podcasts

Creating ‘the Last Hedge Fund’, via the Hardest Blockchain Powered Data Science Tournament in the World, by Richard Craib Founder of Numerai

podcasts

Creating ‘the Last Hedge Fund’, via the Hardest Blockchain Powered Data Science Tournament in the World, by Richard Craib Founder of Numerai

August 2020

Posted by

Jamie Burke

CEO and Founder

As an early investor in Bitcoin and Ether Jamie went ‘all in’ during 2013 founding Outlier Ventures, Europe’s 1st venture fund and platform dedicated to blockchain and Web 3....read more

We talk about building a 150 year startup in Web 3 from first principles with Richard Craib Founder of Numerai and why he chose to apply himself to solving what he sees in the ultimate mathematical problem: creating ‘the last hedge fund’, through ‘the hardest data science tournament in the world’. And his approach of crowdsourcing equities trading models through a network of data scientists and the power of the convergence of blockchain + AI in open data marketplaces.

Posted by Jamie Burke - August 2020

August 2020

Posted by

Jamie Burke

CEO and Founder

As an early investor in Bitcoin and Ether Jamie went ‘all in’ during 2013 founding Outlier Ventures, Europe’s 1st venture fund and platform dedicated to blockchain and Web 3....read more

Key Themes:

  • Applying first principles to startups
  • Building a 150 year company, that transcends the founder
  • Decentralised Data Marketplaces
  • The convergence of AI + blockchain

Listen on iTunes

Transcript:

Jamie Burke
Welcome to the founders of web three series by ally ventures and me Your host Jamie Burke. Together we’re going to meet the entrepreneurs that backers and the leading policymakers that are shaping web three. Together we’re going to try to define what is web three, explore its nuances and understand the mission and purpose the drivers founders. If you enjoy what you hear, please do subscribe, rate and share your feedback to help us reach as many people as possible with the important mission that is web three.

Today, I’m really happy to welcome on Richard Craig, founder of Numerai. I described as building the world’s first open hedge fund by modelling the stock market and what you promised to be the world’s last hedge fund and what you promote as the hardest data scientist All men on the planet. I know that’s kind of a subset of the activity that you’ve done. So welcome to the show. Thank you. So I’ve seen lots of different descriptions of humour, I guess, because there’s, there’s different components to your activities. But the way that people can to ease people into this, the way that they can think about you as as a new kind of hedge fund, built by a network of data scientists, and it’s really this approach that sets you apart. So the reasons why I wanted you on the show is, as I understand that you’re a first time founder, it seems like you’ve only really ever known what we might now call web three, so you’re truly native to the space. And as an entrepreneur, I guess you’ve grown up in with the industry, you’ve evolved with it as an entrepreneur and as a founder. And so I think it’s interesting to get a generational perspective on the space we have a really wide range of founders on here from serial entrepreneurs to people like yourself, and So I’m always intrigued as to how somebody who’s who’s kind of only knows this space have their perspective on it, I think, I don’t know, this is a fairly nebulous term or that might be a kind of stricter definition of it. But you’re you are arguably one of the first defy projects. Before it was cool and important. You’ve done that all out of the US, which is unusual, and try to understand without going to jail, so congratulations on on that. And you’re also one of the first startups to combine blockchain and AI or machine learning. And that’s certainly what caught my attention. Early on, given our convergence thesis at outlier and how we view the world. You’ve also consistently secured backing from big names in the industry from Union Square ventures Coinbase, while the founder of Coinbase, co founder of Coinbase as well as co founder of Renaissance, the kind of centralised finance world or traditional finance world throughout all of these ups and downs, which is They’re really good testaments. And I’m intrigued to understand how you’ve got that, as an entrepreneur, you’ve got that stickiness with the top VCs, you’ve also raised at different stages in the cycle. So in total, you’ve raised about $40 million through two separate rounds. I know you wouldn’t consider it as an Ico because it’s more of a private sale. And one of those was recently for $3 million, literally, June, the second 2020. So it’s going to be good to understand what’s the same and what’s different as you’ve gone through those cycles. So I’m really looking forward to going deeper with you. So if I look at your kind of origins, you came from Cape Town, the accents, not that strong now actually, you’ve kind of Yeah, why is analysed?

Richard Craib
Yeah, I did study in the US. So I think I’ve been here, maybe 10 years total now.

Jamie Burke
Okay, right. So you did mathematics and economics in 2007 to eight University of Cape Town, went to Berkeley as an exchange student studying mathematics. In economics 2008 2009 and then Cornell, where you did a Bachelor of Arts in mathematics 2009 to 2012. And then you went straight in to Numerai, October 2015. And am I correct in understanding that you first time founder, this was your first

Richard Craib
first? No, I was a solo founder, but I’m not a first time founder. Okay. He started companies when I was 17. Even. So I’ve been starting companies for a while. Some are still going actually, there’s a big data science company that that I started in South Africa, which is doing really well just raised around itself. But this was definitely the first time I started something by myself. And it was the sort of my feeling going into where I was I was going to be starting the sort of company I’d work on for 150 years was always my plan.

Jamie Burke
So you’re gonna live for 150 years you but that’s also a plan. Yeah.

Richard Craib
That’s pretty As a part of the game, yeah.

Jamie Burke
So maybe we just pause on that, how the hell are you going to live for 150 years, you’ve gone vegan, you’re doing, you know, the keto diet, what what’s going on.

Richard Craib
So the reason it’s 150 is because my great grandfather started a newspaper, sort of story of like five pounds in his pocket, and he started a newspaper. And then my grandfather took it over and my Uncle Ben took it over. And so there’s this long thing and and kind of lost it like 150 years or so. Each of them worked on it for about 50 years in their careers. And so I think there’s somehow that tradition of doing a company for life somehow isn’t fashionable anymore or something. And I think that’s the that’s the right way to go. It makes you It makes you make better decisions. If you’re, if you’re hiring someone you wondering, I wonder if I can work with this person for 150 years. makes better. Um, Numerai is also kind of famous for having a chronic Onix policy. So we, we allow our employees to sign up for cryonics, and get frozen in the event of their legal death, but you know, not really, when they’re not really dead. So if you get run over by a bus or something like that, that’s tough luck. You probably can’t probably can’t do the cryo. But if you if you’re kind of on your deathbed, and you do it quickly, there’s some hope that you might be be revived in the future.

Jamie Burke
Wow. I mean, I guess if you if you think about it in the context of a family dynasties, or multi generational family business, where you have the crowd near I can try to send you right, and this is one of the main principles, exciting things, I guess, around open source permissionless systems is that as a founder, you you’re there to kind of the Genesis, but these things should evolve and have a life of their own. Does that kind of factor into thinking?

Richard Craib
Yeah, it definitely does. And even the number of people at Numerai sometimes you ask entrepreneur in Silicon Valley, how’s it going? And it’s like, oh, we just added 50 people 50 new engineers. And it’s like headcount is like an important thing. And that headcount means people, you know, inside the office. And for us, we kind of, I kind of imagined us maybe one day getting to like 20 employees, and then we go to 19. And then we go to 18. And we go down to

Jamie Burke
zero, freeze, freeze them all, as they kind of go right.

Richard Craib
But maybe you can do that. And that’s the point like Numerai the real work is being done on the outside of the company by the data scientists. And so we want that to keep growing indefinitely. But we are inspired by Bitcoin. And you know, Bitcoin is kind of like the internet bank. We want to be like the internet hedge fund. And so Bitcoin, the miners are mining to kind of keep the security of the bank and on numerous data scientists or data mining our data Trying to find keep the market efficient. And I think, you know, those similarities will become more clear in the future. We can’t be totally decentralised like Bitcoin, because we actually need to do trading on the real stock market. And so the money has to be somewhere. It’s with us. So, but I think in the future, it’ll be possible to do more and more of Numerai in a decentralised way.

Jamie Burke
Yeah, and I think this was one of the the generational things that I was alluding to, in that, I guess, when you when you were thinking about numerology, we get to get into the Genesis story there. You know what, why that, but I definitely the idea. I mean, it’s an equivalent of a gig economy for the financial services. I know you’ve now extended that beyond financial services for kind of data marketplaces in the data economy and what have you. And I could imagine that and I’ve said this a couple of times in other podcasts that there is a generation born and presumably you’re in it which Probably never once will have a job will always kind of be an equivalent of a freelancer, you know, moving in and out of these these different systems. So, you know, clearly you’ve got this entrepreneurial background actually didn’t know about these earlier startups. So that’s really interesting. And, you know, you’ve you’re kind of very well schooled, you’ve got all the right things on your CV, presumably, the typical direction that that career would go is you’d end up going being a quantitative hedge fund, and probably have a much easier life not having to solve for all of these problems, because it’s kind of solved for in in a centralised fashion. So, what was it that was that have a possibility that you would end up on that career path or why did diverge?

Richard Craib
It is it is how I started briefly. So I was I did start a lot of companies in high school and in college, and then after graduating, I worked at a quant fund. in Cape Town, and it was pretty cool because they, they didn’t have any machine learning. But they had some coolant. They also had fundamental investment. And it was a good place to learn. But at the same time, while I was there kind of learning traditional finance, I was also I remember reading the theory and white paper back then, and, and so I was, yeah, I had sort of enough free time to be able to connect these ideas together early on in like 2014 or so.

Jamie Burke
And so what was the thing that you’ve referenced this idea that you’re collectively solving the hardest data science tournament on the planet? Was it the problem space? or What was it about the problem space that drew you to this because you’re clearly a smart guy. You could have applied that into any other different problems. Why? Why this problem?

Richard Craib
Yeah, the finance is a procedure being being hard is already a compelling thing. To one I work on it because you know all the good arguments for why the markets kind of efficient, efficient already, everybody. If everybody has the same past data, wouldn’t they have found the same patterns that you’re going to find? And wouldn’t they already correct them? So that anything you can learn on the past maybe doesn’t work out of sample on real life trading. So that was a big draw, like, could you how do you set machine learning up to work in that framework? It’s very hard to make machine learning work on it on this kind of non stationary time series data, because you look at the past and it doesn’t look like the future and and there’s sort of no times in history that look quite like the Coronavirus, financial crisis, for example. So everything’s kind of always new. That’s a very inspiring thing for a lot of data scientists too. And so if you go to a data science problem, like on kaggle, or something like that, it’s actually quite easy to you can do these data science problems where you like modelling what Sounds or modelling and the these things actually have solutions and you can get 98% accurate or whatever, on numerology, you’re kind of very lucky to get 53% accurate or something like that. So it’s a much harder problem. And that makes it more interesting.

Jamie Burke
So if we kind of focus on numerous the business first, could you talk us through the fund itself? You know, how much you have under management? And the thesis? I don’t know if you would describe it this way. But the idea that the crowd can solve for this problem more effectively, more efficiently than a centralised equivalent? Is that playing out to you outperforming the market and peers?

Richard Craib
Yeah, so the way we talk about So what we have is this huge test set. And that means like, it’s like four years of data that none of our users have ever seen. And we have our own internal model that we put a lot of research into Frankly, our own until the model is pretty good by itself. And so the question is, can the crowd when you average all the users together? Do they tend to make a model that’s even better than the one we can make internally. And that wasn’t true at noon, right for a while, it took a couple of years for us to get to that point. And now, the user performance is is significantly higher than our internal model. And so the users can take our Sharpe ratio much higher, take our returns much higher. And so it is working quite well. We couldn’t get the Sharpe we had without, without relying on the users.

Jamie Burke
And so during that period, as a founder as an entrepreneur, you know, you had this thesis, and in absence of data, there’s a bit of a leap of faith there that it will play out over that time series. Is there any point you kind of get nervous, you think, Oh, you know, maybe maybe it’s not gonna play out or how do you how do you work through that?

Richard Craib
Yeah. almost to the day that Numerai I started, there’s been something called Conte winter. Before crypto winter, this is like December 2015 or so. And some of the factors that people use to model these things like value in particular, the value factor, just haven’t worked the same way. So the growth of the performance of market neutral hedge funds is kind of like the straight up line. 2014 was great. 2015 was great. And then suddenly, it just flatlines and for the last few years, it’s been much harder to make returns in a market neutral fund. And so that’s been a big challenge in the way we first set up Numerai we were quite exposed to those kinds of factors. Like if those factors weren’t working, we weren’t going to work. Because we were playing in Kwan winter. We had to redesign everything and make almost make the whole problem much harder, so that whatever was learned would work. Kind of all environments. So yeah, it’s been a terrible time to start a corn fund. And I think things would have, things would have gone a bit faster if we had started in a in another time, but also to have the to basically start on the hardmode. More like godmode it was it was very, excruciatingly difficult to make it all work, but I didn’t really ever think it wouldn’t work, because there’s almost like a mathematical principle behind Numerai. So you don’t have to worry that the mathematical principle isn’t right. It’s it’s a proof, which is basically if you can make a number of uncorrelated models that all have performance, that’s going to be better than if you can if you can make one model and that is absolutely true in practice in theory.

Jamie Burke
So, if we look at the business, you manage institutional grade long, short global equity strategy, you In a hedge fund model, what percentage of that is digital assets? versus you know, equities? Is it just pure equity?

Richard Craib
Yeah, zero percent.

Jamie Burke
Wow. Okay, interesting. And why is that?

Richard Craib
Because there’s not enough data yet for, for crypto. And also the strategies that work in crypto apparently like very naive, like, I’m sure you can make a crypto coin fund that goes up, but it will be doing kind of low tech things like I bought it on one exchange and I sold on another exchange, it’s just like arbitrage you don’t need machine learning to do arbitrage. But if you have decades of stock market data and you have 1000 different stocks, you can you can really learn things and that’s where machine learning really shines is if you have a large data set, so the crypto days that’s too too small. I don’t see us trading crypto for a while maybe at some point,

Jamie Burke
maybe 50 years and you know, 150 year journey you start doing it maybe 40 Yeah, yeah, I look, I guess that makes sense. Because otherwise you’re just taking advantage of short term, short term opportunity or small window rather than building this hundred and 50 plus year strategy. So how much is under under management now within the fund? And can you can you walk you tell us about the performance of it?

Richard Craib
We can really talk about those things. Actually, there’s regulations about talking about funds, because we don’t want to be seen to be marketing to people. We also not really open to individual investors. So but yeah, the aim is still pretty low. There’s another regulation, which is kind of in conflict with the first regulation, which is, you have to announce how much money you have in your fund if you have more than 100 and 50 million. So I think it’s under 50 million, so we have less than that. Right?

Jamie Burke
Okay, I get

it. Good, good. Good benchmark.

So Then this data science tournament aspect. I mean, this isn’t. This is a specific tournament rather than how you describe what generally happens in the network, right?

Richard Craib
Yeah, the data science tournament is modelling the price movements of all the stocks in the world.

Jamie Burke
That’s ongoing. Right?

Richard Craib
Exactly. It’s every single every single week users provide new predictions to us.

Jamie Burke
Okay, so you would describe ultimately what’s happening is is one big data science tournament. And so that started in December 2015 had some kind of stats I’ve managed to pick up through through media. At one point it was reported that there were 7500 faceless coders paid in Bitcoin. I can’t remember what year that was and who it was by, but there was this idea that it was his Bitcoins. I imagine it was a while to get an app because presumably you’re paying people out in your native token. Can you talk about the community and It’s scale and how you might describe some of the different types of stakeholders that make up that ecosystem in that game.

Richard Craib
Yeah, I think it’s more like now, even as much as like 40, or 50,000, sort of signed up users. But the number we care about is like number of stakes. And the people who are staking are the ones who are saying, not only do I think my model is good, but you can burn my cryptocurrency if it doesn’t perform well. And that negative incentive is very powerful to have along with a positive incentive. So in the past, people would sign up to new MRI and make lots of models and just hope they get lucky. But if you tell them, you know, you got to stake and will burn then suddenly they make one model. That’s really good. So yeah, the staking has been a really important feature and the people whose stake do a lot better than the people who don’t. The number of speakers right now is something like it’s was about 300 in January, and now it’s about 600. So these are people and they’re not staking a little bit like together, those 600 people have staked over $3.5 million. So some users are staking $200,000 on their models. I think there’s one guy who’s taking half a million dollars. It’s pretty cool to have that. Because, you know, there are many numerous employees who don’t even have that much money in our fund. But those users saying no, he believes in his model. He believes his predictions are going to work. And he’s put a lot of money to say that, that that that’s true. So that is the way to do this. You couldn’t make Numerai without the staking part. That’s why there is no there’s never been a kind of internet hedge fund to date, because no one solved the crowdsourcing problem. And the way we solved that was you have to put skin in the game, just like you would if you were in a hedge fund you You get your employees to put money in the fund. And that’s kind of what the users are doing. except they’re not putting money in the fund. They’re, they’re staking it on their specific models predictions.

Jamie Burke
Yeah. And it’s an often overlooked thing. You know, when people talk about how tokenization can allow for incentivization, they normally think of it in terms of positive incentivization. But obviously, having a cost to play is also a great way to filter and for you ultimately, whilst Of course, you want lots of participants, what’s most important is the quality I guess. So, how do they then participate beyond that particular transaction in numerology is it through the value that the value of the token going up based upon is that some linkage between the performance of the fund itself and how any individual can can share and be rewarded in that performance?

Richard Craib
Yeah, so the token, you know, kind of goes up and down with crypto a lot of the time. And so it isn’t, it isn’t connected to the value of Numerai, or the fund. And it goes up and down, kind of with demand. But what’s cool is that there is actual organic demand. So when people are buying Numerai, they’re buying it to us, like there is no way to use Numerai and earn money. Without staking. If a new user joins Numerai, he has to go out and buy some NMR to stake. And so having that organic demand, I think is the main driver of the token, and then they’re staking on their individual predictions. And that’s kind of important, like I don’t think we want in some of these defy protocols, you sort of stick money in the contract, and then you earn interest kind of for doing nothing. That is not how it works. You’re staking these tokens, and then you’re earning money because of the work you’re doing. And if you stop predicting you won’t earn anything, and so Numerai is kind of more valuable to someone who’s taking it than someone who isn’t. And that’s a very important property.

Jamie Burke
Yeah, interesting. So the recent 3 million that you raised was cited as being for the rollout or development of a razor Bay, which is your kind of next step in evolution. And as I understood it, this is taking some of the principles of what’s been happening in MRI, but applying it to data marketplaces more generally. Is that an accurate description? And can you talk about why that is important? and evolution of what you’ve been doing previously?

Richard Craib
Yeah, so arratia is just like the protocol, the name of the protocol we used to do all the staking and griefing. We made a verse version of numerology, like in 2017 that we released for Numerai and and then decided let’s turn it into a proper like protocol. Maybe other people can develop on to. And certainly, we would want new MRI to run on it too. So it’s kind of like something we’re building anyway, for new MRI. And once we launched it, we we decided, Okay, well, let’s make an app that’s kind of very different from Numerai to prove what the protocol can be used for. And that’s what arrays your Bay is, it’s a way to buy information of any kind over the internet with a bit more trust than you’re used to for the internet. And so the way it works is, someone can go on and put a request, while use a real example. I was I said to someone, I’m looking for batalik beerens home address, and I ate $1,000 or something like that on this request. And now, someone can see that I’m very serious about that. If I just put that as a tweet, they might think I’m joking or might think I won’t pay them if they if they provided to me. So bye bye. staking if they can see that’s a blockchain transaction, that money is actually locked up somewhere. And then they can put the work in required to get that information for me and know that they’re going to get paid. Okay? But that’s only one piece of it. So what about if someone provided the wrong information? So someone did provide the wrong information. He came and he put up his own stake saying I want to I want to give you a metallics home address, and he put up his own stake said, The metallics home address is vitalik dot Eth. He was kind of joking, but that wasn’t what I was looking for. He was given he’s kind of given all the money but I had the rights to burn it. So he was given the thousand dollars for providing the information, but I looked at it as like that I’m unhappy with that. So I had the right to burn it. So I burned the whole the whole thousand dollars and his steak at that transaction shows you the kind of watch The protocol works. So next time he provides data, I don’t think he’s gonna like mess or mess around with me like that. I’ve got a lot of crypto to burn. And then in the good case, it’s it’s usually just, you know, there’s no griefing required because the guy was the guy provided the exact information in the exact right format. So, arrays, your base already shown that, you know, erasure can be used for anything. It’s good for NMR because when the tokens get burned, they’re burning NMR. And it’s also like a neat crypto application that couldn’t have existed a few years ago. I think it’s got some potential for sure.

Jamie Burke
When also ask why you want to metallics address, you don’t want to send out a SWAT team or something that I’d imagine it’d be a more of a premium if you could figure out where Cz lives, but that’s it. That’s a different subject.

Richard Craib
Maybe I should do that now.

Jamie Burke
Yeah. So I guess the natural extension so well, actually. So firstly, is the idea that I mean, you’re just open sourcing this and you’re just gonna organically See how it develops? And yes, there’s some direct application for what you’re doing Numerai I, but other than that, it’s just watch it Watch out an experiment in how the game theory of something like this plays out, and presumably you’re gonna tweak that or you just kind of gonna leave it in the wild.

Richard Craib
Yeah, it’s a good question. I really like things that work automatically. I’ve got no interest in like business development. I like it when you know, and that’s what’s great about hedge funds is if you make a hedge fund that goes up, everybody wants to invest and you don’t have to do any business development, and no one takes their money out. And so it’s it’s really nice. Companies that require business developments, you know, just require different founders. Raise your Bay. I think what’s been quite cool about it is it kind of has grown automatically. It’s sort of like being doubling every month. It’s only been a few months. The stakes are going up. People are getting very interesting things. People have been asking for scans of people. both lungs have COVID. And that’s been successfully provided. People are asking for all kinds of things. And so I like that it is kind of working automatically. We’ve never run ads for raise your pay. And as soon as it’s all integrated with Twitter and all theory, and so it’s really easy to use. And anytime someone makes a post, it gets tweeted. So it’s kind of automatically viral. So I quite like it how it is. But there are things on the horizon that are very intriguing to me, for example, a theorem is making a new story. aetherium is basically a way to sign up a theory and application very easily. And they’re working on a new version that helps for other things that prevent arrays ubay from growing. So one of the things is what if you’ve never used crypto before, how do you get your first die, and that onboarding problem is being solved by other people. And once that’s solved, suddenly, arrays your base numbers will just flip even more. So suddenly, it’s way too easier. So the average person who signs up is way more likely to use it because they know they can get the dye. So those kinds of problems are being solved by a theorem, which is pretty cool. And I think that it’ll get better automatically raise your Bay. But I also have other ideas for other applications of a ratio. And I do quite like the I do kind of believe in the, if you’ve built a protocol, the easiest way to prove it is a good protocol is to is to use it yourself to have Numerai be powered by a ratio and have three and a half million dollars of stakes on it was just like more than auger and melon port and Dharma combined. That’s just running this code that anyone else could use to so I think it’s quite early. I’m not like, I like the approach of make apps that people use and then people will use the protocol.

Jamie Burke
At the outset. Do you have a strong opinion on How that would be governed. So let’s say you do believe that there’s something fundamental that needs to be tweaked for it to be more effective. How would that decision happen currently? And would that be a plan to kind of evolve the governance of

Richard Craib
Eurasia? I kind of think governance is a bit of a scan. At the moment. I think the thing we’ve always cared about, it’s been the least correlated with price. It’s funny, like, the things that no one’s using are the highest priced cryptocurrencies in the world, because they had a focus on press or on Park fake partnerships or something like that. But I think if you focus on usage, you kind of in the long run, you can’t be wrong. So we’d rather gets I think governance matters, you know, when like a million people are using it or something. And it’s very easy to at that stage, allow for that. So, so far, we’re just running it like a We’re trying to make apps that people use. And that’s the sort of framing of, of web two in a way. It’s like, hey, make an app and make people download it. Don’t write a white paper. Don’t do any of that. Just get people to use it. And I think that is that is the starting to become the right approach. Because usage, you can use these things you don’t have to speculate. You don’t have to write a white paper because the technologies are already there. Like no one’s stopping you. You needed to write a white paper, pre aetherium launch. Right? Because if you like Olga, they proposed the August stuff. Long before Ethereum launched, that’s when you needed the white paper. Now, you just needed to make something that people use.

Jamie Burke
Yeah, and I think that makes sense is something that we’ve advocated for in terms of, you know, if you think about these things as just startups, just like in a networked form, and therefore, you know, you would want to take that lean approach, you would want to be able to iterate evolve it fairly easily until you found fit. And then maybe you might consider how you might decentralise it a little bit more or whatever other way that you would measure different forms of governance. That’s actually quite a nice segue. So as I said, You’ve been in the space for a while, I mean, enigmas. 2015. Right. And I’m presuming you, you kind of in and around the space a little bit prior to that you’ve raised in 2015. And as I said, just closed in 2020. What’s the same and what’s changed? Have you seen the industry evolve? You’re just referencing this idea that, you know, pre aetherium a white paper might have been important. What are you finding? investors are most interested in now, as I said, You’ve done a really great job of consistently attracting the best VCs. It looks like they’ve been following on their money throughout. So, you know, can you talk us through how you fundraise and how you do Investor Relations and This kind of stuff.

Richard Craib
Well, so in the more recent rounds, I didn’t think I kind of always think people don’t want to invest. I always just assumed it’s gonna be this very difficult thing. And so, especially during the Coronavirus stuff, like we started raising this recent round, like in March of 2020, so it was kind of like a terrible time. And then, but luckily, you know, I was able to say, guys, this is just a formality, I’m giving you notice, I’m leading the round, I’m gonna take the whole round myself. And, but you know, if you want to invest, you’re welcome to. And then. So I think that puts up and that in this particular kind of environment, it is good to have, first of all, the founder wanting to invest himself. And then second, you know, like, also showing commitment to and be on the exact same terms as the other investors. So I think lately that’s been the main thing that helped us was just me, me being in the round. So I bought a million dollars of Numerai myself in December from the company, and I also bought half a million dollars in June. I think that’s one of the things that’s helping. Yeah, I think people also like investors also like to see that you’re kind of committed to something I think Nomura’s a unique company. In that first principle, since I described it won’t be true five years from now or 10 years from now that ensembling multiple car uncorrelated models is a bad idea. It will always be true that multiple ensembling multiple incarnate models is a good idea. So whether the Uber AI works or doesn’t work fact is that so the question is, is the founder going to live into that truth is, is there a kind of configuration of the company where we are reaping the yield from that first principles situation? That’s why when you have a company with a strong first principles argument and a founder who’s definitely not going to quit, those are the ones that I like. And I’m sure I try to be like that.

Jamie Burke
Interesting. So are you also doing angel investing yourself now?

Richard Craib
I’ve, I’ve only, I’ve only invested in three things. I think aetherium augur and poly chain and I have a good track record. But I also don’t really see myself as an investor, except for the thousands of public equities that I run in my head.

Jamie Burke
Yeah, Greg. Well, look, Richard, it’s been a pleasure talking to you. And I think you know, the fact that you’ve been here since 2015. You’re still going you’re only five years into 150 year journey, but still, it’s it’s a good innings, and you seem to still be smiling, and you have been referenced as one of the nicest guys in the world. I think hopefully the listeners would have that would have come across for the listeners. So, thanks so much for your time and good luck with both Numerai and arrays, you’re back.

Richard Craib
Thanks a lot.

Jamie Burke
If you enjoyed today’s podcast please make sure you subscribe rate and share your feedback to help us reach as many people as possible with the important mission of web three.