The Wolf Of All Streets - Monopolizing the World’s Money with Richard Craib, Founder of Numerai

Episode Date: November 3, 2020

Richard Craib began his career as a quantitative analyst for a hedge fund. During his tenure he formed a vision for a new financial system, leading him to found Numerai, a novel hedge fund that collects trading signals from around the world and pays participants based on their success. The idea is ambitious; Numerai is prepared to be the only hedge fund in the world. Scott Melker and Richard Craib further discussed the definition of a quant, machine learning, million-dollar data sets, managing all the money in the world, crowdsourcing signals, beating the S&P, the Numerai Master Plan, a brand new hedge fund system, Ken Griffin and Citadel, poker players and trading, the ethical good in trading, hedge fund exclusion, accredited investor laws, the data + talent secret, and more.  --- ROUNDLYX RoundlyX allows you to dollar-cost-average into crypto with our spare change "Roundup" investing tool, manage multiple crypto exchange accounts in one dashboard and access curated digital asset content and services. Visit RoundlyX and use promo code "WOLF" to learn more about accumulating your favorite digital assets when making everyday purchases and earn $4 in free Bitcoin. --- EQUOS Diginex is the first company with a cryptocurrency exchange to be listed in the US. That exchange, EQUOS, has been built to institutional standards, but is available to everyone. You can trade Bitcoin and Ethereum spot, as well as Bitcoin perpetuals, and get a 5% discount on all fees, by signing up using equos.com/wolf. --- CELSIUS With the Celsius app you can earn up to 15% APY rewards on over 30 cryptocurrencies. Have crypto but want cash? Celsius also offers the lowest cost loans against your crypto with interest rates starting at just 1% APR. Enter promo code WOLF when you sign up and get $20 in BTC! Users must transfer and hold at least $200 of any coin for 30 days to be eligible for the reward. --- If you enjoyed this conversation, share it with your colleagues & friends, rate, review, and subscribe.This podcast is presented by BlockWorks Group. For exclusive content and events that provide insights into the crypto and blockchain space, visit them at: https://www.blockworksgroup.io

I'd like to thank my sponsors, Celsius, Equus, and Round the X for making this episode possible. Stay tuned later in the episode for more info. What is up, everybody? I'm Scott Melker, and this is the Wolf of All Streets podcast. Today's guest is the founder of Numeri, a company that crowdsources stock signals from around the world and aims to, quote unquote, solve the stock market. In this episode, I intend to better understand how Numerai uses blockchain, crypto, machine learning, and AI to crack the world's financial system and also to learn more about the role of quants in moving markets. And to answer all my questions,

I have the founder and creator of the company, Richard Crate. So Richard, it's a pleasure to have you on the show. Thank you so much for being here. Yeah, good to be here. Thank you. Awesome. So before we get into the questions, once again, you're listening to the Wolf of Wall Street's podcast where twice a week I talk to your favorite personalities from the worlds of Bitcoin, finance, trading, art, music, sports, and politics. The show is powered by Blockworks Group, a media company with over 20 podcasts in their network. You can check them out at blockworksgroup.io. If you like the podcast, you follow me on Twitter, then you should definitely check out my website and my newsletter.

You can find both of those at thewolfofallstreets.io. So now on to what's important. Richard, tell me what it really means to be a quantitative analyst, a quant, because obviously everybody sees it in Hollywood. We see it on the show Billions, but I think most people don't know what that really, really looks like. Yeah, it's a bit confusing. It's kind of misused. There was a time kind of in the 90s, and maybe even the 80s and stuff where people were, they were kinds of people like Black Scholes, you know, who kind of like Black Scholes model was a famous model that

was like a model to sort of use physics principles to come up with a way for pricing options. And that kind of thing isn't really what happens nowadays, where it's much more about data and modeling data. And so the skills you needed back then, maybe you could be like a physicist, and then you could make special kinds of derivatives models. But the skills you need now, what you normally call a quant, is becoming more and more like a data scientist, where they're modeling some data set to create a signal that they can then trade on.

And so what kind of data are you looking for to create accurate signals? So on Numeri, we give out data. And it's more than a decade of data. And it's all obfuscated, and our users model it. And it is kind of conventional quant data, but it is quite expensive to get it for global equities. So the data set we have maybe would cost half a million dollars a year, which kind of prices out normal people from participating. But Numeri gives it all away for free because we obfuscate it. But our new service, Numeri Signals, lets you come with any data that you want.

And you can have an existing model that you've built from tick data, from some alternative data, any kind of data set you want, and you can submit that to Numera as well. Cool. So it's somewhat like democratizing data for anybody who wants access and to be able to trade and create their own models. But more specifically, are we talking about social sentiment data? Are we talking about social sentiment data? Are we talking about previous price action? Are we talking about fundamental news? You know, are we talking about earnings? What kind of data is it that you're obfuscating?

So the data we give out is, yeah, I would call it kind of normal quant data. There's a reason we obfuscate it, right? We don't talk about what it really is. Understood. But it isn't like impossible to find data. It's just very expensive and commercially available. So yeah, inside of it are the typical things, value indicators like PE ratios, momentum indicators like 12-month momentum. Those kinds of things are in it, but there are like 310 of those. So it's a lot of features of that nature. And how did you actually get into quant? Obviously, before Numeri, I believe you had a career

actually doing this with billions under management, correct? Yeah. So I used to be a quant at another asset management firm I basically I graduated with a degree in mathematics and then took some machine learning it was a kind of exciting time to take machine learning when I graduated it was 2012 there were a lot of big breakthroughs and then I went to be a quant and they kind of like let me let me do whatever I wanted to do And especially were very open to me trying new methods. So the company had simple models that they would kind of use. And then no one had tried machine learning at the company. And so I had this experience. And so I spent a few years making a really good

machine learning fund for that company before starting Numeri. So why did you start Numeri? I mean, what are you guys doing that's so different? Well, the key thing is that this idea of like, there's so many small alterations you can make to a model that have a very large effect because the signal to noise is very high, right? So low. So there's lots of noise.

And making just a small change to your model might help it a lot. And having just one new idea might help it a lot. So the question is just like, well, how do I have more of those kinds of ideas at the same time? And the answer seemed like obvious that you'd have to go the crowdsourcing route because there were multiple things, such as competitions on this website called Kaggle, a data science website. They have millions of data scientists that have signed up to participate in various competitions. while being a quant, I did quite well in some of them. But really, you started to realize like, what if all these models were put together at the end? And how good would that be? And then finance, that's particularly valuable, because you can have a model that's not that good. But if it's uncorrelated, it helps you, it helps your shark. And so crowdsourcing is such a natural way to do that. So I started thinking about Numeri around then. And yeah, that was like December 2015 was when we launched. That's funny to hear you say, obviously, it could be a subpar model,

but the fact that it's uncorrelated can make it useful because I think a lot of us who are in the Bitcoin community view somewhat Bitcoin as such, right? Even if it does badly, at least it's not doing badly at the same time as everything else, right? So how important is correlation or finding idiosyncratic, you know, risk and idiosyncratic assets or strategies? It is kind of the whole game. I mean, it's not, it doesn't, it's somehow, it doesn't matter that much for individuals with very long time horizons. So for example, if you are just trying to save your 401k, it's not that stupid to throw it all into the market. And you just have market exposure. You have lots of US risk, you have lots of dollar risk, you have lots of market risk, but it's still in the long run, it's not like the worst idea. But if you're an allocator, like an institutional investor,

you can't do that. Because at any given, we'll say you're a pension fund, at any given time, some of your investors are redeeming. And some of them need the money now. And so you can't have, it doesn't work if you have a 50% or 55% drawdown like we had in 2008. And even recently a pretty big 40% drawdown or something. You just can't run it that way. So you have to put maybe like 60% in stocks and then like 1% in venture funds. And then, you know, like why not put 1% in Bitcoin? And that argument is pretty strong. And certainly you want to put maybe 2%, 3% in hedge funds.

So it's not about, yeah, it's really about making the overall thing much better for large institutions. And what's also good about hedge funds is they're quite liquid. Any investor in our fund can take the money out within 30 days. And that means they can use that almost as cash, they think about it more as cash than market exposure. You said that you were speaking specifically about institutions, but I think the same rules generally apply to your average investor, at least when they're able, because obviously accreditation or issues and people can't be

exposed. So you said that maybe 1% in Bitcoin, but hey, 2, 3, 4% in hedge funds. But I think that almost everyone is excluded from participating in hedge funds with the current model, correct? Yeah. The accredited investor laws make it very difficult for hedge funds to have individual investors. And it's not even that. It's also the tax treatment is very bad. So the cool thing with your big institution, you don't pay tax, like income tax on the gains, because there's lots of trading in a hedge fund, right? But if you're an individual, it's as if it's like normal income. So it's really discouraging, you know, people from investing as individuals in hedge funds. So how does Numerai solve that? Well, we don't really have the fund available to individuals. The fund is basically my money. One of our earliest investors, Howard Morgan, co-founder of Renaissance is in the fund and a few others. And so we're really trying to have kind of like five people

in the fund, but two of them are investing, you know, more than a hundred million dollars. That's the goal to really have, to really have a small handful of institutional investors because that's the way the market is kind of designed from a regulatory perspective. You can't put $5 into our fund. Does that exist for your average investor who just wants some exposure to hedge funds or is that still kind of in the distance?

There's sort of some ETFs that are trying to sort of pretend they're hedge funds. And they say, you know, we, we a little bit like where we mirror a hedge fund or something like that, but they're not the good ones. So none of the good hedge funds are, are kind of like ETFs. But you know, all this regulatory stuff is, is in some ways what's pushing people to crypto. If they can't access the normal financial system, if too few companies are IPO-ing and it's impossible to invest in hedge fund because the accredited investor laws, blah, blah, blah, they just created crypto and now they invest in

that. Yeah. I mean, it's interesting with the present laws, sometimes it's difficult for someone to invest in their friends or family's business. You're not even allowed to give money to a friend and family depending on how it's structured. So I think there's a whole other podcast conversation based on the issues, I guess, with the laws of accreditation, which are evolving. But I want to talk more specifically about what you guys are doing. I read in your medium, really cool story about Ken Griffin. That's his name, right? Ken Griffin. I would love for you to share that and talk about sort of how that applies to what you guys are doing. What you know how that inspired you? Yeah, I read about Ken Griffin. Ken Griffin is the founder of Citadel,

very big hedge fund, $30 billion hedge fund, and himself a billionaire and has been a big participant in the market for a long time. But he started his fund when he was 19. He started trading when he was 19. He sort of got some money from his grandmother and his dentist. Maybe back then they didn't have their credit invested. Yeah, that's okay. And then he got Harvard to install a satellite on his dorm room so he could get trade quotes. So, it's kind of an amazing story. But the sad part about it is it wouldn't be really possible today. So, on the one hand, everyone has the internet now.

No one needs to install a satellite dish on their dorm room. And everyone maybe has Robinhood. So it feels like everyone can trade. But weirdly, they're more trapped out of the real game than ever. So you might have Robinhood and you think you can trade. But guess what? Maybe they're selling the trade, the trade flow. In fact, they are selling the trade flow to Citadel.

So they're selling it to Ken Griffin. So it's very hard for you to credibly say, well, I have an edge as an individual. And because the whole financial system is so built up and it's so hard to say, well, me with my small amount of data can take on these big guys. It's possible, but it's much harder than it used to be. And the way that Numeri solves this problem is kind of saying, well, we can give you the data you need and the tools you need and the system you need to monetize your signals by giving them to us, staking on them so you can earn money on your signals. But you don't

really have to bother with the trading and actually buying data yourself and setting up a whole and setting up a prime broker, which would cost you like $10 million. So there's so many, so many problems that prevent the next Ken Griffin from participating. And I hope that Numeri is where the next Ken Griffin will go, because it'll be the most, the easiest place to submit a signal. So if everybody's submitting their signals, obviously there's incentive for the performance of those signals or how well they do or how much they're used, I guess.

Who's actually trading on those signals? We trade them. So inside of Numerite, we have a hedge fund, and we trade the combination of all the signals, which we call the meta model. And so we of all the signals, which we call the meta model. And so we put all the signals together. And that means, you know, there's some strategies that, like I said earlier, are sort of not very high return, but they would help our existing strategy a little bit. So they're valuable to us, but they're not valuable to the person who has

them. So they might as well sell them to us and earn cryptocurrency. So that's how it works. So interesting. And so, so brilliant. It makes me feel, so I'm a, I'm a retail trader, right? I primarily in crypto, I know there's zero edge in stocks. I, I kind of,, I'm a long term stock investor like anyone else. And I trade occasionally for fun. But I found that, you know, trading crypto, you can at least maintain some sort of edge. But it sounds almost like your average guy looking at a chart somewhat has no chance against what you got against, you know, the bots or the algorithms and the quants. What's interesting is they might get lucky for even a few years in a row. The kind of analogy I like to make is you can sit with the poker pros, the best poker players in the world,

and you can leave the Knights up. Sure. But you couldn't do that a thousand times in a row. And so what happens in investing, people make five trades in a year and they say, I'm beating the S&P by, I'm double the S&P return. Right. And it's like, okay, well, you kind of had twice the beta of the S&P, or you just got lucky on your technology bets or whatever it is. And you can get deceived a little bit. And you can so you can. So in a way, those people who are deceived about their skills, who think they have an edge are kind of feeding the monster of the of the big funds because um when those things turn around when those risks that you're exposed to turn around uh they will they will do fine but you'll do badly so i do see it as like it's possible to do well but it's very hard to know

even if you're doing well whether you're're skilled. Right. Or lucky. Maybe after 10 years, you can see if there's some skill, but earlier than that, it's hard to tell. I talk about this all the time. I mean, it's like two things, right? It's like the whale who goes to Vegas and takes the casino for $4 million in the movie. And so they keep flying him back on the private jet and loading him with expensive champagne and giving him free hotel rooms till he loses it all back. Right. That's A. And B, I mean, the market, you know, random reinforcement is a real phenomenon where it kind of tends to reward bad behavior and to punish good behavior.

You have a system, but it starts losing. Then all of a sudden you go all in and somehow get lucky and you think you're a genius. Right. So you're basically saying that unless you've done it for five, 10, 15 years, you can't have any indication that you're anything but lucky. It depends. It's like, um, you know, in online poker, you can play 10 tables at once. Right. So you get more information and data, right. On stocks, you can make a thousand trades a day. And if you made a thousand trades a day and you did well,

then maybe you're doing, maybe you do have something. Right. But at the end of the day, do you think that all trading is moving towards either algorithms or bots of some sort and that it will basically be all automated eventually? I do think, yeah, I do think so. I do think yeah I do think so I do think it will be it will be like this the system where we just kind of we know that so much like if imagine numeri you know five years from now has a million different data scientists and all the data in the world and like why would you go against that like you would just look at the prices

and be like well I guess those prices are pretty fair and you'd be like I'm not gonna play this game I'm gonna use the market to buy and sell things because I need to take on certain risks so I'm gonna but I'm gonna treat the prices as being well priced at all times and maybe that's where some people's heads are at already but it's gonna going to get even more like that. But that's interesting. It sounds like trading dies and everything just becomes an investment

because if you get to that point, you're so efficient that nobody can really exploit any inefficiencies and it really does come down to everything's fairly priced and efficient, correct? Yeah, but that doesn't mean there isn't new information showing up. So it doesn't mean volatility goes to zero, right?

There'll still be a coronavirus in the future. Probably. There might still be this coronavirus in the future. Yeah, so that's really interesting. So talk about, I guess, those efficiencies in the market. I know also in that same medium, I believe, it talked about how important price is because obviously it allocates capital to the proper places, to the people who need it, to the people who are creating better products for humanity or whatever it is.

Can you talk about the importance of accurate pricing? Yeah. I mean, it is. So can you talk about the importance of accurate pricing? Yeah, I mean, it is very important. And I do see this as something that people overlook. I mean, you think as a trader, what is your job? Your job is to make money. And you're just self-interested. And you'll only care about making money for yourself. But the reason we have the stock markets in the first place, and the reason we set up even like a capitalist society, is that the whole idea is that those selfish trades

where you're trying to make money for yourself, help the world in some way. And how does it do that? Well, it helps the world because the prices go to where they should be. And so I made this example in the medium post. If you have $1 Tesla stock, imagine what that world's like. How could Tesla borrow money to finance a new factory? How could they hire any employees? People would just quit the company. So the price of the stock really matters for what the company can actually achieve and how fast they can grow and all those things. So when a trader comes in and see, and the hypothetical trader comes in and sees a hypothetical $1 Tesla stock, and he buys it and buys it and buys it and buys

it until it's a hundred100, he has done something really good for the world because now Tesla has much more capital. It's not really like, people get confused, not really like Tesla is getting all the money he's putting in. He's just buying it from other traders. But the fact that Tesla's price is much higher means they can issue new shares at that price to finance their company. So that's the, that's kind of the magic of markets. And that's why we do it in the first place. So if you come at it from the perspective of, I like the outcome of this, I care about the outcome of the world getting better, the price prices getting better. If you come at it from

that perspective, instead of the selfish perspective, you start to realize things. So if you're just a selfish trader and you have no data edge, you have no data that other people don't have, what are you really doing trading? How can your insights really be helping? And then maybe you say, well, maybe I'm not going to trade because I'm not helping the end goal here. So I don't want to just get lucky randomly and make money. I want to help the end goal. And somehow if you care about helping the end goal, it helps you become a better trader and a better market participant. And that's why with Numerai, we're hoping that all the trades we make are very differentiated from what other hedge funds are doing because we're doing everything in such a different way. So we're adding original insights

into the market to make the prices better according to how we see it. It's interesting. I've very rarely heard someone make a case for traders being altruistic, but it makes me feel a little better about myself. Yeah. No, it's a strange thing. I mean, it's like even consumption, right? Going out and buying things with the money you earn, it feels like selfish. You're just buying things for yourself, but it's helping all the people you're buying it for. Someone has a job that produced that item. And speaking of efficient markets, we talked about pricing, but I think there's always a lot of confusion when I talk to people about the purpose of derivatives,

options, futures, but those are truly the tools that make the market more efficient, right? Yeah, I mean, those are very important. We don't actually trade. I mean, we do trade some sort of derivatives, like we trade swaps, because we can't buy the equities. But they are very important. And they, yeah, I mean, the I do, when done well, obviously, it makes a lot of sense to have certain certain positions in derivatives. And I know of many pension funds and other things where if you can buy a put option to hedge your market risk, instead of selling all of that stuff, that helps you in many different ways. So derivatives are kind of just as good in some ways as stocks for the world. Right. And obviously they give you an idea of what people

at least assume price will be in the future by setting those futures prices at six months, a year, years, whatever. Right. So you touched on something really interesting earlier. You said in five years, if Numeri has all the data in the world. Okay. I mean, that was kind of, we, we brushed over that, but is that the goal? I mean, are your goals that big that you want to, you know, be managing all of the money in the world or have all of the data in the world? I mean, is that realistic? Is that what you're, what you're thinking about? That is the goal. Um, it would be weird if we said something like our goal is to manage, you know, uh, 1% of the world, like why one,

why not more? Uh, if you got to one, you would want to do two. So I think, um, if we think we have, if we think we can build something that has so much alpha and so much edge, uh, we would want to have the maximum capital applied to it so that the prices were reflective of all of the alpha that we had. What does that look like for every other hedge fund, trader, institution in existence? You're just one huge fund? I see it as Numeri Signals allows other quants or other hedge funds to submit their signal to us. So I see it as they might, let's say, let's say total hypothetical. You know, we make 30% a year for five years in a row. Let's do that. There's another hedge fund out there and they've made 6% every year for five years.

And one year they were down a lot, let's say. Now, why should they continue their business? They can see their beat. They can see their beat. And so why do they need to continue doing what they're doing? And the sad thing, what would happen now is that hedge fund would just shut down. Okay. Now, and those people will go get other jobs. But what I would like instead is, instead of that hedge fund would just shut down. Okay. Now, and those people will go get other jobs.

But what I would like instead is instead of that hedge fund shutting down, they just start submitting their signals to Numeri. So they become part of, part of us instead of sending their trades to the market, they send their signals to Numeri. And then we basically assimilate any intelligence that they did have, because we do want that 6% too. And so we would, we would like, we would basically, so that's how I see it. As we grow more and more hedge funds will submit to us. You consume, I mean, you consume their, their signals and improve your own system and it benefits everyone. So do they make more money by doing that to you through your staking and through the incentives offered by you for them to submit?

Yeah, they do. And it's already quite severe how much this is the case. case so we've seen people upload their signals to to numeri and they can make 200% per year on a signal that has just 1% correlation with our targets so if you can do well on numeri signals which is hard but if you can do well even a little bit the returns would be so much higher than you could make trading it yourself in fact trading it yourself you might not make any money because of trading costs and so on numerize signals how would we how do we get all the data in the world

how do we gather all those signals in the world we make the system that has the best incentives to submit signals to. Right now, the only competitor we have is the market itself, which charges fees. It's expensive to get into. There's huge barriers to entry. On Numeri, it's free. You can upload your signal and stake as much as you want of our cryptocurrency, NMR. Roundthex.com is one of my favorite companies in the entire crypto space. What they do is take all your small purchases and round them up to the nearest dollar and invest that spare change into any of over 30 crypto assets of your choice.

They integrate with your favorite exchanges so that you can view various exchange balances all in one dashboard and round up into different assets all at the same time. And they do all this without ever holding any of your Bitcoin. This is by far the best way to dollar cost average into Bitcoin. Go to roundlyx.com and use the promo code wolf for $4 in free Bitcoin after making your first roundup or purchase. That's R-O-U-N-D-L-Y-X.com and code wolf for $4 in free Bitcoin. Diginex is making huge moves and is soon to be the first crypto exchange listed on the NASDAQ. This means that people will finally be able to invest on a platform they're comfortable with without being directly exposed to Bitcoin. Absolutely massive for mainstream adoption.

Diginex has basically everything investors need under a single roof, including an institutional-grade exchange called Equus. Equus allows institutional and retail investors alike access to an exchange that's on par with platforms they've come to trust in other markets. This means they are compliant with regulation, transparent and fair with regards to fees and orders, secure, and far ahead of the curve in regards to innovation. Go to equus.com slash wolf to get 5% off trading fees. That's E-Q-U-O-S dot com slash W-O-L-F to get 5% off of your trading fees. Sign up now. Unless you've been living under a rock, you've heard about the DeFi craze in crypto. By far the safest and simplest way to passively earn in the space is to hold your coins on Celsius.

You can earn your rewards in the same crypto you're holding, or you can earn even more in their sell token. Right now, I choose to earn 5% on Ethereum in Ethereum and 15% on my stable coins in sell token. It's a little bit better than the sub 1% interest rates you can earn in a legacy bank account. Celsius was founded with the belief that crypto is the opportunity to really shake up the financial system. They're changing the standards for all financial services. They share 80% of their revenue in the form of weekly reward payments. That's how their users are earning up to 15% APY with compounding rewards. They also commit to providing the lowest cost loans on the market. Their loans start at just 1% APR. For just 1% interest, you can borrow cash against your crypto and avoid

selling, which also eliminates the taxable event. It's absolutely huge. High rewards on your holdings and low interest on loans on a platform whose mission you can believe in. Celsius is giving $20 to every new user that joins with promo code WOLF. Just enter the code in the app during registration. $20 is awarded after 30 days of maintaining a wallet balance of $200 or more. Visit Celsius.network, that's C-E-L-S-I-U-S.network, and use promo code WOLF, W-O-L-F. Okay. I mean, that's a perfect segue. So, you know, you're not trading crypto. I mean right you're trading stocks but the the system itself the incentive the incentive system is based on your own cryptocurrency why did you structure it that way and why is that a better way than just giving them a fractional percentage of the fund or

just you know paying them in dirty dirty fiat as we like to say in this space. Yeah. So the first thing is just like why we made it in the first place and it's the staking thing is really valuable. So if you have a normal hedge fund, one of the reasons the hedge funds work is because the people inside the hedge fund invest in the hedge fund. I have to. Yeah. I have a lot of my own capital in numerized fund. My employees, a lot of them have money in the fund as well. So we are very incentivized to, you know, be good actors and try to help the fund go up. But what happens on the internet is the sort of threat model increases. There are people out there who will submit random predictions, who will submit signals that are backwards, so that they're trying to hurt us in some way. And they can also make a thousand accounts on Numerai and submit bad predictions on all of those accounts.

And hope that just one of them got lucky. So how do we solve that problem? Well, we have to solve it the same way the market solves it, which is you have to stake something. You have to put some skin in the game to say that you believe in your signal as well. And so that's why we created Numerare and we gave it away for free to our users

so that they could stake it to prove that they believed in their signals. And what happened? The models that aren't staked do pretty well, but not that well. And then the models that are staked, way, way better. So we even weight our whole model by the amount people stake because it's such a good indicator of how much they think their signal is good. So that was the reason we created Numerare. And then we also pay people in Numerare now.

We used to pay in Ether and Bitcoin. And now we pay people in our own currency because their own cryptocurrency is now on Coinbase. Yeah. I know. For anyone who's not watching is listening. He's wearing a Coinbase t-shirt. Yeah. So, you know, it's, it's kind of, it's much more liquid than it ever was.

So we started paying people in our own currency. And what that means is, yeah, as soon as we pay them, they can put it back into their stake. And that means we've had tremendous growth in staking. Just a year ago, we had $30,000 staked. And now we have $4.5 million staked. Wow. Right. I mean, the entire appeal of proof of stake or staking is that it aligns incentives, sort of what you talked about before with traders, right? I mean, you can be selfish and care about making money, but you're encouraged to behave in a manner that benefits the entire system because if you don't, you get basically kicked out or you get punished, correct? So it's so interesting. So where does your supply of NMR come from and how is that managed yeah so we

like i said we didn't do an ico we gave away coins to our user base so we gave away about a million uh coins to our users and they're 11 million total um and then since, we've given more to our users and sold some to some crypto investors like Union Square Ventures, Placeholder, Paradigm. Big crypto funds have bought some. And so we have the rest is about six million NMRs that we have left. And all of that is kind of allocated to to give a future price basically so it's a lot it's kind of no joke I mean six million the coin is on twenty six dollars six million you know it's like much more than a hundred million dollars yeah we have reserved that you can go and see on the blockchain we have and

it's all waiting to be given to people who submit signals to us. So effectively, you're waiting to give away over $100 million. Yeah. And it might go, as people stake, it's possible the cryptocurrency could go up too. Of course, it's very volatile. It goes down a lot too. One point was down more than 90%. It's extremely crazy out there for anyone in crypto. I don't even encourage you to become a crypto investor. It's pretty scary. But yeah, Numerair has had this incredible growth in staking and our users have started to be like, well, look, even if the cryptocurrency might go down by the end of the year, might be down 30%,

I'm going to make 200% by the time that happens. Also, who cares? I mean, unless you're doing it to sell it tomorrow, if you really believe in the system and you're a part of it, I mean, it's also why you don't sell stock when the market drops 35%. It's not really that different. Right. I think there are users who do want it anyway, and they sort of said, well, I'm willing to put $5,000 into this. And now they have $50,000 worth of it and they're keeping it. They're not selling it down. Right. So I read about the Numeri master plan, which I think is somewhat essential reading for anyone who wants to understand what you were doing and it's phased out. Can you talk about what that

plan is and where we are in the timeline? Yeah. So the master plan is really simple. I remember watching this TED talk, it's pretty good, with Jim Simons, another founder of Renaissance. And he said, it's basic, you know, to make a good hedge fund, it's data and talent. And so that seemed really simple, but it's really, it really is that simple. And so our master plan has three, four steps. The first step is monopolize intelligence. Get as much talent on our platform as possible. Get a good chunk of the top Kaggle users, the top data scientists in the world, onto Numerai, which we have.

Get the winner of the Netflix prize onto Numerai, which we have. Get some of the best Quantopian users onto Numerai, which we have. Get some of the best Quantopian users onto Numerai, which we have. So monopolize intelligence was step one. And step two is monopolize data. So we had, everyone in the step one was just modeling the data we have. And now in step two, you can bring your own data to make your signal from any data set? And step three is monopolize money, which is kind of doing the sales part of the company, going out to investors, getting them to invest in the company, which we haven't really done. The fund is basically just my money. So we need to go get a lot of these institutional investors into our fund. And then the fourth one is decentralize the monopoly. So have it be that the Numerare holders of the, the holders of the cryptocurrency can start to govern and move the whole thing forward. is that we won't have a team. We will slowly have no employees kind of at the company,

a little bit like Bitcoin has no employees. And the master plan is really long. I mean, it's like we're talking about a hundred year plan here. I don't have anything else I want to do. So when I say monopolize money, it sounds kind of crazy, but there are many hedge funds that have gone from zero to half a trillion in a matter of 10 years. Maybe we can do all the money in the world in 100 years. Sounds like the Skynet of finance. By the end, it's a self-feeding machine. We're all going to be in the matrix while you guys are outside trading. But interesting, I mean, as you quoted him from the TED Talk, saying obviously that there were two things, data and talent, right? Well, what about, I mean, there's a lot of hedge funds that have, I mean, I went to Penn, right, in the 90s. So everybody went into investment banking.

It either landed in private equity or a hedge fund at some point you know since i graduated 20 something years ago that's so hard to say i'm sorry um what about the guys you know the the old school hedge fund analysts who call a ceo and ask how it's doing and go play golf and and have dinner and they listen in on earnings calls and those things. Because, I mean, that's still, I mean, that's still a large chunk of how Wall Street operates, correct? Yeah, it is. And I think it's a good, it's a valuable part of how it operates for now, because I do think that those, some of those investors who really understand the companies are adding some information edge to the market. If you really

understand that Google is worth a lot more than $50 because you work there or you understand everything about the advertising business, that's really good that you go out and buy it. The thing I would say though is that is data. Right. But how does that data land in Numeri, I guess, is the real question. Yeah. So what we would want is for those people, let's say they did have, maybe they don't know about every stock in the world, which is kind of the main way we get signals, some data that works on all stocks, like a P ratio or something. But if you did know about 50 stocks, or 100 stocks, you can still submit that to us. And you just give us, you just kind of order them, you say, I'm going to give this one a 0.9 rating and this one 0.8, and this one 0.7. And you're basically getting all of your intelligence into, into the

market. It's not really the way that most fundamental investors kind of think, but because they thinking about maybe a very long-term investment or, but there is nothing in principle to stop all of whatever their insights are being condensed into a signal and submitted to numeri. So that answers the question. So they can provide signals and you kind of dumb down the system for it to be effectively just a simple rating system. Because I mean, a hedge fund analyst, like you talked about, I mean, their job, and they can make millions and millions doing this, is to understand like 50 or 100 stocks. Like you have your telecom guy, you have your, you know, each, you know, your, your internet guy, whatever it is, you know, the guy who understands every phone company and around the world and

they're specialists. So you're saying that eventually those people come to you with their knowledge instead of going to their individual hedge fund and their incentives are aligned to potentially make more money than they would have by trying to walk over to their boss and convince him that they need to buy or sell this stock. Yeah. I mean, we already have, you know, analysts earnings expectations is already a feature even in normal numeri. I mean, it's just a very common thing that quants use. So aggregating the insights of analysts is quite a well-known thing. But yeah, I do think it's probably not, it could be better and signals would be a way to do it too.

That's so interesting to consider that all of this could be effectively consolidated into one place because it's such big business and it's such an old and established system. I mean, where do you get the, I mean, I want to say, where do you get the balls to pursue a goal this ambitious? I mean, it's really, really impressive. It's not like most people would be like, I'm going to go start a hedge fund and I'm going to beat the market. They don't say I'm going to be the only hedge fund. Yeah. I mean, it's, it's feels strange to me too. I mean, if I read the newspapers or the wall street journal or something, it's all this kind of like sort of circus and theatrical like Bill Ackman is doing this.

Warren Buffett is doing this. It's very like human. And I just don't see the world that way. You can make, you can put all this into simple signals and you can take away probably 90% of the costs and the people costs. You know, there's way too many people working in finance and accounting and law firms than there should be. And I think, I don't know, maybe I'm kind of like an alien or something because I kind of see if aliens came to earth and they saw it, it would be so clear to them. This is crazy what you're doing. You're sitting in an office and the guy in the other office is trading against you and you're

not working together in any way. You should really be working together so that you both don't go into the market and both pay costs to trade. It's so peculiar. So I just think it's, I feel like I'm very inspired by, you know, Bitcoin and Ethereum and things like that, where they're thinking about things in a very different way to what was how people were thinking about things. And they're talking about, you know, governance and incentives and all these very interesting things. And there's no reason why, like Bitcoin is a bank, the internet bank or something. There's no reason you couldn't have an internet hedge fund. And that's what Numeri is. Right. And, you know, Bitcoin is obviously a huge threat to the banks, central banks.

I would have to imagine that there has to be some fear in you that you're a huge threat to the Wall Street establishment. I mean, have you gotten any pushback regulatorily or from other funds? Or do you fear that there could be some, you know, reprisal, the more powerful and more money, you know, more AUM and just the larger you get? I mean, yeah, like some of, some, sometimes people say, why are you saying you want to monopolize money? You know, this is like a scary thing to say, but I think I'm talking about it in a kind of,

in a kind of abstract way. Of course, you're not right. Academic way. And ultimately it's extremely competitive industry. Yeah. and we have, you know, so little money in our fund. We are really the little guys in a very, very big competitive space. And I don't think anybody sees us as a threat. But they will. I mean, but, but if you reach even a fraction of sort of your overriding goal, I mean, they're going to take notice. Yeah, I think they will. I think they will. But for now, yeah, we have a good, I don't know. I mean, we do, people do like some of the stuff we're saying.

I think there's a lot of very ambitious blockchain projects. But somehow the pieces don't really connect. But I think when, even when quants look at Numerai, they're like, that could actually work. Not sure it will work, but that could work. And so, yeah, I think it remains to be seen, you know, how things evolve with what's possible. But ultimately, if a new technology comes in, like blockchain and machine learning, and the U.S. doesn't want an innovative new company to change an industry because of those new technologies,

that would be very weird and sad. And I think we will have a lot of support from people to make changes in this industry. Hedge fund industry is not a good industry. It takes away more money than it gives in a lot of ways. The hedge fund managers are taking the money. So we can make a much lower cost system. We could have lower fees than any other hedge fund because we don't have as many employees as other hedge funds because all the people are working on the outside.

So there are a lot of reasons why this is kind of the right way forward for the industry. Yeah, it's so interesting. And it's something you hear across, I mean, across crypto for sure. Like I just had Alex Pshinsky from Celsius on the show and he said, listen, we're not doing anything crazy. We're lending securities. We're just actually giving the money to the people instead of just keeping it like the other banks do. Right. So there is so much fat to be trimmed, obviously, in all these industries. But and what you're saying, I mean, rings true for almost every industry. So when I think about it, it's not so crazy. You would try to do this with a hedge fund. I mean, technology is inherently deflationary in that manner right i mean technology is going to come and take jobs in every industry and it's right it's gonna and theoretically should reduce the prices of things and make you know less salaries and all those things so i guess as crazy as it sounds it's the natural obvious path of where everything's headed yep it is and um and i think the the

the spirit of all these things uh blockchain companies is so good. I mean, look at, I don't know, something like Uniswap or whatever. It's this decentralized thing. They don't charge any fees. You can use it anywhere. And it's coming from the right place. It's not coming from a dark place. And the same with Numerai, like even though we want to make a very good hedge fund ourselves, it will always be open to anyone who wants to contribute and earn just as much money as the next guy. It'll always be a meritocracy. We are always looking for new people to join. So that idea of an open hedge fund is very different to what you see in Wall Street, where it's extremely secretive.

There's all these contracts. If you leave a renaissance, you can never work at anywhere else again. There are people who've quit Two Sigma, who've gone to jail. It's just not a nice place. And that could never happen on Numerai. You can always do whatever you want. You can trade your signal yourself and submit to Numeri. You can use Numeri, you can use our data set for free. There's so many things that are kind of in the right

framing that I find it hard if someone doesn't like what we're doing. It's interesting because, I mean, in theory, Wall Street is supposed to be the beacon of capitalism and this free market. But as you describe it in the way that it actually operates, it's not that at all. But what you're offering is an actual, like you said, a meritocracy, a free market. If you have the best signal, you're going to get paid. Exactly. Exactly. So it, yeah, to me, it's the right, it's the right approach. And what's also nice is I remember being in a quant fund and I had signals that I thought were working great and they were working great.

And for a whole year, my signals were, we were sort of paper trading them and we kept seeing how good they were, but they wouldn't deploy them in the firm until, you know, they had more evidence. And until I got approval from various people inside of the firm and you know they had more evidence and until I got approval from various people inside of the firm and blah blah blah and it was just like this very slow process for my models to reach the market and on numeri every single week we start from scratch because users submit new new signals every single week if no one showed up we wouldn't be able to trade.

Because all of our models are based on our users. So that means any new insights can enter our system immediately. So none of it is internal. It's not your own signals. Oh, wow. So it's all completely crowdsourced. Wow. So you're saying it's done every week.

I mean, that's clearly some sort of AI machine learning. I mean, it's not like a couple of guys in a room crunching data, right? How is that done? And as you scale, how are you able to parse all that data? So yeah, every week people provide basically a CSV of their signal. And so we get hundreds of those across thousands of stocks and we pretty much average them together based on how much they're staking and trade that in our hedge fund. And the reason it's every week is just because we just do a weekly, that's just the way, but it will become every day and it might become less than a day at some point. And then all of those trades go to our prime brokers. So yeah, it's kind of simple in a lot of ways. And really the good, the interesting things are happening outside. It's like, how did this person make a signal so good? How did they do this? We don't know.

We never see the code because like I said, all they do is send us a CSV of their signal. So we don't know what they use to create it. It's so interesting. I mean, with your own background, I would think that you guys would be trading your own signals as well, to some degree. Well, there we were. But then my signals are worse than the meta model by a long way. Right. So, I mean, it makes sense. As you said, if you can improve it by 1% with even a crappy signal that just works in the system and, you know, it optimizes it. And

wow, man, it's just brilliant. Really kind of blowing my mind a little bit. I have to ask you, so why do you only trade stocks? Like why not? You know, obviously like Forex is the highest volume, you know, market in the world. And certainly you could trade Bitcoin or other things. Why, why are you only applying it to stocks? It's sort of like length and breadth. So length, there's a lot of stock data. It goes back about two decades of really good global equity data. And then there are also lots of stocks per month. So there's a lot to choose from.

So both of those things. So you need length. You need a lot of data to kind of do machine learning. If I were to say I've made an amazing model on the price of Ether or the price of NMR, well, it's necessarily only got three years of data because NMR is only three years old. Right. So you want to have more data and you want to have more breadth. So, yeah, I mean, global equity equities is a very good place to start if you if you care about that. Now, there's some some other types of models. If you have some kind of momentum trading like trend following kind of system where you're just using the price graph to make your trades. That's different. We're using all sorts of fundamental data and there's lots of fundamental data available

about stocks, but not really about crypto. Where do you store all this data? In the cloud. So you don't have it. It's not like 20 years ago. You don't have a guy in the basement who's, uh, sweating with the fans on the, on the servers. It's just, that world is over now. Um, how much of your fund is, how much of your capital is deployed at any given time? Uh, about 500% of it. Yeah. So we have five times leverage. Yeah. Uh, and, uh, and, uh, yeah, so uh and uh and uh yeah so two and a half times long two and a half times short we're always neutral neutral neutral yeah neutral to everything

we're neutral to um value momentum size any bar a factor and we're also neutral to the market we're dollar neutral we're beta zero it's really a kind of pure quant fund. We don't take, we don't say, we don't have any views. Like we don't have a view on where the market's going to go or anything. Don't care. Yeah. Right. So your signal, you're not contributing to the signals, but the strategy is very clearly has been defined and developed by your team. Exactly. So one way of thinking about it is the data we give

is in some way the definition of the strategy. Because if we give a data set where the target value is one month forward returns, then there's no way to use our data to make a model that's one week. For five years. Yeah, right. So we're in some ways putting in some of our knowledge into the data so that any model trained on that data is decent.

And then you can make really good models on that data as well. So that's how we see it. It's kind of like the data is deciding the data we give is deciding the kind of nature of the strategy. No one can come and say, oh, I have a model that's really good on a certain subset and use that on numerized data. Maybe they can use that in signals. We'll see. But right now, numerize kind of like set, in a way, set the strategy and the users are optimizing the strategy. Right. So you're not going to get like one week where there's 2000 long signals are submitted and no short signals

and then not be able to play with your strategy or something. Exactly. So they are, if you think, if you look at the target values, there are as many ones as there are zeros. And the reason for that is we want to be as long as we are short. And if there's no tilt there, so that's a kind of way that no machine learning algorithm would ever want to

do, not do long short because they would. Right. So obviously it doesn't play into your strategy completely, but what do you think of the market right now as a human person, as a human being, as an American citizen, not as a, uh, you know, as a quant. Yeah. Well, I, um, I did, I do, you know, sometimes look at the markets. I was, uh, I did, I invested in some cryptocurrencies and sold, uh, uh, out of a cryptocurrency hedge fund in the end of 2017. So I ended up having some cash well-timed like you can't you can't get away from being a macro trader right in your personal life you always have to so

I had this money and I never wanted to put it into market and then when the covert 19 remember seeing the VIX on 80 yeah it's on a river it's gonna revert to the mean eventually right yeah put it all Put it all in. Oh, nice. And what, what, when you put it all in, are you buying just the stocks you're interested in? Did you do it with funds? Did you do it with options? I mean, or was it just like this is time to buy companies that are good companies that are down 40%. Yeah, no, it was good time to book, to just buy. I mean, you want to buy,

you don't want to buy options when it's so volatile, it'll be expensive options. So you want to just options when it's so volatile. It'll be expensive options. So you want to just put the money in. So yeah, I put it in the S&P. I also don't like buying individual stocks because I have to like, we trade individual stocks. So there's a compliance issue. Oh, right.

So you just like buy SPY or something? Or I mean, whatever you call it. Maybe I could have been more thoughtful and put it in maybe the NASDAQ, but SPY was a pretty good return. Oh, yeah, pretty decent. Yeah, a good decade in three months. If you timed it right, right?

I mean, even people who did nothing are fine, which is amazing. The real pain is the people like we talked about earlier who panic on March 23rd instead of on, you know, February 23rd or even April 23rd. Yeah. No, I remember the Vanguard. I have the Vanguard account and the CEO, they sent like a message like, yeah, don't panic, don't panic. And you can see they're panicking. Yeah. I got to say, I have Vanguard as well. I got the exact same email and thought the same thing. Like don't ever use the word panic when you're talking to your customers

it's like okay so you can see they're panicking and they're telling you not to panic so it makes you want to panic so what do you do don't panic yeah that's actually the lesson that's a signal all right yeah the like the like four-piece string quartet on the Titanic is not convincing me that I need to stay on this boat. Yeah, it's amazing. So I know we're up against it a bit with time. Are there any parting thoughts that you have? Anything you want to tell us about like what we can look forward to in the immediate future and then also where people can participate and follow you personally as well? Yeah, well, I would say anyone with a signal, like a quant signal, go to signals.numer.ai,

try it out. You can upload your signal on historical data and get an assessment quickly on how good it is. We do all this stuff, neutralization, to figure out if your signal is original. So it's also a very interesting way to do research. Just check against all of our data if your signal has any edge. And yeah, try that. Or go to Numeri to try download our own data and do machine learning on it. And yeah, I mean, I'm very happy with signals. It's only 11 days old. It's already as big as Numeri was a year ago. Oh, wow. In terms of amount stake, there's about $30,000 stake there now or something. So I think it's going to do really well and it's going to be really helpful to our fund and master plan. Can someone buy NMR on Coinbase and stake it without participating at all? I

assume yes, that you can still stake the coin. Not really. I mean, we have people say, well, I kind of just stake someone else. I don't know anything about, well, that kind of ruins the point of it because we need to know you believe in your model. That's why we're asking you to stake. It's not a way to just, I don't know, some of these DeFi things, we just stick it in, you get 300% a year. I mean, that's why, but I think when you say staking, most people think it's just literally parking your money and, and having the confidence that it, it's going to grow, right? I mean, that's what it is. It's not really as active or you vote every month. Something comes up. It's like, should we do this or no? Yes. Okay. I'm staking, right? Yeah, no, exactly. It's a, it's

usually pretty idle capital. And what's nice about all the stakes on Numerai, they're all powering real models that are important to us. So if you really wanted to, you could kind of stake our example predictions, which is just an example model we built on the data, if you wanted to start that way.

But mainly it's for people who are contributing, and we want to be the coin that the people who hold it are the users. We don't want to be a coin where the people who hold it are the speculators. So there is a sort of sense where we're intentionally trying to make it be real and not a speculative coin, a real utility token. Very cool. Well, it sounds like you definitely are a potentially successful

david and goliath's story um i hope you can take down uh the wall street goliath for sure it would be a quite a story to tell for for generations to come thank you so much for uh taking the time to do this and oh and by the way where can people you? I don't know if you said that. Oh, yeah. I do use Twitter, Richard Crave, at Richard Crave. And then Numerize Accounts, at Numerize. Awesome. Well, thank you so much for taking the time. Go back to changing the world. Thank you.

Have a good one. Let's go. Bye.