Hi guys. Welcome to the Fintech Coffee Break. I’m your host, Isabelle Castro, and today I shared my coffee break with Benjamin Gabay, Co-Founder of CrunchDAO.
CrunchDAO brings together thousands of data scientists and PhDs, pooling their collective ideas and strategies through gamification to create predictive financial insights. Now collaborating with financial institutions, they leverage the decentralized nature of the DAO to build more accurate machine-learning models that can power the institutions’ growth.
Heavily involved in improving understanding of DeFi in the wider community of France, they have also been involved in building the country’s first DeFi certification.
I spoke to Benjamin about why they chose the DAO structure and gamification to approach financial machine learning and predictive insights.
Isabelle Castro – Hi, Benjamin.
Benjamin Gabay – Hi, Isabelle, nice to meet you.
Isabelle – Nice to meet you as well. So welcome to the show.
Benjamin – Thank you.
Isabelle – First of all, I want to ask you what gets you up in the morning?
Benjamin – There are many things. But basically, to be honest, today, it’s only my project because I believe that CrunchDAO is one of the most beautiful things I’ve done. And this is really driving me these days. And so maybe that can be boring for some people, you know, just to think about work. But for me, I think this last year, it was definitely this project.
Isabelle – Nice, nice. It is a beautiful, amazing project. And I’m looking forward to learning more and letting our listeners learn more. So what brought you to founding CrunchDAO?
Benjamin – Okay, so I met my two co-founders, actually three years ago, and they were working in insurtech. So basically, they were building insurance, and they exited their company. At the same time, I was doing the same with my company, it was a marketing company. My co-founder, is an ex-Lehman Brothers employee, he worked also at Deutsche Bank, and he really wanted to go back to his roots, yeah, to what he loved, basically, and my other co-founder is a specialist in machine learning.
And so they started together, building this project, because they thought that machine learning, Artificial Intelligence, and finance could be huge, you know, and we can, can talk about that later. But it’s a very, very challenging thing, in the finance industry. They started together, and then, I joined them a few months later. And this is how the project started.
Isabelle – Nice. So tell me a bit about CrunchDAO’s approach. What does it do?
Benjamin – Yeah, so basically, at Crunch, we are a research team of 2000 data scientists, and a bit more 400 PhDs and we predict the market. That’s what we do. So, basically, every week
our community sends us their predictions that we will assemble, and then we will sell this prediction to financial institutions.
So, that was the I would say version one of CrunchDAO. And now, what we are doing is that we are turning any bank and any financial institution into community-led research company, because they can now give us their data set. So, all their data, and we will provide to our community the data. The Committee will build machine learning models on their data. So basically, now, we are not just selling financial predictions, but we are also powering many financial institutions with our protocols and with our frameworks.
Isabelle – I’ve looked on your website. And you describe the predictive predictions as a tournament. So you’re kind of bringing gamification into the whole predictive insights. Tell me about this, why this approach?
Benjamin – Yes. So, I can explain to you how that works. So, basically, we are using the competition for the data scientists and the machine learning experts, because they love to compete, they love to compete, and this is what is driving us and this is what drives them as well. So, basically, if you are a machine learning expert or two data scientists or quant researcher, then you can work during your week, you know, at the best financial institutions or the best research center, but during the weekends, what you do is that you will come and you will work with us, and you will crunch some data because you like it, you like it, you love the competition, and you want to improve your skills. So this is how we attract the best, you know, and this is working very, very well. So this is why we adopted these tournaments.
Also these tournaments allow us to collect a lot of predictions. So it’s not just about finding the best prediction, but to get as many predictions as we can, because this is the best way it can work in finance.
Maybe I can explain a bit how, what is challenging in the finance industry at the machine learning level. So I mean, artificial intelligence in finance is very challenging, but it yields very good results. And why is it challenging? It is challenging because there is a high signal-to-noise ratio. So it means, for example, if you want to do face recognition, it’s very easy, you know, you provide it with an algorithm, and he can detect the face quite easily, like almost 200%, it can be sure about 100% around that.
But in finance, you’re working over 1000s of parameters. And due to the complexity of the system, so it’s for the stock market is what we call a complex system. And because it’s very, very difficult with a lot of parameters to take into account. And then you need to have a lot of prediction, to extract the signal from the noise. And this takes a lot of time, and is very difficult. There is a lot of research that is needed. And this is why we need to be a lot. And this is why the tournament is very important to us.
Isabelle – Yeah. Now I can see how it kind of how all those 1000s of predictions, kind of feeds in, and then you kind of collate these and then sell them to.
Benjamin – Yes, exactly. So what we do, basically, is that we collect all these predictions that we will collect, we assemble them because today, it’s the best way to understand.
We say to have Alpha, so to get an information coefficient on the market to get financial information, you know, and to extract this information from the market. And then we sell it
to, to the clients basically.
So we have this, but especially today, now they can also just upload their data set on our platform. Of course, of course, the dataset is obfuscated, it means that we can’t know what is inside the dataset. And then the machine learning experts from the community will work on this obfuscated dataset. It’s a pure machine-learning exercise. We just help the financial company, turn their investment ratio into a quantitative strategy, let’s say, and then they can just get 1000s of AI models on the fly to power their own financial institutions.
So yes, we are now at the transition, you know, where before, we are just collecting predictions and setting these predictions to the financial institutions. Now, it’s another step in the DAO, because now it’s not just about selling this prediction. It’s also about hosting other data sets and powering as many financial institutions as we can. So this is huge. And this is new in the market. It’s not I don’t know any other project that does this, like we do.
Isabelle – No, to me, it’s unique. You’re positioned in Web3 using Web3 technologies. Why did you use the DAO structure? Rather than other structures.
Benjamin – So Web3is all about making people move together, about cohesion, you know, and with the token, we can incentivise people to work with us and we can drive this community, you know. So basically, it’s also about the reward system of the DAO.
If you are a data scientist, you come to our platform, and you will download the set, you will do your machine learning exercise, and then you will be rewarded based on your actual performance. We say it’s an out-of-sample performance because maybe you will say Apple will go up in the next 30 days, then we check if Apple goes up in the next 30 days. And then, if it does, then you get points because we have a leader board and then you need to climb the leaderboard, and you’re rewarded according to your place on the leaderboard.
Isabelle – Okay. Okay. And what is it about the DAO that allows you to improve on other methods to produce predictive insights?
Benjamin – Yes. First there is this token part, which is a huge incentive. And second, it’s, the solution is so difficult to build that one is not smart enough, and we need to be 1000s to build these kinds of things. There are so many difficulties that we need to face. And third, this is my last point, what we are doing is, we call it decentralized science. So basically, we do research together. If, for example, at Crunch, I mean, in the Crunch team, we don’t have an answer, then we ask the community, and the community can work on the project can send us their mathematical answer to our questions. And so, we need to work incrementally to build knowledge incrementally. And the DAO structure for this is really well done.
Isabelle – Okay, that’s amazing. You’re also involved with the Paris Dauphine University, and France’s first DeFi certification. Tell me about this. Why did you get involved?
Benjamin – It’s so first about the certificates, I would say. I mean, we did not create it. Of course, it’s a professor called Hervé Alexandre who did that. And I need to say first, that the certificate is really nice. I mean, maybe, if you can check what they’re actually doing during the certificate. It’s beautiful. I would love to do all this stuff. So that’s the first thing.
The second thing is that because the certificate is so nice, and actually they were looking for people to help build the program. And we know Hervé Alexandre. So we decided to partner together to help him build these certificates.
Of course, for us, it’s also very important because it’s because we are a scientific project. So we need also this academic, you know stance, let’s say, and this is why also for us, it’s important.
So this is the reason why we decided to partner with Dauphine to build the certificate.
Isabelle – Okay, nice. Do you feel like it’s like the creation of this kind of certificate is a turning point for France in terms of how they approach Web3 technologies and DeFi?
Benjamin – Yes, I think that it’s, it’s a turning point, because it allows executives in service companies like the Big Four, and all the big consulting companies, and banks to understand Web3, and more specifically DeFi. So basically, even lawyers, they can go to this certificate and go back to their company and implement this with practice. So DeFi practices. So this is a turning point. And it proves that there is a lot of interest.
Also in France, I would say I need to say that. But I think that the government is doing quite well on Web3 and is very proactive. I had the chance to talk to the minister three months ago, four months ago, and I really saw that they really want to push on DeFi and on Web3. So it’s really, really interesting. It’s definitely proof that it’s a turning point.
Isabelle – Do you think it’s going to have much effect on the financial sector having this DeFi element around?
Benjamin – Yeah. Yes, I mean, DeFi will reshape the industry. We are DeFi projects, because we believe because, I mean, we need to define what is DeFi, you know, but just the fact that we are doing decentralized finance, decentralized prediction, this is a revolution for the industry.
And then also the disintermediation, removing centralization. This is also a turning point, and more than that, for example, just to explain how it works for people that don’t know, DeFi but
just to understand the goal of Web3, if you have Uber, or if you have an Airbnb, these platforms are taking a margin, which is 20%-25%. And we say that their margin is our opportunity because we are removing the centralization points. So it’s more money that is distributed to everyone. It’s more fair, and the finance industry will have to deal with that, in one way or another.
Isabelle – Okay, it’s really exciting stuff. What’s one piece of advice that you would give to someone approaching DAO contribution for the first time?
Benjamin – So basically, if you are in the case of CrunchDAO, those who want to come to the DAO what do you do?
In the DAO, socials are very important. The social contribution in those is very important. So I would suggest going to the discord or to the forum or the rocket chat, and to introduce yourself, and then to ask how you can help because we need people. In CrunchDAO, I try to meet every person that enters a discord I asked him to try to meet them all face to face. Because we are an exclusive community. So I want to know who is coming, why he’s coming. What do you want? You know? And then what I tell them is, please introduce yourself in the community to the community because they would be happy, you know. And, yeah, so that would be the only thing I would say. Just get stuck in. Yes.
Isabelle – All right, nice. Now, your curveball question. What are the first three words that come to your head when you think of Web3?
Benjamin – SEC. Regulation. And not world but a sentence ‘Regulation is Good.’
I think many people will not be happy when I’m saying this, I have many people that are jumping at me. But that’s true. Of course, we need regulation. And Web3 is the most beautiful thing we’ve done in the last case because we are able to give back the value of the credit content to the people that are actually creating the content. So that’s the value with Web3. Let’s not throw it that. Let’s not throw it away. I mean, this is a huge game for everyone.
Isabelle – Nice message. How can people get a hold of you? How can people get a hold of you on social media or follow you?
Benjamin – Oh yeah, sorry. Yeah. Of course, on Twitter you can follow us @CrunchDAO. This is the handle, and Telegram also is CrunchDAO. So if you want to enter the Discord just hop into our website crunchdao.com; you can just find the Discord there, and you join, and we’ll be happy to talk to you.
Isabelle – Okay, cool, super simple. Thank you for taking the time to talk to me, and have a good rest of your day.
As always, you can chat with me on my LinkedIn or Twitter at @IZYcastrowrites. But to access great daily content, check out Fintech Nexus on LinkedIn, Twitter, Facebook, or Instagram. You can also sign up for our daily newsletter, bringing news straight to your inbox.
For more fintech podcast fun, check out the website, where you can find more fascinating conversations hosted by Peter Renton and Todd Anderson.
That’s it from me. Until next time, enjoy your downtime.
