, sometimes I just read the reviews of the Robin Hood credit card app before bed because I haven't seen reviews like this for any financial product. They're better than free trading initially. It's just all five stars, people saying this is the best credit card they've ever used. And the only one-star reviews are the people that don't have it yet, that wanted so bad that they, like, are leaving one-star reviews on the app out of frustration.

Can you do something similar as you did with the free trading for the 3% and explain to us what you learned, what the traditional model has been, what the governing variables have been, how you're changing things and being innovative. The card itself looks beautiful. I'm sure it's a great experience, et cetera. But like talk us through like the back end physics of this and what you learned and what you've done. One thing that has been helpful is looking at existing credit card companies and their economics. So Capital One is a great. one. It's like a credit card business. They do a great job. They're efficient. If you take their revenue divided by their total transaction volume, and this is just simple surface level, they generate

revenue from a variety of sources. But if you just simply look at total revenue from card divided by transaction volume, you'll get to something like 6%, which basically means there's like a 6% revenue yield on transactions. And of that, I think they give out about two to three percent in rewards closer to two percent. The very profitable business, they've got a healthy profit margin, gives us a little bit to play with. And credit cards are some of the most profitable businesses out there. If you can do it right, it's a great business. And so it really just came out of looking at that. So there's some profit margin,

and maybe we don't want to do, can we go a little bit over the 2%. Now, we probably don't want to go to 5%, even though that would be really, really cool because maybe that's a little bit tight and we don't expect to be as efficient at running a credit card business as Capital One out the gate.

They've had decades of experience. But if you could give more rewards and get 3%, then there should be plenty of monetization to make this a good business to us and simultaneously put more money in customers' pockets. And I think the shocking factor is that this is like a greater reward than the interchange economics that you would typically get. So interchange is like between two and two and a half percent. That's the rebate we would get from the network on each transaction. So we're actually paying a greater reward than we get from the network. And we compensate for that in

two ways. One of them, which is what Capital One has, is interest on people carrying balances, so revolving activity. And I think that's going to be enough, in the long run, at least, to make the card portfolio very, very profitable on a standalone basis. But there's also another thing that basically, to be conservative, we sort of underwrite at close to zero. But if you look at Fidelity. Fidelity has this product, which is a 3% rewards credit card that they give to their highest net worth customers. So if you have something like $2 million under management with Fidelity Advisory Services, you can get the Fidelity Rewards Credit Card. It's a very nice green card. So you get 3% for that card, but you have to have millions under management. Now, what that should

signal to you is I don't think they're doing this out of the goodness of their hearts. It's a profitable enterprise. Their customers probably, if they were there for advisory services, they would be happy with just an at-market credit card that wasn't losing money for them. So I don't think it's losing money for them. And what they're likely seeing is that the secondary effects on the rest of the financial relationship for being the primary credit card account for someone, particularly if they're a high-quality, high-fico individual that's pretty.

predominantly using the card as a transaction tool is accretive. And so we think the ecosystem second order effects will actually be very significant for us, or we think there's a good chance that once someone has a primary card relationship with Robin Hood, they'll invest more money, they'll deposit more money in our gold offering, and it'll help drive the gold flywheel that has been working very effectively for us. If you had to like ultra simplify the business as it exists today, like an understanding from an investor's perspective, let's say, of the business. I'm curious how you would do it. I ask that first so that I can ask the next question, which is about this cross effect that you just described and what the vision is for all the things you might do for the Robin Hood consumer customer over the longer term.

But maybe you start by orienting us in the business today just in terms of like where the revenue comes from, what the customer base looks like, however you want. to do it in simple format so that we can build on that in terms of where the business might go from here. Yeah. So basically the business is focused on three things. The first one is being number one in active trading. And these are options traders, equity traders, active trading drives the majority of the revenue of the business. Crypto is kind of in that category. It's being the number one place for retail trading. And then the second part of the business is growing wallet share with millennials. You can think of it as Robin Hood has more millennial customers than all the major brokers combined, more than Schwab, Fidelity, Vanguard, e-trade put together for millennial customers. And we think there's

going to be close to $100 trillion in wealth passed down from baby boomers. And it's already started, but it's really going to accelerate over, say, the next one to two decades. So close to $100 trillion in wealth transfer. And so we're one of the few financial companies that actually has tens of millions of millennials who are trusting us with multiple products. And we think we can be the beneficiary of that. So it's sort of like building the tools and the capabilities and putting the incentives so that the lion's share of that ends up in Robin Hood accounts over the long term.

then that's what we mean when we say we want to grow wallet share with our customers, predominantly millennials. All of your money should be in Robin Hood. All of your financial transactions should go through us. And that's why you see our work in retirement, the credit card, which will eventually expand into a broader neobanking offering. All of your money will eventually be best served in Robin Hood accounts. High yield savings, eventually advisory. We're going to serve all of your assets and all of your financial needs. And then the third category, international, recently started. We launched in the UK and the EU, so still small, but I think over a 10 to 20 year time scale, we're going to be live in all major markets. All of our products will be available there.

And I think there will benefit from the rapid wealth building in the developing world, where if you look at some of these overseas markets, Eastern Europe, Southeast Asia, Latin America, Africa, right now, they don't have very much disposable income, but over time, that will increase. The standard of living worldwide will continue to increase, and they're going to be increasingly interested in investing. So core business active trading, the second priority is setting ourselves up for this wealth transfer with products that help us build wallet share and help our customers build their wealth and then expanding internationally. Can you double click on the active trading portion since that's the core original piece and

predominant piece still today? And then we'll get into some of the future stuff. Can you just explain how it works? I feel like everyone knew the answer to this question like three years ago and payment for order flow and all this other stuff was like in the news all the time. But I would love you just to explain like the model of how the business works now to understand the foundation that you're going to be building on. Yeah, for sure. So you're primarily asking in terms of revenue, what drives the revenue that? business. Yeah. Options trading is a large piece of it. So options trading is annualized in the 500 million to one billion revenue range, if you look at the last few quarters. Crypto trading,

last quarter crypto trading was ballpark 150 million in quarterly revenue. So that's a big business as well. Equity's trading is smaller, predominantly because the payment for order flow take rate that we collect is quite small. That's, I think, ballpark 100 million in quarterly revenue. And then we generate revenue from these customers holding cash on the platform. So we collect a small interest rate spread on their cash. Securities lending also is an additional revenue source. And a couple of other smaller things.

Like if people want to withdraw money instantly to their bank account, we offer that service and charge a small transaction fee there. And if you break it down in another lens, transaction revenue versus interest revenue. A couple of quarters ago, interest revenue actually flipped transactions. It used to be very, very heavy on the transaction side. So payment for order flow was much bigger than net interest margin back in the zero rate environment.

But now the interest rates have gone up and we've retooled our business to have more asset building products than net interest margin actually flipped transactions as a greater contributor to revenue. And then we also have a subscription offering, Robin Hood Gold, which ties all these together. And we really think of Robin Hood Gold as the product version of growing wallet share with our customers. Once you get sufficiently engaged into Robin Hood and you find value, you almost always end up being a gold subscriber. And then becoming a gold subscriber actually creates a strong incentive to try out all of our other products, where you now get industry leading economics. And gold subscriptions are now a nine-figure revenue business as well. I think the last number

we announced at the quarterly earnings was 1.7 million gold subscribers, which was an all-time high. And the majority of those customers are paying us just $5 a month to be gold subscribers. I love this idea in businesses. It's so cool. You're a public business. You've got all these segments. The numbers are big. It's great. You made it. You did it. You built the big business. on the way to doing so, I love the idea of crucible moments, like the incredibly hard choices that companies have to make or difficult situations they have to work through to get to that promise land. If you had to think of one or two that the episode of it itself was the most impactful on you, or you just feel it the most as you think back on it, which one or two pop most immediately to mind?

I'd love to just like hear the story of how you processed a hard moment on the way to building a big business. A lot of people, their minds when they hear about Robin Hood go to GameStop for better or for worse, and they're like, oh, that must have been very, very difficult for you. It was in a certain way, but it was difficult in an acute way. So we had this crisis. We had a certain set of steps that we had to take, make sure that we serve our customers, raise a bunch of capital, strengthen the balance sheet, and then deal with the people. PR fallout and the congressional hearing and all of that.

Not a lot of uncertainty in how we operate there, other than how best to message what was actually happening to the press and to the public, which was challenging. I don't think we did a particularly great job of that. It's sort of like a meta problem rather than an actual problem with the product or the business. I think what was more difficult for me is navigating 2022, where we had a company that had recently gone public. The entire macro environment was reversing from very, very loose monetary policy to the tightest monetary policy in multiple decades. When that happens, when interest rates get

hiked up very quickly, trading basically slows down to much, much lower levels. And so we were seeing that. And we had a business that was dependent on trading. And we were acting in 2021, like those trends would continue indefinitely. So there was a huge amount of headcount growth and lots and lots of investments in kind of ancillary things that weren't going to generate value to the business anytime in the next five years, let's say. So huge investments in like very, very long-term projects. And when the rate environment shifted and our core business was weakened, we had to basically change everything about the company. I mean, we had to cut headcount quite dramatically. We had to focus and prioritize on the core business, which was the most active customers, less so the novice

customers that were actively trading, but doing so more as a novelty. We were focusing more on pro-sumer, very active options traders and equity traders that were more sophisticated and driving the bulk of the revenue. And that was a big shift. We couldn't just focus on the core business because we had to diversify away from trading as well. It was clear that there wasn't going to be a huge amount of growth in the active trading segment for possibly two or three years. And we had to focus on the core while also making sure that the business in the medium term was diversified away from the core and that we had another way to grow customers and grow assets outside of active trading. That was all very tough. It played out over a long period of time and pain inflicted

was much less acute than GameStop, but it burned over a long period of time until relatively recently, actually, when the market and the public started realizing that, hey, these things that Robin Hood has been doing since 2022 are actually working. And we left this company for dead, but now they've actually like strengthened the core, retooled and diversified and growing and expanding aggressively like they did in 2021, but under a very different environment. Let's say tomorrow, a young entrepreneur comes into your office that you know and like a lot and you want to be helpful to. And they say, look, I'm in a situation that sounds like the one you just described in early 2022. And you know that lots needs to change.

but nothing's changed yet, and it's like all in front of you. What would you tell that person in terms of like how to start biting away at that problem piece by piece? Like, what were the good elements of the methods that you use to get through 2022 that you think might be portable to other entrepreneurs, all of whom, if they're going to build something huge, go through hard trying times? What worked and what would you do differently? A couple of entrepreneurs have asked me this. And one exercise that I found useful for myself was, was to sit down and imagine what the company would look like if a new CEO came in to run it. And a new CEO who's very, very good.

This might be a great chat GPT or Claude Prompt, too. And at the time, what I used was Frank Sludman, you know, because he was kind of like the business book CEO who everyone pointed to as hardass, came in and fixed companies and got them right on track. he's my archetype of central casting CEO coming in to fix things and turn things around, even though I've never met the guy. I would ask myself, all right, if Frank Sleuteman came in here to clean this place up, what are like the 10 things that he would do?

I came up with a list. And for us, it surfaced these obvious things that we weren't doing, but that we, for a variety of reasons, maybe we were like clinging on to decisions that we had made in the past, that we wanted to be consistent with. And those are the most dangerous because it's very, very difficult to go up in front of the company

and say, hey, this thing that I just told you that we were going to do three months ago, I was wrong. That was the wrong decision. We're just going to completely reverse that. Much easier for a new CEO to come in and do something like that because all you have to do is say the old guy was an idiot and everything they did was wrong.

So we're just going to change it. So I think that was the usefulness of the thought exercise. And it's like, what was the answer to that? You're going to have to cut headcount. Focus on the core of the business. At the time, we had very few people actually working on options trading, which was driving a ton of our revenue.

And so we looked at that and said, a smart capital allocator would put more and more people on options trading, make sure those customers are very, very happy with the service they're getting because they're generating the lion's share of our revenue. And we were in a situation where you look at the numbers, the more active a trader you were on Robin Hood, the more revenue you were generating for us, the less satisfied you were with our product. There was an inverse relationship between revenue and NPS. And when we kind of groked that, we were like, this is a five alarm fire. We have to

solve this problem immediately. And we did. We solved that problem within a couple of quarters, much, much faster than I thought possible because we just put our best people on it. Yeah, what did they literally do? What were the changes that were made to make those customers happier? A lot of those changes were like incredibly simple. So for example, one thing that was a source of a lot of problems was that if you were trading a lot, you were marked a pattern day trader. And if you were marked a pattern day trader, for a while, we basically would not let you trade options. Because we hadn't built the mechanisms to like softly off-ramp you into other products, you just basically had to turn to a competitor once you became a pattern day trader.

And the solve for that was building cash accounts, which is a different account type, non-margin. And once someone became a pattern day trader, we had like a simple, ramp to off ramp them into a cash account and they could continue trading with us. And that alone was both solving a pain point, increased NPS for customers, and led to them staying on the platform. And these were some of our most active people. And this was below the line before because we just had a very small team of people working on this. And so little paper cut improvements, even significant ones were just not getting prioritized when you have a handful of people working on options, which was a multi-hundred million revenue business.

So from a revenue per engineer standpoint, before we did this, it looked great. But we were very nomadic in how we built. People would build the old things. Then they would move on to new things. And we had lots and lots of projects that weren't going to generate revenue for the next two or three years at least, and we basically had to put all of those and make sure we fed the core that was driving the business first and expand outward from there. So there's huge changes in the service model, huge changes in the product roadmap. And I guess the other advice

I'd give entrepreneurs is it can seem kind of daunting if you see a lot of things that you have to fix. But it really is true that you tend to underestimate what you can do in a year. Everyone Once things done in two weeks, you can do a lot of stuff in two weeks, but pretty much all of these problems can be fixed if you give yourself a long enough time horizon. And one benefit we had at Robin Hood is we had a very large balance sheet because of all the capital that we raised in 2021. And so we did have some leeway and we wanted to turn things around quickly, but we had $6 billion of cash in the bank, which can also solve a lot of problems. It's pretty ironic that the one crisis led to help solve the solution for the next crisis. It's pretty interesting silver lining I hadn't considered before. I had never really thought of it that way, too.

But yeah, we navigated the GameStop crisis and left it with a reputation that we had to repair, but also a lot of capital because that crisis led to us growing a lot. and we grew our users in revenue tremendously. And it kind of propelled us to go public as a company right as the window was closing. I mean, I think we were kind of at the tail end of the time companies could even go public. So if we'd waited a couple more months, we probably still would be private. What are the keys to repairing a reputation? I think it really is true that reputations can be lost very, very quickly.

We see this in our data. GameStop obviously had a negative effect. And you saw that in NPS. It took a dive pretty much immediately. In our NPS survey that we ran right after the GameStop events, it was at the bottom. And then you could actually kind of see it gradually increasing over the next several years. So it got to like a decent level. The biggest increases were right after GameStop.

You saw it go from very negative to like slightly less negative. but really you do good things, roll out good products, those have a gradual effect on NPS. And so it's been three years. We're probably now at the point where we're pretty much on par with free GameStop Robin Hood. So it's been a journey. And obviously doing good things can speed it up marginally,

but I think the biggest thing is time. And particularly in our space, If you're handling someone's money, there's no getting around the fact that time is important. And the timescale that you've been a company, particularly a public company, it factors into people's decisions for where to put their very serious finances. One advantage that the incumbents have on us, they've been around for multiple decades. Customers have more confidence that they'll continue to be around for decades since they already have been. And that's a uphill battle that we have to climb.

But the benefit is it gets easier over time. So there's just this tailwind behind us where our brand and reputation and trust, provided we don't have missteps, knock on wood, time actually helps us in financial services. And now we can say we've been around for 10 years. We've been public for three and look at what we've been able to accomplish. And we have seen that reiterating that message does itself build trust. One of the interesting things about your business is the data set that you sit on top of. You just see what people do in investing in active trading.

And if you go back to when Robin Hood started, around the same time was this crop of robo advisors. They were called venture funded like Robin Hood was and had big exciting addressable markets. And the reality is, I think the pitch of those firms was this is really simple, low-cost, easy responsible way to invest for the long term. And you shouldn't be actively trading. You should just set it and forget it. And it's interesting to me that like those firms, there isn't like a big dominant one. They didn't really achieve venture scale outcomes. Robin Hood did. What does that teach us about how people are with trading, with their money, with investing? What have you just learned sitting on top of watching all this trading for a long time now, just about people in the nature of

trading in markets? Robo advisors are interesting. Robo advisors were kind of the original application of AI to investing. At least that's how it was marketed. But they really weren't doing anything sophisticated in terms of AI. And I think the reason that those products have failed to gain significant market share or grow super quickly is that automating the actual asset allocation is the easy part. And I think that's a commodity service. And even if you have a human advisor, they spend a very, very small percentage of their time. Doing investing.

Yeah. Automating your portfolio. Yeah. And the value is in all of these other things, helping you consider your financial situation holistically, helping you with your will and inheritance and trust and estate planning, giving you advice on your budgeting. situation, doing your private banking, so helping you file your taxes. And I think that there's so many different services that registered investment advisors and financial planners and private

bankers provide to customers outside of just the portfolio allocation that AI was misapplied there. I think what people really want is the experience of a private wealth manager or a private banker delivered to them digitally in a lower cost so that many, many more people can access it. And I think that that's the really valuable and really interesting application of AI to our industry and one that Robin Hood can really help drive. Robo advisors themselves have some advantage here relative to the incumbents. They also utilize technology. But yeah, I think you're going to start to see generative AI.

and this new intelligence applied to advisory in the next couple of years. And I think that's going to be a big transformation in the wealth management industry. Do you think that you'll play a big role in that? Do you have ambitions to play a big role in that? We do. Yeah, I think there's two big technology trends that could affect financial services. One is crypto, the other is AI, and Robin Hood will be a leader in both. I mean, we're already a leader in crypto, and we're making lots and lots of investment.

investments to make sure we're the leader in AI. I don't think the competition is particularly formidable, but I also think that's not an excuse to get complacent. So yeah, I do think we're going to be doing a lot. And we also just recently announced an acquisition of Pluto, which is basically an AI-based investment research product. So you can put in information about companies and investments. And they built a very sophisticated system that can pull public filing, synthesize the information, and answer questions and help you with investment research. So we're going to be looking to integrate that capability and actually build awesome tools

for customers, both self-directed and in the managed account space with advisory. Where do you think it all might go with both? crypto and AI, like I'm interested in both separately and maybe the intersect or overlap or whatever, but it seems like the pace of change of this stuff is just hard to wrap one's mind around. If you think not just two years hence, but five or 10 or 15 years hence, where do you think these two categories bring us? What does financial services look like that's most different from today as a result of these two technologies? I think that with crypto, the short-term thing that has happened and I think has led to maybe some sullying of the brand of crypto is that people

see all the meme coin activity and lots and lots of things that don't have any fundamental backing being created just because it's very easy to create a new token on a blockchain. And so that's, I think, distracting a lot of people from the underlying potential of the technology. Now, at the same time, here's what we can see. So abroad outside of the U.S., if you look at transaction volumes on crypto blockchains, they've been going up. And what's driving that increase is largely dollar stablecoin volume on low-cost networks. So for a while, USDT, which is a tether stable coin on Tron, was a huge amount of actual crypto transaction volume outside of the U.S. And this is people that want to hold U.S. dollars.

They want to hold U.S. dollars. In a lot of cases, they're in local markets where their currencies aren't doing very well. They're rapidly depreciating. The dollar has been a safe haven for a long time. They'd rather hold that than Bitcoin because they trust it more. And it's got a great brand and perceived stability. And so the best way that they found to do that is to use crypto. So they're using not custodial wallets and getting dollar stable coins.

because it's cheap, it's efficient, it's easy, it's available on any blockchain. It's rapidly surpassed the kind of traditional financial infrastructure. I think that you can extend that to stocks. You'll see tokenization of stocks, and that'll get very, very interesting. And you'll see tokenization of pretty much any asset. What I think about a lot is I started thinking about this when there was that whole ICO, mania that happened in 2017, where companies were ICOing and the SEC came in and cracked down on that. And Robin Hood was a small startup at the time, and we were kind of looking at what could

happen in the space. I think a very disruptive force in capital markets is companies starting out and issuing tokens upon their formation. So imagine you're a software engineer and you're considering two different companies. One has traditional cap table. Maybe they're using carda or something and they've got their traditional legal docs and you get stock that vests over four years and you get liquidity either when the company gets so big that they can do a tender offer or they go public. And you tend to discount the value of that equity as someone that's applying for jobs because you just never know if it's actually going to materialize. And you have a choice On the other side, it's a startup doing a substantially similar thing, but they've been incorporated

and they have a crypto token and they're crypto-native. And maybe their business doesn't have anything to do with crypto. It's the same business, but you just get a token right away and they can tell you, hey, our token is liquid. Once it vests periodically, you can just exchange it for dollars or whatever you want. I think that if that were possible, pretty much everyone would choose the latter option. And that's going to be a force that changes how startups incorporate and manage their equity. And then if you see that continuing in a large portion of new startups doing that, there's going to be less companies going public in the traditional way. And companies will do ICOs instead. And that could fundamentally change the equity markets.

Now, the unfortunate thing is there's regulation that doesn't want this to happen and a lot of embedded special interests. So it looks like that might happen outside of the U.S. first. But as the U.S. tends to do, it will eventually catch up. And you will see crypto infrastructure getting more and more embedded into non-token, non-crypto-native projects. There's clear benefit. Look at the dollar stable coin analogy. We see the efficiency gains of using crypto infrastructure because we have a crypto business and a traditional brokerage business, the cost of offering those services is dramatically lower on the crypto side. I think it's very, very likely that we move off of mainframes and onto blockchains,

and there's like a efficiency benefit outside of all of these things that I've mentioned. I think it's inevitable, and Robin Hood can help drive that, both on the product and technology side and the policy side, because we want the U.S. to be a leader here and for us to benefit from all this technology and not clutch our pearls because we are the market leader already in traditional finance. So how does AI then play a role in some of this and or drive totally different things for your product roadmap, your future, the future of financial services in general? Crypto's impact is going to be more on the market.

Infrastructure and on the market side. AI's impact is likely going to be a little bit different. I think that when you think about the impact of AI, you should think about what are the types of things that a human is doing for me in financial services that I'm paying a decent amount of money for it. I think there's a lot of distraction in the AI realm. There's speculative technologies, things that people aren't really doing. And some of those might work and there might be new behaviors. But the things that I know will work are things that you're paying a human, a decent amount of of money for already that could be done by AI. And that's both on the company side,

where are we investing our CAPEX, where are we spending money on things that could be automated, but also on the customer side, what are services that are provided by humans that could be automated by this technology or if they're not automatable already with a bunch of elbow grease and you can trace the curves and assume the technology is going to be better and make less hallucinations and less errors, what's going to be automatable in two years. And I think what you come up with is near-term customer support. The global customer support and market is going to see a major dislocation. I mean, not just in finance, but across the board.

I was looking at this market. It's one of the biggest markets. I think something on the order of $500 billion is spent annually on call center. And that's going to be reallocated in the next 10 years. You can kind of see that happening very, very, very, directly, those folks, their jobs are fundamentally changing. They're going to have to do different things. And then in the long run, the really interesting opportunity in finance, I think, is private banking and private wealth management and advisory. If you look at the dollars and the

revenues, it's quite dramatic. And I think there's a user experience issue on both sides of the barbell of this product. So on the one side, if you're not a high, net worth individual and you have $50,000 or less and you're going to your local Edward Jones or something of that sort, you're paying a very high percentage, 2%, 3% on that capital because it has to be worth it to have a human helping you with this. And you're getting not the best service. It doesn't take much to be a financial advisor at one of these firms and you're not getting elite quality service there. And so you're unhappy. And then on the other, side of the spectrum, if you're a high net worth individual and you have hundreds of millions of

dollars with a financial advisor and your wealth is growing over time, you're being charged a fixed percentage. So basically, the service you're getting isn't going to be very much different at all if you have $10 million, $100 million or a billion dollars, but you're just paying 10 times more. And so the more money you make, the more dissatisfied you are with your financial advisor relationship. And it really should be the opposite. The more money you have, the more you're utilizing that product, you should be happier. It's inverted in a very strange way. So I think the whole space is going to be disrupted.

And I think AI will play a big role in that. If I re-invoke Sleutman again and he storms into your office and spends a week, tearing apart your business. What do you think he would say is the worst part of the Robin Hood business today after all this improvement that happened after 2022? I mean, I think there's a lot of things that are still not perfect and we're on the journey of making them better. Yeah, for example, there's probably still a little bit too much bureaucracy, a little bit too much process or maybe like 25% of the way to building the company that I want to build from a culture standpoint. And maybe I should articulate what that is.

So some things I believe, we want to have a relatively small number of the highest quality people. I would rather have a small number of just elite people that I pay more. It sounds obvious, but not a lot of people in my space believe this. and actually a lot of the incumbents and public companies, if you look at it, they invest heavily in offshoring, which is having large and large numbers of low-cost, mediocre people. And there's a benefit to that to a certain extent. But my preference would be a small number of amazing people that are elite in their fields

that we compensate very, very well. And I just think for me personally, I would have more fun being in that environment. I think people around me will have more fun. And if you're in small groups of elite people, they kind of like push each other. And the end result is better than what they could do individually, which I don't think is the case for people that are less talented. So that's one thing.

I want as little bureaucracy as possible. Things should be effortless to do here. It should be a joy to deploy code and get products to customers. and I want compensation to be sort of close to perfectly tied to performance. So it's very, very clear who's getting promoted and who's not based on how well you're achieving your goals and doing the things that you're supposed to be doing and setting out to do. And in particular, I want to de-emphasize, I think a lot of times when we think about

promotions and leveling and recruiting outside people, one of the, common things that people look for is the scope of their existing role, which translates into how big of a team did they lead? Oh, if someone has led a team of a thousand people, surely they're good enough for Robin Hood where they can lead a team of 500 people or more. I think at a lot of companies that creates an incentive to hire empire builders that are inefficient. And what I instead want is someone that's done big things with as few people as possible. Those are kind of the North Stars, a culture of high performance, top talent, in person. I think in person is valuable. And I'd say I'm remote tolerant, but I don't think the future is fully remote and distributed

companies, at least while humans are involved. I mean, maybe if we get to mathematical super intelligence, things will look different. I think there's a huge benefit. And I think if Frank Sluutman came in, I don't know exactly what he would say, but I'd say we're maybe like 25 or 30 percent along these journeys and there's still ways we can improve. I don't think there's anything major that hasn't been started yet. If you could know anything that you don't know the answer to, like what question would you have answered? Just like in general about any of the stuff we've talked about or what answer would you get? I don't really even have to think very much. I'll tell you what got me really interested in physics and in math in the first place was I would read

these books about the Big Bang and string theory. And it's all getting to the fundamental question. First, what are the rules of the universe and how did they arise? What is that fundamental theory that explains the laws of physics and why the forces and the subatomic particles are the way they are. That's what got me interested in physics to begin with. I don't think string theory provides a very compelling answer. I think there's probably something simpler. But I think that's really the fundamental question. Why these rules of physics, what really happened before the Big Bang? If we can figure out the rules of the machine that we live in, I think that's inherently useful and valuable. And that's kind of what drives me. Yeah, ultimately, it's that long before you're at

about a creator. It's amazing how the deeper into science and physics you go, the closer you get to that fundamental question and like trying to glean the nature of or existence of a creator. It's a bizarre search. There's a lot of other related problems like the Fermi paradox. Where are all of the advanced civilizations? Why can't we see any? See all of these stars and all of these planets and humanity has been around for not that long. There should be civilizations that have had the best. benefit of billions more years of evolution. And you'd think in that amount of time, they would leave their solar system. Think about what we've done in 100 years. We've gone from basic airplanes to sending Voyager out of the solar system in less than 100 years. So if you add a billion years of

that, we should be populating the entire galaxy, right? We don't see any civilizations populating entire galaxies. Hopefully the answer is not AI kills us off. No. That's one of the explanations, right? I've read too much Nick Bosterham, probably for my own good. Vlad, this has been so much fun. I feel like we could talk about just about any topic, and you're obviously a super curious guy. Every time I interview someone, I ask the same traditional closing question, what is the kindest thing that anyone's ever done for you? Oh, well, I'm sure many, many people have done kind things. One thing that I found myself reflecting on,

recently is when I left my PhD program, I left my PhD program in 2008 to start my first business. And I remember being quite nervous about it because I had to go into the office of the department vice chair, who was this guy, Dimitri Schlatenko, this is at UCLA, and Russian guy, which means intimidating in his own right, but he was a very nice guy. And I had to ask for permission to take a leave of absence. And at that point, I didn't really have any safety net. Of course, my parents could support me, but I was very rightful. And to me, having my parents have to support me was just not very acceptable. And I didn't have money. And I was a grad student. And I had this idea that was very risky of starting a new business with my partner. I wanted to ask for a leave of absence.

And I was like sweating. I was preparing for this meeting for a long time. What am I going to say. And I thought that there was a pretty good chance you'd say, yeah, you can leave, but taking a big risk and you're not going to come back. But this guy was so encouraging. He's like, oh, well, you'd like to become an entrepreneur. I wish you well. And by the way, if it doesn't work, we love you here. We'd always welcome you back. And that made me feel so good. And it's sort of like empowered me to feel comfortable leaving, just knowing that I could, come back and that they wouldn't make it difficult for me. And so a year later, they sent me an email and I had to figure out whether I was going to try to extend my leave of absence. But they're

like, are you coming back? And a year later, nothing was working. I was an entrepreneur. My first business was failing. Everything was in disarray and on fire. But I had loved being an entrepreneur so much and I knew it's what I wanted to do that I was just like, I'll figure it out. I'm not going back. This is it for me. I'm an entrepreneur and there was no going back. But over that year, a lot of learning about myself and growth had to happen for me to get to that point. And I just think that conversation was absolutely defining for me because I was still like very uncertain about whether I should even do it. And then about 10 years later, I was invited to give the commencement at UCLA Math. So I was invited to give a talk in front of the students and they invited me to

this dinner where I think Dimitri Schlottenko was there. And we talked about this thing. And he remembered the meeting very fondly too. He was like, you know, I remember you came into my office. You asked for a leave of absence. You said you were. starting a business. Very, very few people actually from the mathematics department start businesses. And I always use it as an example to people. You don't have to be a mathematician. Everyone thinks like it's a dead end career because if you're not a mathematician, it doesn't make sense to do the major. And one of the reasons we invited you is very few people actually start businesses and we want more people to do that. I thought that was just very,

very nice and he remembered it. Yeah, I think sometimes you have to ask for stuff and you'll be surprised what happens. What an incredible, cool closing story. I don't think we've had a story quite like that in 400 of these. So thanks for the story and thank you so much for your time. It was a blast. I really appreciate it. Thank you for taking time. If you enjoy this episode, check out join colossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly, where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.